TABLE OF CONTENTS

Preface........................................................................................................................................ vi

Overview ..................................................................................................................................... 1
Audience and Objectives.......................................................................................... 1
Key Recommendations ............................................................................................ 2
Energy Impacts.......................................................................................................... 4
Design Guide Organization ..................................................................................... 6

Introduction ................................................................................................................................ 9
Objective ...................................................................................................................... 9
Role of the Designer................................................................................................10
Market Share............................................................................................................11
Integrated Design Issues.......................................................................................14
The Role of Simulation in Design ........................................................................16
HVAC System Selection..........................................................................................22
Location and Size of Airshafts..............................................................................28
Return Air System...................................................................................................30
Auxiliary Loads........................................................................................................33
Design Airside Supply Temperature ..................................................................34
Code Ventilation Requirements ...........................................................................36
Determining Internal Loads..................................................................................37
Simulation and Performance Targets ................................................................47

Zone Issues...............................................................................................................................50
Thermal Comfort .....................................................................................................50
Zoning and Thermostats .......................................................................................52
Demand Control Ventilation (DCV).......................................................................53
Occupancy Controls ................................................................................................57
Window Switches....................................................................................................57
Design of Conference Rooms................................................................................58

VAV box selection....................................................................................................................60

VAV Box Selection Summary................................................................................60
VAV Reheat Box Control.........................................................................................61
Minimum Airflow Setpoints..................................................................................67
Sizing VAV Reheat Boxes ......................................................................................75
Other Box Types.......................................................................................................81
Other Issues..............................................................................................................88

Duct Design...............................................................................................................................93
General Guidelines..................................................................................................93
Supply Duct Sizing ..................................................................................................99

XiX
 Return Air System Sizing....................................................................................104
Fan Outlet Conditions ...........................................................................................105
Noise Control..........................................................................................................106

Supply Air Temperature Control.......................................................................................110

Optimal Supply Air Temperature ......................................................................110
Recommended Sequence of Operation.............................................................112
System Design Issues ..........................................................................................113
Code Requirements ..............................................................................................114

Fan Type, Size and Control .................................................................................................115

Fan Selection Criteria ..........................................................................................115
Visualizing Fan Performance ............................................................................119
Fan Selection Case Studies .................................................................................126
Comparing Manufacturers .................................................................................146
Fan Control .............................................................................................................147
Conclusions............................................................................................................161

Coils and Filters ....................................................................................................................163

Construction Filters..............................................................................................163
Pre-Filters...............................................................................................................163
Final Filter Selection ............................................................................................163
Filter Area...............................................................................................................164
Extended Surface Area Filters ...........................................................................164
Monitoring Filters .................................................................................................164
Coil Selection..........................................................................................................165
Coil Bypass .............................................................................................................167

Outside Air/Return Air/Exhaust Air Control.................................................................168

Control of Minimum Outdoor Air for VAV Systems.......................................168
Isolation Zones ......................................................................................................176
Design of Airside Economizer Systems...........................................................176
Economizer Temperature Control ....................................................................181
Economizer High-Limit Switches ......................................................................182

Appendix 1 – Monitoring Sites...........................................................................................184

Site 1 ........................................................................................................................184
Site 2 ........................................................................................................................186
Site 3 ........................................................................................................................188
Site 4 ........................................................................................................................189
Site 5 ........................................................................................................................190

Appendix 2 – Measured Fan Performance......................................................................192

Energy Benchmark Data .....................................................................................192

xiiX
Appendix 3 – Airflow in the Real World.......................................................................... 196
Interior Zone Airflow........................................................................................... 196
Perimeter Zone Airflow ...................................................................................... 198
System Level Airflow........................................................................................... 200

Appendix 4 – Cooling Loads in the Real World.............................................................. 203

Appendix 5 – DOE-2 Fan Curves ....................................................................................... 206

Appendix 6 – Simulation Model Description.................................................................. 212

Assumptions.......................................................................................................... 212
Results .................................................................................................................... 215

Appendix 7 – VAV Box Minimum and Maximum Flows for 2 Manufacturers ....... 220

Appendix 8 – How to Model Different VAV Zone Controls in DOE-2.2..................... 221

Single Maximum................................................................................................... 221
Dual Maximum-VAV Heating............................................................................. 226
Dual Maximum with Constant Volume Heating............................................. 231
Tips for Accurate Modeling of Zone Controls ................................................ 234

Appendix 9: Simulated Performance of Three Zone Control Sequences ................ 237

Summary................................................................................................................ 237
Schematics of Modeled Zone Control Sequences.......................................... 238
Model Description ................................................................................................ 239
Basecase Results.................................................................................................. 243
Parametric Analysis............................................................................................ 246

Appendix 10: How to Model DCV Control In EQUEST................................................... 252

Fixed Minimum Ventilation Control ................................................................. 252
Multi-zone DCV Using Sum of Zone Air Method ............................................ 256
Multi-zone DCV using Critical Zone Method................................................... 258
Sample eQuest Simulation Hourly Output ...................................................... 259

Appendix 11: Analysis of DCV in Densely Occupied Zone of a VAV System .......... 262
Basecase Modeling Assumptions..................................................................... 263
DCV Modeling Assumptions .............................................................................. 265
Simulation Results............................................................................................... 266

Appendix 12: Fan Systems ................................................................................................ 268

Overview ................................................................................................................ 268
Fan System Models .............................................................................................. 268
Motor Model........................................................................................................... 282
Variable Speed Drive Model............................................................................... 284
Belt Model............................................................................................................... 286

XiiiX
 Fan Type and Sizing.............................................................................................288
Fan Staging and Isolation ...................................................................................288
Supply Pressure Reset.........................................................................................288
Appendix 12 References .....................................................................................288
Saftronic VSD Data................................................................................................290

Appendix 13: Four Inch and Odd Size VAV Boxes.........................................................292

Appendix 14 - Zone Demand Based Reset Case Studies..............................................294

Supply Air Temperature Reset ..........................................................................294
Static Pressure Reset ...........................................................................................298
Static Pressure and Supply Air Temperature Reset Case Study ...............305
Static Pressure Reset Case Study – Underfloor System...............................306
Demand Controlled Ventilation..........................................................................307

Appendix 15 – References ..................................................................................................310

General ....................................................................................................................310
Controls ...................................................................................................................311
Supply Air Temperature .....................................................................................311
Night Flushing .......................................................................................................312
Load Calculations..................................................................................................312
VAV Box Sizing ......................................................................................................312
Fans and Fan Systems.........................................................................................313
Filters.......................................................................................................................314
Outside Air Dampers............................................................................................315
CO2 and DCV ...........................................................................................................315
Project Reports ......................................................................................................316

xivX
This report was prepared by Pacific Gas and Electric Company and funded by California
utility customers under the auspices of the California Public Utilities Commission. Neither
Pacific Gas and Electric Company nor any of its employees and agents:

1. Makes any written or oral warranty, expressed or implied, regarding this report, including,
but not limited to those concerning merchantability or fitness for a particular purpose.

2. Assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of
any information, apparatus, product, process, method, or policy contained herein.

3. Represents that use of the report would not infringe any privately owned rights, including,
but not limited to, patents, trademarks or copyrights.

March 2007

XvX
 PREFACE

The Advanced Variable Air Volume (VAV) System Design Guide (Design Guide) is written
for Heating, Ventilation, and Air-Conditioning (HVAC) designers and focuses on built-up
VAV systems in multi-story commercial office buildings in California.

The Design Guide recommendations include best practices for airside system design,
covering fans, air handlers, ducts, terminal units, diffusers and controls, with emphasis on
getting the air distribution system components to work together in an integrated fashion.
Key topics critical to optimal VAV design and performance are addressed in the following
chapters: 1) early design issues, 2) zone issues, 3) VAV box selection, 4) duct design, 5)
supply air temperature reset, 5) fan type, size and control, 6) coils and filters, and 7) outdoor
air, return air and exhaust air. The intent of the information is to promote efficient, practical
designs that advance standard practice, achieve cost effective energy savings and can be
implemented using current technology.

This is the second edition of the Design Guide. The original was published in October 2003
and was developed as part of the Integrated Energy Systems — Productivity and Building
Science project, a Public Interest Energy Research (PIER) program administered by the
California Energy Commission and managed by the New Buildings Institute. Development
of this second edition was managed by Pacific Gas & Electric as part of the Energy Design
Resources program.

Highlights of new material in this second edition include:

ƒ Additional guidance for simulating VAV system performance
ƒ Updates for 2005 version of Title 24
ƒ Added cost effectiveness analysis for demand controlled ventilation
ƒ Guidance on using window switches for integration with natural ventilation
ƒ Information on stability of VAV box controls at low flow
ƒ Updated guidance on use of electric reheat
ƒ Updated supply air temperature control sequences
ƒ Updated supply air pressure reset control sequences
ƒ More information about dealing with rogue zones
ƒ Updated appendix with DOE2 fan curves

In addition eight new appendices have been added, covering the following topics:
ƒ VAV box minimum and maximum flows for two manufacturers
ƒ Modeling of different VAV control sequences
ƒ Simulated performance of different VAV controls options
ƒ Modeling guidance for demand control ventilation

XviX PREFACE
ƒ Energy savings estimates for demand control ventilation
ƒ Fan system model
ƒ Four-inch and odd-size VAV boxes
ƒ Zone demand based reset case studies

Lead author for development of this second edition was Jeff Stein of Taylor Engineering,
with contributions from Anna Zhou and Hwakong Cheng of Taylor Engineering. Editing
and production assistance were provided by Erik Kolderup, Camren Cordell and Zelaikha
Akram of Architectural Energy Corporation. Andrea Porter and Ken Gillespie of Pacific Gas
& Electric served as project managers. Valuable review and feedback were provided by
Steven Gates, Dan Int-Hout of Kruger, and Kent Peterson of P2S Engineering. Their
assistance is greatly appreciated.

Principal investigator for the first edition was Mark Hydeman of Taylor Engineering.
Research team members included Steve Taylor and Jeff Stein, Taylor Engineering and
Tianzhen Hong and John Arent of Eley Associates (now Architectural Energy Corporation).
The review committee for the first edition included the following: Karl Brown, CIEE;
David Claridge, Texas A&M; Paul Dupont, Dupont Engineering; Ken Gillespie, Pacific Gas
& Electric; Tom Hartman, the Hartman Company; Henry Lau, Southern California Edison;
and David Sellers, PECI, Inc.. Project managers for the first edition were Cathy Higgins,
Program Director for the New Buildings Institute and Don Aumann, Contract Manager for
the California Energy Commission. Additional review was provided by Alan Cowan and Jeff
Johnson, New Buildings Institute.

PREFACE XviiX
 OVERVIEW

Audience and Objectives

The Advanced VAV System Design Guide (Design Guide) is written for HVAC designers
and focuses on built-up variable air volume (VAV) systems in multi-story commercial office
buildings in California. The Guidelines are written to help HVAC designers create systems
that capture the energy savings opportunities, and at the same time feel comfortable that
system performance will meet client expectations. This is a best practices manual developed
through experience with design and commissioning of mechanical and control systems in
commercial buildings and informed by research on five case study projects.

The recommendations address airside system design, covering fans, air handlers, ducts,
terminal units, diffusers, and their controls, with emphasis on getting the air distribution
system components to work together in an integrated fashion.

The Design Guide promotes and employs the concept of early design decisions and
integrated design, meaning that the job of designing and delivering a successful mechanical
system is a team effort that requires careful coordination with the other design disciplines,
the contractors, the owner and the building operators.

A primary emphasis of this manual is the importance of designing systems and controls to be A PRIMARY EMPHASIS
OF THIS MANUAL IS
efficient across the full range of operation. This requires care in the sizing of the system
THE IMPORTANCE OF
components (like terminal units) to make sure that they can provide comfort and code DESIGNING SYSTEMS
AND CONTROLS TO BE
required ventilation while limiting the fan and reheat energy at part load. It also requires EFFICIENT ACROSS
careful consideration of the system controls integrating the controls at the zone to the THE FULL RANGE OF
OPERATION.
controls at the air-handling unit and cooling/heating plants to make the system respond
efficiently to changes in demand.

The Design Guide also presents monitored data that emphasize the importance of designing
for efficient “turndown” of system capacity. Measured cooling loads and airflows for several
buildings show that both zones and air handlers typically operate far below design capacity
most of the time.

The intent of the information is to promote efficient, practical designs that are cost effective
and can be implemented with off the shelf technology.

OVERVIEW X1X
 Key Recommendations

The Design Guide presents recommendations that are summarized per Chapter in Table 1
below.

TABLE 1: Integrated Design 1. Engage the architect and structural engineer early to coordinate shafts
for low pressure air paths.
KEY RECOMMENDATIONS
2. Work with the architect to evaluate glazing and shading alternatives to
mitigate load, glare and radiant discomfort while providing daylight,
views and architectural pizzazz.
3. Prior to starting the mechanical design for any space, first consider the
potential to reduce or minimize the loads on each space.
Early Design Issues 4. Use simulation tools to understand the part-load performance and
operating costs of system alternatives.
5. Employ a system selection matrix to compare alternative mechanical
system designs.
6. Consider multiple air shafts for large floor plates
7. Place the air shafts close to, but not directly under, the air-handling
equipment for built-up systems.
8. Use return air plenums when possible because they reduce both energy
costs and first costs.
9. Design the HVAC system to efficiently handle auxiliary loads that
operate during off hours.
10. Select a design supply air temperature in the range of 52°F to 57°F.
11. Size interior zones for 60°F or higher supply air temperature to allow
for supply air temperature reset in mild and cold weather.
12. Avoid overly conservative estimates of lighting and plug loads.
Zone Issues 13. Consider demand control ventilation in any space with expected
2
occupancy load at or below 40 ft /person.
14. For conference rooms, use either a VAV box with a CO2 sensor to
reset the zone minimum or a series fan power box with zero minimum
airflow setpoint.
VAV Box Selection 15. Use a “dual maximum” control logic, which allows for a very low
minimum airflow rate during no- and low-load periods.
16. Set the minimum airflow setpoint to the larger of the lowest
controllable airflow setpoint allowed by the box (~10% of design flow)
and the minimum ventilation requirement (often as low as 0.15
cfm/ft2).
17. For all except very noise sensitive applications, select VAV boxes for a
total (static plus velocity) pressure drop of 0.5 in. H2O. For most
applications, this provides the optimum energy balance.

X2X OVERVIEW
Duct Design 18. Run ducts as straight as possible to reduce pressure drop, noise, and
first costs.
19. Use standard length straight ducts and minimize both the number of
transitions and of joints.
20. Use round spiral duct wherever it can fit within space constraints.
21. Use radius elbows rather than square elbows with turning vanes
whenever space allows.
22. Use either conical or 45° taps at VAV box connections to medium
pressure duct mains.
23. Specify sheet metal inlets to VAV boxes; do not use flex duct.
24. Avoid consecutive fittings because they can dramatically increase
pressure drop.
25. For VAV system supply air duct mains, use a starting friction rate of
0.25 in. to 0.30 in. per 100 feet. Gradually reduce the friction rate at
each major juncture or transition down to a minimum friction rate of
0.10 in. to 0.15 in. per 100 feet at the end of the duct system.
26. For return air shaft sizing maximum velocities should be in the 800
fpm to 1200 fpm range through the free area at the top of the shaft
(highest airflow rate).
27. To avoid system effect, fans should discharge into duct sections that
remain straight for as long as possible, up to 10 duct diameters from
the fan discharge to allow flow to fully develop.
28. Use duct liner only as much as required for adequate sound
attenuation. Avoid the use of sound traps.
Supply air 29. Use supply air temperature reset controls to maximize economizer
temperature benefit and avoid turning on the chiller/compressor whenever possible.
30. Continue to use supply air reset during moderate conditions when
outdoor air temperature is lower than about 70°F.
31. Reduce the supply air temperature to the design set point, typically
about 55°F, when the outdoor air temperature is higher than about
70°F
Fan Type, Size 32. Use demand-based static pressure setpoint reset to reduce fan energy
and Control up to 50%, reduce fan operation in surge, reduce noise and to improve
control stability.
33. Use housed airfoil fans whenever possible.
Coils and Filters 34. Avoid using pre-filters.
35. Specify final filters with 80 percent to 85 percent dust spot efficiency
(MERV 12).
36. Utilize the maximum available area in the air handler for filters rather
than installing blank-off panels.
37. Use extended surface filters.
38. Consider lower face velocity coil selections ranging from 400 fpm to
550 fpm and selecting the largest coil that can reasonably fit in the
allocated space.
39. Consider placing a bypass damper between coil sections where the
intermediate coil headers are located.
Outside 40. For outdoor air control use a dedicated minimum ventilation damper
Air/Return with pressure control.
Air/Exhaust Air 41. Use barometric relief if possible, otherwise relief fans (rather than
Control return fans) in most cases.
42. For economizer control, sequence the outdoor and return air dampers
in series rather than in tandem.
43. Specify differential drybulb control for economizers in California
climates.

OVERVIEW X3X
 Energy Impacts

For buildings designed with the practices recommended in the Design Guide HVAC
electricity savings are estimated to be reduced 25% below standard practice, corresponding
to 12% of total building electricity consumption. Natural gas heating savings are estimated
to be 41%. Careful design could exceed these savings. Additionally, building owners and
developers can expect reduced maintenance and improved ventilation and occupant comfort.
HVAC ELECTRICITY
SAVINGS ARE
Expected annual savings are about 1.5 kWh/ft2 for electricity and 8.5 kBtu/ft2 for gas, with
ESTIMATED TO BE
25%, corresponding annual utility cost savings are about $0.20/ft2 for electricity and $0.07/ft2 for
CORRESPONDING TO
12% OF TOTAL
gas, based on 2003 PG&E rates. 1
BUILDING
ELECTRICITY
The savings fractions for fan energy (57%), cooling energy (14%), and heating energy (41%)
CONSUMPTION.
that are listed in Table 2 are based on simulations comparing standard practice to best
practice for a 50,000 ft2 office building, with most of the savings from supply air pressure
reset controls and sizing of VAV boxes to allow for 10% minimum flow.

TABLE 2: Standard Best Savings Savings

Practice Practice Fraction
SIMULATION RESULTS
San Francisco (Climate Zone 3)
AND END USE SAVINGS
Cooling (kWh/yr) 111,522 89,428 22,094 19.8%
FRACTIONS
Fan (kWh/yr) 33,231 12,613 20,618 62.0%
Heating (kBtu/yr) 456,000 237,368 218,632 47.9%
Sacramento (Climate Zone 12)
Cooling (kWh/yr) 131,788 120,889 10,899 8.3%
Fan (kWh/yr) 38,158 18,432 19,726 51.7%
Heating (kBtu/yr) 528,800 347,901 180,899 34.2%

Average of San Francisco and Sacramento

Cooling (kWh/yr) 14.1%
Fan (kWh/yr) 56.9%
Heating (kBtu/yr) 41.1%

Typical vs. Best Practice Performance

Significant fan and reheat energy savings are possible through the design strategies promoted
in this Design Guide. The potential savings are illustrated in the graphs below which present
simulation results; in this example the “Standard” case is a reasonably efficient code-
complying system and the “Best” case includes a number of the improvements suggested in
this guideline. The result of this simulation show that fan energy drops by 50% to 60%, and
reheat energy reduces between 30% and 50%.

1
See the Statewide Energy Impact Report (Deliverable 3.4.1), August 2003 at
http://www.energy.ca.gov/reports/2003-11-17_500-03-082_A-22.PDF.

X4X OVERVIEW
This example is by no means comprehensive. For example these savings do not include the
impact of reducing duct pressure drop through careful design, the impact of properly
designing 24/7 spaces and conference rooms, or the potential savings from demand based
ventilation controls in high density occupancies. The assumptions in this example are
presented in Appendix 6 – Simulation Model Description.

Most of the savings are due to the efficient “turndown” capability of the best practices design
and the fact that HVAC systems operate at partial load nearly all the time. The most
important measures are careful sizing of VAV boxes, minimizing VAV box supply airflow
setpoints, controlling VAV boxes using a “dual maximum” logic that allows lower airflows in
the deadband mode, and supply air pressure reset control. Together these provide substantial
fan and reheat savings because typical systems operate many hours at minimum (yet higher
than necessary) airflow. Appendix 6 provides more details about this comparison, and the
importance of turndown capability is emphasized by examples of monitored airflow profiles
in Appendix 3 and cooling load profiles in Appendix 4.

FIGURE 1:
4 12
SAN FRANCISCO
10
3
kBtu/yr/f t2
kWh/yr/ft2

8
Cooling
2 6 Heat
Fan
4
1
2
0 0
Standard Best Standard Best

FIGURE 2:
4 12
SACRAMENTO
10
3
kWh/yr/ft2

kBtu/yr/ft2

8
Cooling
2 6 Heat
Fan
4
1
2
0 0
Standard Best Standard Best

OVERVIEW X5X
 Design Guide Organization

The Design Guide Chapters are organized around key design considerations and
components that impact the performance of VAV systems.

Appendices to the Design Guide present monitored data that emphasize the importance of
designing for efficient “turndown” of system capacity. Measured cooling loads and airflows
for several buildings show that both zones and air handlers typically operate far below design
capacity most of the time.

The diagram in Figure 3 shows the Design Guide content followed by brief descriptions of
each of the Chapters.

FIGURE 3:
OVERVIEW OF
G U I D E L I N E CO N T E N TS

X6X OVERVIEW
Chapter Descriptions

Introduction

The HVAC designer faces many challenges in the design of a high performing HVAC
system. This chapter describes the objective of the guidelines, the role of the designer and the
market share of VAV systems in California.

Early Design Issues

According to an old adage, “An ounce of prevention is worth a pound of cure.” This holds
true for building design. An extra hour carefully spent in early design can save weeks of time
later in the process, not to mention improve client relations, reduce construction costs, and
reduce operating costs.

Zone Issues

Comfort is a complex sensation that reflects the heat balance between the occupant and their
environment but is tempered by personal preferences and many other factors. This chapter
covers zone design issues such as thermal comfort, zoning, thermostats, application of CO2
sensors for demand control ventilation, integration of occupancy controls, and issues
affecting the design of conference rooms.

VAV Box Selection

Selecting and controlling VAV reheat boxes has a significant impact on HVAC energy use
and comfort control. This chapter examines the selection and control of VAV boxes to
minimize energy usage (both fan and reheat) while maintaining a high degree of occupant
comfort. Guidelines are provided for a range of terminal units including single duct boxes,
dual-duct boxes and fan powered terminal units.

Duct Design

Duct design is as much an art as it is a science; however, some rules of thumb and guidelines
are presented to help designers develop a cost-effective and energy-efficient duct design.

Supply Air Temperature Control

This chapter covers the selection of the design temperature set point for VAV systems in the
climates of California. It also addresses energy efficient control sequences for reset of supply
temperature to minimize central plant, reheat and fan energy.

Fan Type, Size and Control

A number of factors need to be considered when selecting fans, including redundancy, duty,
first cost, space constraints, efficiency, noise and surge. This chapter discusses how to select

OVERVIEW X7X
 fans for typical large VAV applications. Information includes the best way to control single
and parallel fans, as well as presentation of two detailed fan selection case studies. Supply air
pressure reset control sequences are discussed in detail.

Coils and Filters

Selection of coils and filters needs to balance energy savings against first costs. This chapter
examines those issues as well as coil bypass dampers.

Outside Air/Return Air/Exhaust Air Control

Ventilation control is a critical issue for indoor environmental quality. Maximizing “free”
cooling through economizers is a cornerstone of energy management. This chapter describes
the design of airside economizers, building pressurization controls, and control for code-
required ventilation in a VAV system.

X8X
 INTRODUCTION

Objective

The intent of the Design Guide is to promote efficient, practical designs that advance
standard practice and can be implemented successfully today. The goal is having HVAC
systems that minimize life-cycle cost and can be assembled with currently available
technology by reasonably skilled mechanical contractors. In some cases, as noted in specific
sections, increased savings might be captured through more advanced controls or with
additional construction cost investment.

This document focuses on built-up VAV systems in multi-story commercial office buildings
in California or similar climates. 2 But much of the information is useful for a wider range of
systems types, building types, and locations. Topics such as selection guidelines for VAV
terminal units apply equally well to systems using packaged VAV air handlers. And
recommendations on zone cooling load calculations are relevant regardless of system type.

This guide addresses airside system design, covering fans, air handlers, ducts, terminal units,
diffusers, and their controls with emphasis on getting the air distribution system components
to work in an integrated fashion. Other research has covered related topics that are also
critical to energy efficiency such as chilled water plant design 3 and commissioning of airside

2
California has 16 climate zones.

3
See CoolTools at www.energydesignresources.com.

INTRODUCTION X9X
 systems. 4 The design of smaller packaged HVAC systems has also been addressed through
another PIER project. 5

Following the practices in this Design Guide can lead to major improvements in system
performance, energy efficiency and occupant comfort.

Role of the Designer

Built-up HVAC systems are complex custom assemblies whose performance depends on a
range of players including manufacturers, design professionals, installing contractors, Testing
and Balancing (TAB) agents, controls technicians and operators. The designer stands in the
midst of this process coordinating the activities of the various entities in producing a product
that works for the owner within the design constraints of time and budget. Due to the
complexity of the process, the lack of easily accessible analysis tools and the limitations in fee
and time, many choices are made based on rules-of-thumb and experience rather than
analysis. In most cases, these factors lead to less than optimal performance of the resulting
system.

Risk is another powerful force influencing HVAC design decisions. The penalty for an
…PRODUCING A
PRODUCT THAT uncomfortable zone is almost always greater than the reward for an optimally efficient
WORKS FOR THE
OWNER WITHIN THE
system. If a system is undersized, the designer may be financially responsible for the
DESIGN CONSTRAINTS remediation, even if it is due to a change in occupancy requirements or problems in
OF TIME AND BUDGET.
installation. Even if the designer avoids these out-of-pocket expenses, he or she will likely
lose future business from an unsatisfied client. As a result, the designer is likely to be overly
conservative in load calculations and equipment selection.

The design of high performing built-up VAV systems is fraught with challenges including
mechanical budgets, complexity, fee structures, design coordination, design schedules,
construction execution, diligence in test and balance procedures, and execution of the
controls and performance of the building operators. 6 With care however, a design
professional can navigate this landscape to provide systems that are cost effective to construct
and robust in their ability to serve the building as it changes through time. The mechanical

4
See The Control System Design Guide and Functional Testing Guide for Air Handling Systems, available for
download at http://www.peci.org/ftguide/.

5
Small HVAC Package System Design Guide available for download at www.energy.ca.gov/pier/buildings or at
http://www.newbuildings.org/guidelines.htm.

6
A great treatise on the issue of barriers to design of efficient buildings is presented in “Energy-Efficient Barriers:
Institutional Barriers and Opportunities,” by Amory Lovins of ESOURCE in 1992.

X10X INTRODUCTION
design professional can also align their services and expertise with the growing interests of
owners and architects in “green” or “integrated design” programs.

These guidelines are written for HVAC designers to help them create systems that capture
the energy savings opportunities, and at the same time feel comfortable that system
performance will meet client expectations. This is a best practices manual developed through
experience with design and commissioning of mechanical and control systems in commercial
buildings and informed by research on five case study projects.

Market Share

Share of Commercial Construction

The California Energy Commission predicts large office building construction volume of
about 30 million square feet per year over the next ten years, equal to 20 percent of new
construction in California. A reasonable estimate is that about one-half of those buildings
will be served by VAV reheat systems. Therefore, these design guidelines will apply to
roughly 150 million square feet of new buildings built in the ten-year period between 2003
and 2012. This estimate equals roughly 10 percent of the total commercial construction
forecast.

FIGURE 4:
School CO M M E R C I A L N E W
5%
CO N ST R UC T I O N
Restaurant Large Office
BREAKDOWN FORECAST
3% 20%
BY FLOOR AREA , TOTAL
Warehouse 157,000,000 FT /YR.
2

18% SOURCE: CALIFORNIA

Small Office E N E R GY CO M M I SS I O N
6%

Hospital
4%

Retail College
16% 3%

Hotel
5%
Other
16% Food Store
4%

Other data sources indicate that the market share of VAV systems could be even higher.
Direct survey data on air distribution system type are not available, but studies indicate that

INTRODUCTION X11X
 chilled water systems account for more than one-third of energy consumption in new
construction 7 and for about 45% of cooling capacity in existing buildings 8. A majority of
these chilled water systems are likely to use VAV air distribution. In addition some of the air-
cooled equipment will also serve VAV systems. Therefore, an estimate of 10 percent of new
commercial construction is likely to be a conservative estimate of the applicability of the
Design Guide and prevalence of VAV systems.

Share of HVAC Market

It is important to note that chilled water systems account for only a small fraction of the total
number of all commercial buildings, roughly 4%. Yet these few number of buildings account
for a large amount of the statewide cooling capacity. Thus, the individuals involved in the
design and operation of these buildings have a tremendous ability to affect statewide energy
use based on the performance of their systems.

A review of PG&E’s 1999 Commercial Building Survey Report (the CEUS data) indicates
the following distribution of HVAC cooling capacity:

Direct expansion systems (55% of total cooling capacity):

ƒ 44% direct expansion
ƒ 10% heat pump

Chilled water systems (45% of total cooling capacity):

ƒ 28% centrifugal chillers
ƒ 15% reciprocating/screw chillers
ƒ 2% absorption chillers

The fraction of the number of commercial buildings with each system type is as follows (note
that sum is greater than 100% because some buildings have more than one type of cooling
system):
ƒ 78% have direct expansion cooling
… SLIGHTLY MORE
THAN 20% OF ALL ƒ 28% have heat pump cooling
COOLING CAPACITY
WOULD BE PROVIDED
ƒ 4% have chilled water cooling systems (including 2% centrifugal chillers, 2%
BY CHILLED WATER, reciprocating/screw chiller, and 0% absorption)
VAV REHEAT
SYSTEMS.
The CEUS data do not indicate the fraction of chilled water cooling system capacity that
also corresponds to VAV reheat systems, but the amount should be at least 50% according to

7
California Energy Commission, 2003

8
Pacific Gas & Electric, Commercial End Use Survey, 1999.

X12X INTRODUCTION
the opinion of industry experts. Based on this estimate then slightly more than 20% of all
cooling capacity would be provided by chilled water, VAV reheat systems.

INTRODUCTION X13X
 EARLY DESIGN ISSUES

According to an old adage, “An ounce of prevention is worth a pound of cure.” This holds
true for building design. An extra hour carefully spent in early design can save weeks of time
later in the process, not to mention improve client relations, reduce construction costs, and
reduce operating costs.

This chapter includes those items that provide the greatest leverage for energy efficient airside
system design. Each of these issues is described in detail in the following sections.

Integrated Design Issues

Traditional design is a fragmented process where each consultant (architect, mechanical

engineer, electrical engineer…) works exclusively on the aspects of the design that fall under
their scope of services. Integrated design is a process that has a more collaborative
multidisciplinary approach to better integrate the building design, systems and controls.

The purpose of this section is to emphasize the importance of teamwork in the design of
high performing buildings. Issues that are not traditionally the purview of the mechanical
designer none the less have great impact on the cost, efficiency and success of their design.
For example the glazing selected by the architect not only impacts the thermal loads but
might prevent occupants in perimeter spaces from being comfortable due to visual glare or
excessive radiant asymmetry. Use of high performance glazing or shading devices can
drastically reduce the size of the mechanical equipment and improve occupant comfort.

Similarly there can be a reduction in project cost and improvement of operation if the
lighting and mechanical controls are integrated in a single energy management and control
system (EMCS). Consider the issue of a tenant requesting lights and conditioning after
hours. With separate systems the tenant would have to initiate two requests, one for the

X14X INTRODUCTION
lighting and another for the HVAC. Similarly the building operator would have to maintain
two sets of software, hardware and parts. The building manager would have to track two sets
of reports for billing. In an integrated system a tenant could initiate a single call to start both
systems, there would be only one system to maintain and one set of records to track.

Achieving optimal air-side efficiency requires more than just selecting efficient equipment
and control schemes; it also requires careful attention to early architectural design decisions,
and a collaborative approach to design between all disciplines. An integrated design process
can improve the comfort and productivity of the building occupants while at the same time,
reducing building operating costs. A high performance building can be designed at little or
no cost premium with annual energy savings of 20%-50% compared to an average building.
Paybacks of only one to five years are common. This level of impact will require a high level
cooperation between members of the design team.

HVAC and architectural design affect each other in several ways. Table 3 identifies a number
of coordination issues as topics for early consideration. While the list is not comprehensive,
it provides a good starting point for discussions between the HVAC designer and architect.

INTRODUCTION X15X
 TABLE 3: Shaft size, Larger shafts reduce pressure loss and lead to lower fan energy. Early coordination
coordination with the Architect and Structural engineer can significantly relieve special constraints
HVAC AND
and location and the resulting system effects at the duct transitions into and out of the shaft. See
ARCHITECTURAL the section titled Location and Size of Air Shafts and the chapter on Duct Design.
COORDINATION ISSUES Air handler Larger face area for coils and filters reduces pressure loss. Adequate space at the fan
size outlet improves efficiency and may allow the use of housed fans, which are usually
more efficient than plenum fans. See the chapter Coils and Filters as well as the
section titled Fan Outlet Conditions in the Duct Design Chapter.
Ceiling height Coordinate early with the architect and structural engineer for space at duct mains
at tight and access to equipment. See the chapters on VAV Box Selection and Duct Design.
locations
Return air Plenum returns are more efficient than ducted returns, but they require fire-rated
path construction. See the Return Air System section in this chapter.
Barometric Barometric relief is more efficient than return fans or relief fans but requires large
relief damper area and has a bigger impact on architectural design. See the chapter
Outside Air/Return Air/Exhaust Air Control.
Outside air Sizing and location of outdoor air dampers are especially important in California due
intake to the savings available from air-side economizer operation. See the chapter Outside
Air/Return Air/Exhaust Air Control
Acoustics Coordinate with the architect, acoustical engineer (if there is one) and owner early to
determine acoustic criteria and acoustically sensitive spaces. Work hard to avoid
sound traps in the design. See Noise Control in the Duct Design chapter.
Window Reduction or elimination of direct sun on the windows offers several benefits in
shading addition to the direct cooling load reduction. Ducts and VAV boxes serving
perimeter zones can be smaller and less expensive due to lower peak air flow
requirements. Perhaps more importantly, the glass will stay cooler, improving the
comfort of occupants near the windows (see the thermal comfort discussion in the
Zone Issues section).
Window Favorable orientation can be the most cost effective solar control measure. Avoid east
orientation or west-facing windows in favor of north facing windows and south facing windows
with overhangs.
Glass type Where exterior shades and/or good orientation are not feasible, use spectrally
selective glazing with low solar heat gain coefficient (SHGC).
Zoning Grouping spaces with similar ventilation requirements, cooling loads and occupancy
schedules can provide first cost savings (due to fewer zones) and energy savings (due
to opportunities to shut off portions of the system). See Zoning and Thermostats in
the Zone Issues chapter.

The Role of Simulation in Design

Standard design and design tools focus on equipment and system performance at “design
conditions,” a static condition that occurs rarely, if at all, in the life of a mechanical system.
In fact, the weather data used for mechanical heating and cooling loads is described by a
metric that indicates how few hours of a typical year that design condition is expected to be
met or exceeded. These design conditions may indicate performance of the mechanical
equipment on peak, but they do not inform the designer on the cost of operating the
mechanical system over the entire year. To understand the operating energy costs of systems
and system alternatives, the designer is strongly encouraged to use simulation tools.

X16X INTRODUCTION
To deliver a high performing system the designer is strongly encouraged to use simulation
tools. These tools assess the annual operation of building systems and design alternatives and
provide a unique perspective of system performance.

Mechanical system operating costs are strongly dependant on the equipment installed, the
equipment’s unloading mechanism, the design of the distribution systems and the way that
equipment is controlled. Consider the complexity of a built-up VAV reheat system. Energy
use is a function of all of the following: the selection and staging of the supply fans; the
selection and control of VAV boxes; the VAV box minimum setpoints; a duct distribution
system whose characteristic curve changes with the response of the economizer dampers and
VAV boxes; economizer design including provision for minimum ventilation control and
building pressurization control; a pressure control loop that varies the speed or capacity of
the fan(s); and possibly a supply temperature setpoint reset loop that changes the supply
temperature setpoint based on demand or some proxy of demand. It would be nearly
impossible to evaluate the annual energy cost impact of the range of design options by hand.

Simulation tools can be used to evaluate system part load operation. The results of the SIMULATION TOOLS
CAN BE USED TO
analysis inform the owner and design team of the importance of a design feature, such as the PERFORM THE
installation of DDC controls to the zone, for example. Research indicates savings can be IMPORTANT
EVALUATION OF
realized of about 50% of the fan system energy by demand-based reset of supply fan pressure SYSTEM PART LOAD
(Hydeman and Stein, 2003). That energy savings, along with the improvement in comfort OPERATION.

and diagnostic ability to detect and fix problems, may be an important part of convincing an
owner to pay the premium for installation of these controls (a premium of approximately
$700/zone over pneumatic or electronic controls) 9.

Simulation can also be used to perform whole building optimization. For example it can
demonstrate the integrated effects of daylighting controls on the lighting electrical usage and
the reduced load on the HVAC systems. It can also be used to assess the reduction in
required system capacity due to changes in the building shell and lighting power density.

So, if simulation tools can help to evaluate and improve designs, what is the resistance in the
marketplace to using them? Here is a list of possible concerns:
1. The tools are expensive.
2. The tools are complex and take too much time to learn.
3. The time that we spend doing these evaluations will not be compensated
in the typical fee schedule.
4. The owner doesn’t really value this extra effort.

This is not a complete list, but it does cover a range of issues. The points below address each
of these in turn.

9
Prices based on cost comparisons of recent projects.

INTRODUCTION X17X
 1. Tool Expense: Simulation tools are no more expensive than other
engineering and office software that engineers currently use, and some
programs do not have any cost at all. The California utilities have
developed a powerful simulation tool called eQuest that is distributed free
of charge (see http://www.energydesignresources.com/tools/equest.html).
Market based products are typically between $800 and $1,500 per license,
a common price range for load calculation tools. Both Trane and Carrier
have simulation tools that can be added to their popular design load
software for an additional cost.
2. Tool Complexity: Many of the current simulation tools have simple
wizard driven front-ends that can be used to quickly develop building
models and descriptions of mechanical systems. Both eQuest (see above)
and VisualDOE (http://www.archenergy.com/products/visualdoe//) have
well developed wizards that allow users to build a multiple zone model in
15 minutes or less. In addition both of these programs can import
AutoCAD DXF files to use as a basis for the building’s geometry. Trane’s
Trace and Carrier’s HAP use the same input as provided to their load
calculation programs to do simulation analysis, and California PIER
research has produced GBXML protocols to link Trane’s Trace to
AutoCAD files (see http://www.greenbuildingstudio.com and
http://www.gbxml.org/). On the horizon, a group of software
programmers are developing a protocol for building industry software
interoperability (called the International Alliance for Interoperability
(IAI), the Building Services Group (BSG), http://www.iai-
international.org/index.html). These protocols have already been
demonstrated linking 3-D CAD programs, thermal load programs,
manufacturer’s diffuser selection software and programs for sizing
ductwork. All of these programs utilize the same geometric description of
the building.
3. Concerns about Time and Fees: Many firms currently perform simulation
analysis as a routine part of their design practice with no increase in design
fees. This is due in part to the advent of simpler software and interfaces,
as well as increased market demand for these services. Both the Green
Building Council’s Leadership in Energy & Environmental Design
(LEED, http://www.usgbc.org) and the California utilities’ Savings By
Design Program (http://www.savingsbydesign.com/) require building
simulation as part of their applications. In the case of the Savings by
Design Program, incentives for the design team can more than make up
for the additional time needed to do simulation. Simulation is also
required for compliance with California’s Title 24 building energy code
when the building fails to meet one or more prescriptive requirements,

X18X INTRODUCTION
 such as if glazing areas exceed the limits of 40% window-to-wall ratio or
5% skylight-to-roof ratio.
4. What Owners Value: Owners value projects that come in under budget,
generate high degrees of occupant satisfaction, and result in few headaches
throughout the life of the building. During the California electricity
curtailments of 2000 and 2001, owners were acutely aware of the
efficiency of their buildings and performance of their mechanical systems.
Owners with mechanical and lighting systems that could shed load did
and appreciated the design features that allowed them to do so. New
utility rates are in development to provide huge incentives for owners with
systems that can load shed on demand from the utility. Although design
fees are paid before the building is fully occupied, relationships are made
or broken in the years that follow. Buildings that don’t work well are
discussed between owners at BOMA (Building Owners and Managers
Association), IFMA (International Facility Managers Association) and
other meetings, and between operators in their union activities and
contractors in their daily interactions with one another. Owners value
buildings that work.

To get high performing buildings, building energy simulation should be an integral part of
design at all phases:

In schematic design (SD), it plays a pivotal role in the selection of mechanical

system (see next section) and in analysis of the building envelope. It can also be a
powerful tool for communicating with architects and owners about sound glazing,
shading and orientation practices that not only reduce energy use but increase
occupant comfort as well.

In design development (DD), simulation can be used to refine design decisions

such as evaluation of subsystem alternatives (e.g., evaporative pre-cooling),
equipment selection, and distribution system alternatives.

In the construction document (CD), phase, simulation is invaluable for

evaluation of control algorithms and compliance with energy codes, rating systems
like LEED, and utility incentive programs like Savings By Design.

In the construction administration (CA), acceptance, and post occupancy,

phases simulation tools can be used to verify system operation and troubleshoot
problems in the field.

The use of simulation tools in the design process is depicted in Figure 5. This figure also
shows the relative roles of simulation and verification in the development of high performing
buildings. Verification in this graphic includes documentation of design intent, design peer
reviews, acceptance tests on systems and post occupancy monitoring and assessment.

INTRODUCTION X19X
 Much of this analysis is supported by the utilities through the Savings By Design program
and verification is in part supported by Pacific Gas & Electric’s Tool Lending Library
program and the California Commissioning Collaborative 10.

esign
FIGURE 5:
gs By D
THE ROLE OF Savin
Concept
S I M U L AT I O N I N D E S I G N
Building

ANALYSIS
SD
Simulation

DD

VERIFICATION
CD

CA

Commissioning
Acceptance

Post-
y
Occupancy ibrar
ing L
Lend
Tool

Using Simulation

What is important in doing simulations for evaluation of mechanical system and

architectural alternatives? How much detail is required? A study on the uncertainty of cost-
benefit analysis for central chilled water plants (Kammerud et. al., 1999) found that the
accuracy of the analysis is a relatively weak function of the actual load profile but a strong
function of both the equipment model accuracies and economic factors (like energy costs,
discount rates, etc.). In schematic and design development studies the overall building
geometry needs to be correct but general assumptions for internal loads and operation
schedules can be used. A reasonably accurate weather file is also needed. The details of the
mechanical system should be as accurate as possible including: the design efficiency of the
equipment; the part-load curves for fans, pumps, cooling and heating equipment; the
controls; the zoning; and the terminal unit settings.

10
The California Commissioning Collaborative The California Commissioning Collaborative is an adhoc group of
government, utility and building services professionals who are committed to developing and promoting viable
building commissioning practices in California. More information can be found at www.cacx.org.

X20X INTRODUCTION
Design for Part-Load Operation

Monitored loads illustrate the importance of designing for efficient part-load operation.
F I G U R E 6 shows that the HVAC system may operate at only one-half of the design
airflow for the bulk of the time. This is quite typical for office buildings. The design
aiflow for the monitored building is 0.83 cfm/ft2. During cool weather, the airflow doesn’t
exceed 0.4 cfm/ft2, and in warm weather airflow is seldom greater than 0.5 cfm/ft2. Figure 7
shows similar results for cooling delivered to that floor. For additional examples, refer to
Appendix 3 – Airflow in the Real World and Appendix 4 – Cooling Loads in the Real
World.

80%
Fraction of Operating Hours (%)

70%
60%
50%
40%
30%
Design Airflow
20% 0.83 cfm/sf

10%
0%
0.0-0.1

0.1-0.2

0.2-0.3

0.3-0.4

0.4-0.5

0.5-0.6

0.6-0.7

0.7-0.8

0.8-0.9

0.9-1.0

1.0-1.1

1.1-1.2

1.2-1.3

1.3-1.4

Airflow (cfm/sf)

Warm (8/8/02 - 9/7/02) Cool (12/12/02-1/11/03)

FIGURE 6:
ME A S UR E D CO O L I N G D E L I V E R E D BY A I R H A N D L E R , SI T E 3
(LIGHT BAR INCLUDES AUG-OCT 2002, DARK BAR COVERS NOV 2002 – JAN 2003)

60%
Frequency (% hrs)

50% Peak AHU Capacity

40% = 4.5 W/ft2
30%
20%
10%
0%
0.1-0.5

0.5-1.0

1.0-1.5

1.5-2.0

2.0-2.5

2.5-3.0

3.0-3.5

3.5-4.0

4.0-4.5

4.5-5.0

5.0-5.5

5.5-6.0

Cooling Load, W/ft2

FIGURE 7:
MEASURED SYSTEM AIRFLOW, SITE 3

INTRODUCTION X21X
 Simulation Guidance

Simulation program documentation and help systems are getting better all the time. The
DOE-2.2 Language Dictionary (http://doe2.com/DOE2/index.html), for example, is an
excellent resource for eQuest users. eQuest and VisualDOE both allow users to quickly run
hourly reports on hundreds of variables. Guidance on how to model some of the measures
recommended in this Guide are included in several sections of the Guide, including:
ƒ Static Pressure Reset – Appendix 5
ƒ Supply Air Temperature Reset – Appendix 6
ƒ Dual Max Zone Controls – Appendix 8
ƒ Demand Control Ventilation – Appendix 10

HVAC System Selection

Mechanical system selection is as much art as science. The choice that the designer makes
must balance a wide range of issues including first cost, energy cost, maintenance effort and
cost, coordination with other trades, spatial requirement, acoustics, flexibility, architectural
aesthetics, and many other issues. First costs depend on local labor rates for various trades,
and operating costs depend on climate and energy costs. Most senior engineers over time
develop a feel for what works based on past experience with the building type, climate,
location, and client requirements. Although this allows them to make a decision on a timely
basis it doesn’t necessarily lead to the right decision in terms of optimal performance. On
the other hand a pure life-cycle cost analysis ignores substantive but hard to quantify issues
like ease of maintenance, occupant satisfaction and architectural aesthetics.
A SYSTEM SELECTION
MATRIX CAN
Like beauty, performance is in the eye of the beholder. What engineers need is a method to
ACCOMPLISH THIS
COMPARISON, compare mechanical system performance over a wide range of quantitative and qualitative
PROVIDING BOTH
QUANTITATIVE AND
issues that can be customized and adjusted to the preferences of particular clients and jobs.
QUALITATIVE A system selection matrix can accomplish this comparison, providing both quantitative and
ASSESSMENTS.
qualitative assessments. An example selection matrix is presented in Table 4. This matrix
allows attributes of different systems to be compared by weighting the importance of each
attribute and providing a ranking of each system with respect to each attribute. The product
of the attribute weight and the system rank for each attribute and each system are then
summed and compared. The higher the total score, the better the system.

The system selection matrix works as follows:

1. Performance attributes (important system performance characteristics) are
listed in the leftmost column. These include the considerations previously
discussed like costs, acoustics and aesthetics.
2. In the next column is a weight representing the relative importance of
each attribute. Selection of these weights will be discussed in detail later.

X22X INTRODUCTION
 3. A short list of alternative systems (typically two to four) is selected by the
engineer in conjunction with the other project team members.
4. For each HVAC system, a rank is assigned for each attribute. The scale
ranges from 1 (worst) to 10 (best). A score of 0 could be used for total
non-compliance. These scores can be on an absolute scale with a rank of
10 representing the perfect system. More commonly a relative scale is used
where the system that performs best for each attribute is awarded a rank of
10 and other systems are ranked relative to that system.
5. A column is also provided for commentary on each system as it applies to
each attribute.
6. The first row (System Description) is provided to give a text description of
each system.
7. The bottom row is the sum of the weight times the rank
(∑ weight i × rank ) for each system.
i

Table 4 provides an example of a selection matrix comparing three systems (single fan VAV
reheat, dual-fan dual-duct VAV and underfloor VAV with VAV fan coils) for a high-tech off
ice building in a mild climate. This example is not a definitive comparison of these three
system types for all applications but is specific to how these system types compared for a
particular application using attribute weights agreed upon by the owner and members of the
design team. The purpose of the example is to illustrate the process.

Table 4 reveals that this project put a high emphasis on first cost, as indicated by the very
high weight (20) assigned to this attribute. By comparison, energy efficiency and
maintenance were assigned weights of only 10 each. Clearly this owner was most concerned
about bringing the project under budget, which is typical of most commercial projects.
Other heavily-weighted categories are impact on the other trades (general contractor),
comfort, and indoor air quality.

The selection of weights is meant to reflect the relative importance of each attribute to the
owner. Although the weights could be assigned at any relative level, the total of the weights
should be limited to 100. This has two important effects: 1) it forces the team to reflect on
the relative importance of the selection criteria, and 2) the weights represent a percentage of
total score across attributes. Often in assigning the weights the team discovers attributes that
are unimportant and can be eliminated.

Walking through the example in Table 4, the first row has the descriptions of the systems
being compared. The second row contains a comparison of the first cost of these three
systems. In our example, this attribute has a weight of 20 (out of 100 total). The VAV
reheat and dual-fan dual-duct VAV systems were awarded the same rank of 8 out of 10. As
indicated in the comments, the core and shell costs for VAV reheat are lower than the dual-
fan dual-duct VAV system but the dual-fan dual duct system has lower zone costs (due in
part to the differential in labor cost between sheet metal and piping). Overall installed costs

INTRODUCTION X23X
 of these two systems are about the same but they are higher than the underfloor system (for
the HVAC costs). The under-floor system has significantly lower core and shell costs, lower
internal zone costs but higher perimeter zone costs. It received a rank of 10 out of 10. For
this row the scores are the weight times the system score, or 160 for the VAV reheat and
dual-duct VAV systems and 200 for the raised floor system.

Adding up the weights times the system ranks for each row produces the final scores in the
last row: 860 for the VAV reheat; 875 for the dual-fan dual-duct VAV system, and; 811 for
the under floor system. The system with the highest score “wins.”

The advantages of this method are:

1. The design team and owner are forced to focus and agree on what system
features are most important for the project. This is embodied in the
weights that are applied to each attribute and in the selection of the
attributes to consider.
2. Both soft and hard factors can be compared in an objective manner.
Scores can reflect relatively precise factors, such as simulated energy
performance and first costs, as well as hard to measure factors such as
perception of comfort.
3. It inherently documents the design intent. It also communicates the
design intent to the other design team members.
4. It has more rigor than simply choosing a system based on “experience.”

Similar matrices can be used to select contractors. Experience has shown that it does not
take much time to set up or evaluate and that owners and architects appreciate the effort. It
also has been a learning experience that sometimes provides unexpected results: what the
designer expects to be the answer is not necessarily the end result in each case. The process
of developing the matrix and filling it in informs designers about the strengths and
weaknesses of various systems and alternatives.

X24X INTRODUCTION
Performance Weight VAV Reheat System Rank Dual Fan Dual Duct Rank Raised Floor System Rank
TABLE 4:
Attribute System
System Central cooling fan systems Central cooling fan systems Central cooling fans supply EXAMPLE SYSTEM
Description on roof supply 55°F to on roof supply 55°F to 60°F 63°F to 65°F air to 14 in. to SELECTION TABLE
60°F air in ceiling air, and central heating fans 18 in. raised floor plenum
mounted ducts to VAV supply 95°F to 100°F air, in using minimal ductwork. Air
reheat boxes in perimeter ceiling mounted ducts to to interior zones is delivered
zones, cooling-only or dual-duct VAV boxes in by individually adjustable
reheat boxes in interior perimeter zones, cooling-only “swirl” diffusers. Perimeter
zones. Return air by or dual-duct boxes in interior zones are served by underfloor
ceiling plenum. Cooling zones. Return air by ceiling variable speed fan-coils that
fans have 100% outdoor plenum. Cooling fans have draw air from the underfloor
air economizers. 100% outdoor air plenum. Return air by
economizers. Heating fans reduced height ceiling plenum
supply 100% return air. or by central shafts with no
ceiling at all. Cooling fans
have 100% outdoor air
economizers.
HVAC First Costs 20 Low shell & core costs. 8 Low zone costs usually offset 8 Elimination of ductwork 10
Highest zone costs. higher shell & core cost typically results in lowest shell
resulting in slightly lower & core costs. Interior zone
overall costs compared to costs lowest due to eliminated
VAV reheat VAV boxes and ductwork.
Perimeter zone costs highest
due to cost of fan-coil and
small zones. Overall costs
should be $1 to $2/ft2 or so
lower than others.
Impact on Other 10 Smallest equipment rooms 10 Larger penthouse space 9 Raised floor raises cost 1
Trades: or wells and shafts. Furred required for heating fans. significantly ($7 to $8/ft2).
General columns required for hot (Net overall add including
Contractor water piping. mechanical and electrical is
about $3/ft2). Penthouse
space similar to reheat system.
Typically more vertical shafts
required.
Impact on Other 5 Fewer units to wire 7 Slightly higher cost 7 Perimeter fan-coils require 10
Trades: mechanically. Poke- compared to reheat due to power. Underfloor wiring
Electrical through system for tenant added heating fan, often reduces tenant improvement
Contractor improvement. offset by eliminating boiler. wiring costs, particularly with
Poke-through system for future revisions.
tenant improvement.

INTRODUCTION X25X
Performance Weight VAV Reheat System Rank Dual Fan Dual Duct Rank Raised Floor System Rank
Attribute System
Floor Space 5 Smallest shafts required. 10 Somewhat larger shafts 9 More shafts required in order 9
Requirements required for additional to properly distribute air with
heating duct. minimal underfloor ducts;
total area slightly larger than
VAV reheat.
Ceiling Space 5 Significant duct space 9 Usually extra heating duct 8 May reduce floor-to-floor 10
Requirements required above ceiling. can fit into same space as height a few inches if exposed
cooling duct (with cross- structure (no ceiling). Works
overs between beams) but very well with concrete flat-
will not work well with flat slab without ceiling.
slab structure.
Energy Efficiency 10 Reheat system causes high 8 Reduced reheat and heat 10 Reduced duct losses provide 9
Normal Operation heating costs. recovery from recessed lights central fan energy savings,
reduces overall energy costs offset somewhat by perimeter
compared to reheat system fan-coil fan energy.
Previously it was thought that
UFAD systems could operate
with higher SAT resulting in
better economizer
performance. It turns out,
however, that SAT similar to
overhead systems are
necessary with UFAD.
Reduced reheat in exterior
spaces due to low minimum
volumes required due to floor
supply.
Energy Efficiency 2 VAV boxes may be used to 10 VAV boxes may be used to 10 No VAV boxes to isolate flow 8
Off-hour isolate unoccupied areas to isolate unoccupied areas to to unoccupied areas. Each
Operation minimize off-hour usage. minimize off-hour usage. floor may be isolated using
smoke dampers.
Smoke Control 3 Outdoor air economizer 10 Same as VAV 10 Same as VAV. 10
(7 story buildings) and relief fans may be used
for smoke control.
Acoustical Impact 5 Noise problems may occur 9 Same as VAV reheat 9 Noise problems may occur 10
near fan rooms and shafts. near fan rooms and shafts.
Slight VAV box noise and Very quiet interior zone
hiss from diffusers supply. Perimeter fan-coils
quieted by heavy floor, low
velocity supply.

X26X INTRODUCTION
 Performance Weight VAV Reheat System Rank Dual Fan Dual Duct Rank Raised Floor System Rank
Attribute System
Indoor Air 10 Outside air economizer 8 Reduced outdoor air supply 7 Excellent ventilation 10
Quality allows 100% fresh air most in winter due to 100% efficiency with floor supply.
of year. return air on heating fan, but Perception of improved air
minimum overall circulation quality in interior zones due
rates can be higher. to control and floor supply
Comfort 10 Good cooling performance 10 Same as VAV reheat 10 Recent experience has shown 5
on exterior zones. Fair that mass coupling, leakage
heating performance due (at air highways, through
to stratification. Can only floor, walls, etc.) and controls
maintain uniform complexity have combined to
temperatures in interior make it more difficult to
open office zones; maintain comfort conditions
individual control not in UFAD buildings.
possible.
Maintenance 10 Only rooftop equipment 8 No water above ceiling 10 No VAV boxes in interior, 7
Costs and requires frequent reduces risk of water damage. but perimeter fan-coils
Reliability maintenance; VAV boxes Dual duct boxes require require most maintenance,
occasional maintenance. slightly less maintenance especially if fitted with
Risk of water damage due than reheat boxes. optional filters. Risk of water
to piping above ceiling. damage due to piping below
floor.
Troubleshooting hot
complaints is more difficult
with UFAD due to variations
in supply air temperature
under the floor.
Flexibility 5 Any number of zones may 7 Any number of zones may be 8 Outlets may be moved easily 10
be used, but at high cost used and zone costs are less to accommodate changing
per zone. than for reheat interior layouts. Air tends to
be naturally drawn to high
heat load areas.
Total 100 860 875 811

INTRODUCTION X27X
 Location and Size of Airshafts

The location and size of airshafts is an extremely important coordination item to begin early
in the design process. The issue can have tremendous implication on the cost and efficiency
of the mechanical systems as well as architectural space planning and structural systems.
Poor shaft design or coordination will result in higher system static pressure and fan energy
use.

There are a couple of general principles to employ in sizing and locating shafts:
1. Keep shafts adjacent to the building cores but as close to the loads as
possible. The architect will generally prefer the shafts near the cores where
there are some distinct advantages for access, acoustics, and servicing.
2. Consider multiple shafts for large floor plates (e.g. greater than 15,000 to
20,000 ft2) and under-floor systems. This can greatly reduce the installed
cost of mechanical systems and reduce problems coordinating services at
the shaft exits.
3. To the extent possible, place the shafts close to, but not directly under, the
air-handling equipment. Leave plenty of space to fully develop airflow
from the fans prior to the ductwork turning down the shaft. As described
in the section on air handlers, the best acoustics result from a lined,
straight horizontal run of duct before turning down the shaft. If using
relief fans or return fans, prevent these fans from having line of sight to
the shaft to minimize fan noise transmission down the shaft.
4. Decide on a return air scheme, either fully ducted from the fan to each
return air grille, ducted only in the riser with the ceiling cavity used as a
return air plenum on each floor, or fully unducted using both the ceiling
cavity and architectural shaft as a return air plenum. See additional
discussion in the following section. This may have an impact on the shaft
area required.
5. Size the shaft for the constraints at the floor closest to the air handler.
This is where the supply, return and exhaust airflows and ducts will be
largest. Shaft size can be reduced as loads drop off down the shaft, but this
is typically only done on high-rise buildings for simplicity.
6. Be conservative when sizing shafts initially. It is always easier to give up
space than expand the shafts in the late stages of design. Also there will
almost always be other items like tenant condenser water piping, reheat
piping, plumbing risers, and toilet exhaust risers that will make their way
into the shaft.
7. Make sure to leave ample room between the supply duct riser and the
shaft wall at riser taps to provide space for a fire/smoke damper and a
smooth transition from the riser into the damper. Typically at least 11 in.

X28X INTRODUCTION
 is required between the inside of the shaft wall and the edge of the duct
riser. This provides 6 in. for a 45° riser tap, 3 in. for the fire/smoke
damper sleeve, and 2 in. to connect the tap to the sleeve with a slip
connection. (See Figure 9.) The more room provided between the tap
and the fire/smoke damper, the lower the pressure drop through the
damper since the air velocity profile will be more uniform through the
damper. However, the longer duct tap blocks the return air shaft and
increases lost shaft space.
8. Coordinate with the structural engineer early on to make sure that the
ceiling space where ducts tap off of risers is not blocked by beams.
Structural engineers will typically select the lightest and deepest steel
beams to reduce steel costs, but where added space is essential such as at
shafts, beams can be made heavier and shallower with only a minor
structural cost impact.
9. Look beyond the inside dimension of a duct or opening. It is critical in
shaft sizing to account for physical constraints like duct flanges, hanging
brackets, transitions, fire damper flanges and fire damper sleeves. If the
shaft is serving as an unducted return air plenum, be sure to account for
the free area lost by horizontal ducts tapping into supply and exhaust
risers (see Figure 8).

FIGURE 8:
TYPICAL DUCT SHAFT
WITH UNDUCTED
RETURN

INTRODUCTION X29X
 FIGURE 9:
TYPICAL DUCT RISER

Return Air System

It is important to establish a return air system design scheme very early in the design process.
It has a significant impact on the cost and complexity of the mechanical system, the size of
the shafts, coordination of fire and smoke zones, space requirements for the penthouses or
mechanical rooms, and operating efficiency of the mechanical system. The three most
common options are:
1. Fully unducted using both the ceiling cavity and architectural shaft as
return air plenums.
2. Partially ducted return, generally ducted from the fan, down the riser, and
part way onto each floor into local return air plenums. (This option may
be used when floors are substantially blocked by full height walls, making
a low pressure fully unducted return more difficult.)
3. Fully ducted return from the fan to each return air grille.

These options will also impact the type of economizer relief system selected. Plenum returns
have a very low pressure drop in general and thus either non-powered (e.g., barometric) relief
or low pressure relief fans may be used. With partially or fully ducted return air systems, the

X30X INTRODUCTION
pressure drop of the return air path will be relatively high, favoring the use of return air fans.
For more discussion on this issue see Taylor, S. “Comparing Economizer Relief Systems,”
ASHRAE Journal, September 2000.

Fully unducted plenum returns have the following advantages:

1. Plenum returns reduce energy usage due to the following factors:

a. Reduced fan static pressure (plenums are essentially a very large ducts) will
reduce fan energy. Typically, plenum returns have static pressure drops in
the range of 0.25 in. to 0.75 in. H2O compared to a range of about 1 in.
to 2 in. for fully ducted returns.

b. Some of the heat gain from recessed lighting and envelope will be picked
up by the return air rather than becoming a space load. This reduces
supply fan energy and, by increasing return air temperatures, it can extend
the effectiveness of airside economizers and improve the efficiency of
packaged cooling equipment.

c. Non-powered relief or relief fans are viable options due to the low pressure
drop of the plenum return, and these types of relief systems use less energy
than return air fans.

2. Plenum returns significantly reduce installed mechanical costs due to the

elimination of all return air ductwork, reduced fan motor and VFD
horsepower, and reduced relief system costs (non-powered relief and relief fans
are less expensive than return fans).

3. Plenum returns are essentially self-balancing and thus obviate the need for
balancing labor. For VAV systems, this feature also ensures that individual
spaces will not be negatively pressurized as supply air flows change. With fully
ducted returns, return airflow does not track supply air flow changes at the
zone, and as a result air balance to spaces and floors varies with changes in
supply airflow.

4. Return plenums typically reduce the required depth of the ceiling space and
shafts can be smaller because the entire free area of the shaft and ceiling are
available for return airflow.

5. Return plenums greatly reduce ceiling coordination among trades by

eliminating the large return air ducts and the need to cross over supply and
return mains to serve zones.

INTRODUCTION X31X
 However, there are some distinct disadvantages to plenum returns:
1. Using building cavities as return air plenums can draw them below
atmospheric pressure if not properly designed, causing outdoor air to be
drawn into the building fabric. In humid climates, this can result in
condensation of moisture from outdoor air within architectural cavities,
and consequently result in mold and mildew growth. Ensuring that
building space pressurization (e.g., 0.05 in.) exceeds the pressure drop
from the space to the return air plenum (e.g., <<0.05 in.) so that all
building elements remain pressurized above ambient air will mitigate this
problem.
2. Most building codes only allow architectural cavities to be used as air
plenums if the materials exposed to the plenum meet certain flame spread
and smoke generation limits. This means that ceiling plenums that are
exposed to wood joists or plywood decks usually cannot be used as return
air plenums.
3. Exposed wiring (electric, control, and telecommunication) must be
plenum rated.
4. Individual space pressurization control is not possible, which is a critical
issue in laboratories and health care buildings.
5. Care must be taken at full height walls to be sure that adequate openings
are provided for return air transfer and that the openings are acoustically
treated where necessary (e.g., with lined elbows or boots).
6. Some indoor air quality (IAQ) experts have concerns that return air
plenums can lead to IAQ problems due to the debris and dust that can
accumulate on ceiling tiles, etc. This same dirt can also accumulate in
return air ducts, of course, but ducts are more easily cleaned than large
plenums. The counter-argument is that the return air plenum is upstream
of particulate filters in the supply air system so dirt entrained in return air
can be substantially filtered out. No studies we are aware of have shown
that return air plenums result in higher particulate concentrations than
ducted return air systems.

Clearly, plenum returns should not be used where codes prohibit them (e.g., due to
combustible structure) or in occupancies where individual space pressures must be controlled
(e.g., hospitals). They also should not be used in humid climates without very careful design
to ensure that all parts of the building remain pressurized. In other applications, return air
plenums are recommended because they reduce both energy costs and first costs.

X32X INTRODUCTION
 Auxiliary Loads

Most buildings will require auxiliary cooling systems to serve 24/7 process loads, such as
server rooms or telecom closets, and other loads that do not operate on the normal HVAC
system schedule. It is important to evaluate the performance of the HVAC system when
serving only these loads, which typically are a small percentage of the total building load.

There are a number of options to serve these loads:

1. Dedicated chilled water fan-coil units. With this design, the chilled water
plant must be able to operate efficiently at the lowest expected auxiliary
load. Typically this will require variable flow (2-way valve) distribution
with variable speed pumps and unequally sized chillers or perhaps a small
“pony” chiller sized for the 24/7 loads alone. Having the smaller chiller
will also generally improve the chilled water plant part-load performance
during low load conditions. Because these loads occur even in cold
weather, energy efficiency can be improved either by installing a water-
side economizer to reduce chiller load and number of operating hours or
by recovering condenser heat to serve heating loads that use low-
temperature hot water (90°F to 110°F), such radiant floor heating or
domestic water pre-heating.
2. Dedicated water-cooled AC units. These units are served from either the
main cooling tower serving chillers (appropriately up-sized) or a dedicated
cooling tower or fluid cooler. Using a heat exchanger or fluid cooler to
create a closed-circuit loop is beneficial from a maintenance standpoint
since it reduces condenser cleaning requirements. If the main cooling
tower is used, tower cells may need to be fitted with weir dams or low-
flow nozzles to allow for adequate water distribution across the tower at
low flow rates. Also, head pressure control must be considered if the
tower is controlled to provide low condenser water temperatures for
optimum chiller operation (e.g., for variable speed chillers) or waterside
economizer operation. This is most easily done by installing head pressure
control valves at each air conditioning (AC) unit or by providing a
controlled tower bypass to the loop serving the AC units. Consider using
waterside economizer pre-cooling coils, offered on water-cooled computer
room units as a standard option. This option is probably the most
common for speculative buildings because it is inexpensive to oversize
towers for future loads, and any number of auxiliary AC units can be
added to the system without affecting efficiency (unlike option 1 above
where the plant will not be efficient unless loads are sufficient to load the
smallest chiller to 25% or so.) Like the first option this system can be
configured to utilize heat recovery. It can also be configured to use a
waterside economizer but that should be carefully evaluated as the extra

INTRODUCTION X33X
 coil cost and air-side pressure drop will offset the benefit of compressor
cooling.
3. Air-cooled split systems. This is not the most efficient option but is
inexpensive for small, distributed loads in low rise buildings. It is not
usually practical in buildings over 5 stories or so due to distance
limitations between rooftop condensing units and fan-coils.
4. VAV boxes from the central VAV air system. This option can either be
the most efficient or least efficient depending on the details of the system
design. To be efficient, first the system must have the ability to shut off
unoccupied areas so that only auxiliary loads can be served without
wasting energy serving unoccupied areas. This is easily done with modern
DDC controls at the zone level; VAV boxes are simply commanded to
close (or temperature setpoints are set back/up and minimum airflow
setpoints set to zero) when spaces are scheduled to be unoccupied.
Second, the central VAV fan system must be evaluated to see if it can
operate stably when only serving auxiliary loads. If the fan system has
variable speed drives, it can operate very efficiently down to about 10% of
design airflow. (Note that some VFDs are configured from the factory
with very high minimum speeds, such as 30% to 50%. These minimum
setpoints should be reduced to 10%, which is all that should be required
for motor cooling.) If the fan system has multiple fans with backdraft
dampers, the fans may be staged to provide efficient operation at even
lower airflow rates. Third, the cooling plant must be capable of operating
efficiently at low loads as described for option 1 above. If all three of
these capabilities are provided, this option can be the most efficient
because fan energy is very low (the variable speed drives will provide cube-
law performance as the airflow drops, partially offset by reduced motor
efficiency at low load) and the central airside economizer can be used to
provide free cooling in cool and cold weather (which is a common
condition during nighttime operation). However, if large areas must be
conditioned to operate the fans or chiller plant stably, this option becomes
the least efficient.

Design Airside Supply Temperature

What’s the best choice for supply air temperature? A designer needs to answer that question
at a fairly early stage in order to calculate airflow requirements and equipment size. The short
answer for VAV systems in California is, “somewhere around 55°F”, which happens to be a
common rule-of-thumb. The long answer is a bit more complex; a designer might say, “it
depends…” It depends on factors like chilled water system efficiency at different
temperatures. It depends on the cost of real estate (i.e. space for shafts and ducts within the

X34X INTRODUCTION
building). It depends on the local climate and the number of potential economizer free
cooling hours and the need for dehumidification. Therefore, choosing an optimal design
temperature can involve a complex tradeoff calculation.

It turns out that “somewhere around 55°F” (e.g. 52°F to 57°F) is a good choice for air THE OPTIMAL SUPPLY
handler design in California office buildings. It results in a good balance between efficiencies TEMPERATURE IS
USUALLY IN THE MID
of the chilled water plant and the air distribution system at peak cooling conditions. The 50S AT PEAK
CONDITIONS.
exact selection is not critically important. If physical space for the air handling equipment is
very constrained or humidity may be a concern during cooling conditions, then choose on
the lower side. Or choose a lower temperature if the building has relatively high loads in
order to avoid the need for excessive peak airflow at the zone level. Otherwise, a temperature
close to 55°F is appropriate, which allows the chilled water plant to operate more efficiently
(through higher chilled water temperature and/or lower chilled water flow). It also reduces
the likelihood that reheat will be required in some zones. A higher temperature also saves
some energy by reducing unneeded latent cooling (this is only a benefit in fairly dry climates)
and by extending the number of hours the economizer can handle the entire cooling load,
which reduces the number of hours the chiller plant operates at low loads.

What happens at higher or lower supply air temperature? If the air handler is selected to
provide higher temperature, say 60°F, at peak periods, then the additional fan energy
typically exceeds the savings from more efficient chiller operation and extended economizer
operation. If the supply air temperature is lower, then fan energy drops while chiller and
reheat energy increases. Systems designed for very low air temperatures (40° to 50°F) are
generally not a good choice in mild California climates. (See Bauman et al.) Low supply air
temperature can be a better choice in warm and humid climates where there are fewer
potential economizer hours and dehumidification is important.

Lower SAT Higher SAT TABLE 5:

Less fan energy, due to lower airflow. However, Less chiller energy due to greater chiller
TRADEOFFS BETWEEN
if series fan-powered boxes are used to prevent efficiency at higher chilled water temperature.
the direct supply of cold air to spaces, fan energy LOWER AND HIGHER
will be higher than 55°F air systems. SUPPLY AIR DESIGN
More dehumidification (desirable in humid Less reheat energy. TEMPERATURE (SAT)
climates, but a potential waste of energy in dry
regions).
Lower construction cost (potential for smaller Potential for higher chilled water delta-T, which
air-side system components that require less leads to lower pumping energy.
space and may be less expensive).
Larger airflow capacity increases opportunity for
economizer savings under mild conditions.
Higher airflow rates which may improve indoor
air quality and comfort.

An important point to note is that supply air temperature reset control is ultimately more
important than the choice of design air temperature. A system designed for 55°F can still
operate at 60°F. Appropriate reset control strategies are described in detail in the section
Supply Air Temperature Control.

INTRODUCTION X35X
 Supply Air Temperature for Interior Zone Sizing

For interior zones, the supply air temperature used for zone airflow calculations and VAV
box sizing should be the fully reset (warmest) temperature. If interior zones are not sized for
a warmer supply air temperature then one or more interior zones could require 55°F air
when the perimeter zones have little cooling load, causing excessive reheat and increased
chiller operation.

Code Ventilation Requirements

For commercial buildings in California other than UBC type “I” occupancies (principally
prisons and hospitals), Title 24 sets the ventilation requirements. Section 121 (b) 2. requires
mechanical ventilation systems be capable of supplying an outdoor air rate no less than the
larger of:

A. The conditioned floor area of the space times the applicable ventilation rate
from Table 121-A; and

B. 15 cfm per person times the expected number of occupants. For spaces
without fixed seating, the expected number of occupants shall be assumed to
be no less than one half the maximum occupant load assumed for exiting
purposes in Chapter 10 of the CBC. For spaces with fixed seating, the
expected number of occupants shall be determined in accordance with Chapter
10 of the CBC.

The outdoor air requirement thus has two components: an occupant-based component and a
building- or area-based component. The design outdoor air rate must be the larger of the
two. Based on the lowest allowed occupancy density assumption allowed by Section 121 (b)
2. A., the code-minimum ventilation rate is calculated for a few common occupancy types in
Table 6.

TABLE 6: Ventilation based on Occupants

CBC Occupant Ventilation Overall
MINIMUM VENTILATION Space Type Load Factor from Expected Occupant Outdoor air based on Area Minimum
(without fixed Table 10-A Load = Twice CBC Rate from Table Ventilation
RATES FOR A FEW 2 2 2 2 2
seating) (ft /person) (ft /person) (CFM/ft ) 121-A CFM/ft Rate (CFM/ft )
OCCUPANCY TYPES
Auditorium 7 14 1.07 0.15 1.07
Conference room 15 30 0.50 0.15 0.50
Classroom 20 40 0.38 0.15 0.38
Office 100 200 0.08 0.15 0.15
Retail - ground
floor 30 60 0.25 0.2 0.25
Retail - upper
floors 60 120 0.13 0.2 0.20
Library - reading
area 50 100 0.15 0.15 0.15

There is a very important exception to Section 121 (b) 2. that states

X36X INTRODUCTION
 EXCEPTION to Section 121 (b) 2: Transfer air. The rate of outdoor air required by
Section 121 (b) 2 may be provided with air transferred from other ventilated spaces if:

A. None of the spaces from which air is transferred have any unusual sources of indoor air
contaminants; and

B. Enough outdoor air is supplied to all spaces combined to meet the requirements of
Section 121 (b) 2 for each space individually.

This exception simplifies the calculation of outdoor air rates by assuming that once
outdoor air is brought into a system, it will be properly distributed to each zone
served by the system. Two results of this exception include:
ƒ The minimum outdoor air to be provided at an air handling system is equal to
the sum of the ventilation requirements of each zone served by the system. So-
called “multiple spaces” effects (see ASHRAE Standard 62-2001 section
6.1.3.1) do not have to be taken into account.
ƒ The minimum rate of air supplied to a space is equal to the minimum
ventilation rate even if the supply air is partly or fully composed of air returned
or transferred from other ventilated spaces. It need not be air supplied directly
from the outdoors.

Even with the simplified assumption at the zone level, the Title 24 ventilation rates
are very similar to the rates from ASHRAE Standard 62.1-2004 so they should
result in acceptable air quality. Also in the mild climates of California the
ventilation rates at the system level often exceed the minimum due to operation of
air-side economizers.

Determining Internal Loads

An understanding of internal loads (lighting, plug loads, and heat from occupants) is
important both for sizing equipment and for determining the required part-load
performance. This section provides guidance on estimating internal heat gains from lighting,
plug loads and occupants. In addition to addressing peak loads, this section also addresses the
frequency and range of internal loads to emphasize the importance of designing systems for
efficient part-load operation. Appendix 4 includes measured loads from several buildings,
providing examples of cooling load profiles.

Oversizing

The effects of equipment oversizing depend on the system or component being considered.
An oversized chiller or boiler will have higher first costs, require more space, have higher
standby losses, and may use more energy than a properly sized unit depending on how it
unloads. As discussed below, an oversized fan may be more efficient at design conditions,
but if it is not properly controlled it will use more energy at part-load and may spend

INTRODUCTION X37X
 significant time in surge. Oversized cooling towers and coils in general will reduce operating
costs but may cause control problems at low loads due to unstable heat transfer
characteristics. Oversized ductwork and pipes will always reduce energy cost but at a first
cost premium. In all cases, oversizing will cost the owner more and gross oversizing will be
easily recognized in the field by observation of equipment performance. Most owners don’t
like paying for equipment that sits idle. However, they also want the flexibility of systems
that can accommodate changes in building loads or operation.

Systems should always be designed to turn down efficiently. This can accommodate
moderate over sizing, reduction of loads due to changes or reductions in tenant spaces and
operation at off designing conditions. Almost all systems will have some amount of over
sizing due to inaccuracies in the load assumptions and techniques, the desire of engineers to
be conservative to avoid liability and the need of owners to have future flexibility. Designs
can accommodate some over sizing and turndown without a significant energy penalty by
doing the following:
ƒ Provide multiple pieces of central equipment (chillers, boilers, towers, fans and pumps)
in parallel to allow staging at low loads. Staging reduces standby losses and inefficient
operation at low loads from motors and fixed speed equipment.
ƒ Using variable speed drives which can effectively reduce equipment capacity
automatically and very efficiently down to very low loads. With variable speed drives
equipment staging is less of an issue.

Simulation can be used to evaluate system operation over the range of anticipated loads and
to test the system performance over a range of design loads.

Lighting Loads

Lighting is the easiest of the internal heat gains to predict. Data regarding the energy
consumed by lighting systems is widely available, and lighting power limits are specified by
Title 24. Even the controls—the one uncertain aspect of lighting load—are fairly easy to
characterize.

There has been a steady downward trend in lighting power, and traditional lighting load
assumptions may no longer be valid (see Table 7). Many office spaces are now designed with
less than 0.8 W/ft2 of lighting. Recent technologies such as higher output “second
generation” T-8 lamps and high efficiency electronic ballasts reduce lighting power by more
than 15% compared to the industry standard T-8s and electronic ballasts. In addition,
motion sensor and daylighting controls are becoming more prevalent due in part to better
controls and field proven technology.

Two numbers are needed to estimate peak lighting loads for sizing calculations: the total
installed lighting power and the diversity factor accounting for controls. The installed
lighting power comes from the lighting plans and fixture schedule (which must indicate the
input power for each type of luminaire). At early stages of design, plans may not be

X38X INTRODUCTION
complete, and in that case the assumed lighting power should be no greater than the code
allowance for each space type. Although this is a start, the final lighting power densities
should be lower and the engineer should request revised numbers from the lighting designer
at each stage in the project development. Ideally, the design team will set lighting power
targets at the beginning of the design process, and there are references available to help
provide reasonably attainable targets that are significantly lower than code allowances.11

Standards Office Building LPD (W/ft2) TABLE 7:

ASHRAE 90.1-2004 1.0
LIGHTING POWER
ASHRAE 90.1-2001 1.3
ASHRAE 90.1-1999 1.3 ALLOWANCES FOR
ASHRAE 90.1-1989 1.5 to 1.9 depends on building size OFFICE BUILDINGS
ASHRAE 90.1-1975 2.2
CA Title 24-2005 1.1
CA Title 24-2001 1.2
CA Title 24-1999 1.2
CA Title 24-1995 1.5
CA Title 24-1992 1.5
CA Title 24-1988 2.0

The diversified lighting loads on central equipment is generally lower than the total installed
power; not all lights are on all the time. Recommendations for lighting load profiles and
diversity factors were developed as part of ASHRAE Research Project 1093 “Compilation of
Diversity Factors and Schedules for Energy and Cooling Load Calculations.” That research
provides a set of schedules and diversity factors for energy simulations (the 50th percentile
schedules) and design cooling load calculations (the 90th percentile schedules). Figure 10
shows the data for offices based on measured data in 32 buildings. The schedules are grouped
by building floor area:
ƒ Small: 1,001 - 10,000 ft²,
ƒ Medium: 10,001 - 100,000 ft², and
ƒ Large: > 100,000 ft².

These results show that the appropriate diversity factor for energy simulations is roughly
80% and for design cooling load calculations is about 90%.

11
Advanced Lighting Guidelines, www.newbuildings.org.

INTRODUCTION X39X
 Weekday Lighting, Small Office Weekend Lighting, Small Office

FIGURE 10: 1.0 1.0

0.8 0.8
MEASURED LIGHTING
0.6 0.6
SCHEDULES (90TH
0.4 0.4
PERCENTILE FOR DESIGN
0.2 0.2
LO A D C A LC U L AT I O N A N D
0.0 0.0
50TH PERCENTILE FOR 1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23

E N E R GY S I M U L AT I O N S ) Hour Hour
load calc. simulation load calc. simulation
FOR SMALL, MEDIUM
AND LARGE OFFICE
Weekday Lighting, Medium Office Weekend Lighting, Medium Office
BUILDINGS – ASHRAE
1.0 1.0
1093-RP
0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23
Hour Hour
load calc. simulation load calc. simulation

Weekday Lighting, Large Office Weekend Lighting, Large Office

1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23
Hour Hour
load calc. simulation load calc. simulation

Where applicable, it is recommended that lighting load calculations include the impact of
daylighting controls. These controls are likely to have the greatest impact during the cooling
peak. If credit is taken for this peak reduction, it is critical to get buy in from the owner and
design team and to clearly document the fact that load calculations assume functioning
lighting controls. Therefore, they can take some equipment downsizing credit for the
lighting controls, and they will understand that eliminating the controls will require that
loads be recalculated. Some may argue that peak load calculations must assume that
automatic daylighting controls are not working (i.e., the lights are always on). This was not
an unreasonable assumption in the early years of daylighting controls due to their notorious
lack of reliability. Modern control systems are, fortunately, much more reliable particularly if
thoroughly commissioned.

Motion sensor lighting controls have a big impact on lighting loads, and though it may not
be appropriate to assume the lights are off for purposes of zone air flow calculations, it is
appropriate to assume some level of diversity at the central equipment.

X40X INTRODUCTION
The recommendations for lighting load assumptions can be summarized as follows:
ƒ Use peak load assumptions no greater than the installed lighting power, or no greater
than the energy code allowance if lighting designs are not complete.
ƒ Encourage the design team to set lighting power targets that are lower than code,
accounting for improvements in lighting technology. Use those targets for load
calculations. Use simulations to show the HVAC system impact and the economic
benefits of a low lighting power density.
ƒ Include a diversity factor for lighting loads because it is rare that all lights are on at the
same time. Include consideration of occupancy sensors if they are part of the lighting
design.
ƒ Account for daylighting control savings in peak load calculations.

INTRODUCTION X41X
 Monitored Lighting Loads

Measured lighting loads (15 minute intervals from 9/14/2001 to 8/15/2002) for Site 1 show
a peak of 0.43 W/ft² for the third floor office area of 32,600 ft², while the installed power is
about 1.2 W/ft2. Therefore, actual lighting power never exceeds about 1/3 of the total
installed power. These results are not representative of all buildings, but they represent what
may be encountered in the field. During the monitoring period, the office areas were only
about 60% occupied and every office had occupancy sensors to control the lights. The
measured profile for weekdays and weekends are shown in Figure 1 1 and Figure 1 2 .

0.7
0.6
0.5
Lighting W/sf

0.4
0.3
0.2
0.1
0.0
1 3 5 7 9 11 13 15 17 19 21 23
Hour

Figure 1 1 :
Measured Weekend Lighting Profile – Site 1 Office Area
Showing Average (line) and Min/Max (dashes)

0.7
0.6
0.5
Lighting W/sf

0.4
0.3
0.2
0.1
0.0
1 3 5 7 9 11 13 15 17 19 21 23
Hour

Figure 1 2 :
Measured Weekday Lighting Profile – Site 1 Office Area
Showing Average (line) and Min/Max (dashes)

X42X INTRODUCTION
Plug Loads

The energy consumed by office equipment such as computers, printers, and copiers is harder
to predict than lighting loads because the quantity of equipment is seldom known with
certainty. However, there is a great deal of information and data available to assist with
estimates. Almost all studies where plug loads were measured show that actual loads are
much lower than what is indicated from nameplate ratings and much lower than commonly
used design values.

An important point to remember is that equipment nameplate power is not the actual power
consumed by the equipment either at peak or part load conditions; it is typically just the
rating of the power supply. The heat gains from internal equipment are always much less
than the nameplate power (Figure 13).

The general recommendation to the designer is that zone airflow be sized to handle
reasonably anticipated peak plug loads (which requires some judgment), but that the more
likely typical plug loads be used to evaluate system performance at normal loads.

FIGURE 13:
OFFICE EQUIPMENT
LOAD FAC TO R
CO MPAR I S O N –
WILKINS, C.K. AND
N. MCGAFFIN. ASHRAE
JOURNAL 1994 -
ME A S UR I N G CO MP UT E R
EQ UI P MEN T LOADS IN
OFFICE BUILDINGS.

Although new commercial buildings have more office equipment installed, the equipment
consumes less energy. For standard PCs and copiers, Energy Star compliant “idle” modes
reduce energy usage when the equipment is unused for a period of time. Loads are also
falling due to the increased use of LCD computer monitors rather than traditional CRT
monitors, laptop computers rather than desktop computers, and shared network equipment
such as printers.

The final report of ASHRAE Research Project 1093 reported that equipment power density
(EPD) ranges from 0.18 to 0.66 W/ft² for office buildings based on measured data from
eight buildings. Another study measured EPDs ranging from 0.4 to 1.1 W/ft² based on 44

INTRODUCTION X43X
 typical office buildings with a total floor area of 1.3 million ft². 12 These data indicate that
typical assumptions of 2 to 5 W/ft² are far off the mark. If no better data are available for
EPD, the following tables provide EPD estimates for different building types from a few
different sources.

TABLE 8: Building Type EPD (W/ft2)

Large (>=25,000 ft2) Small (<25,000 ft2)
EQUIPMENT POWER
Office 1 1
DENSITY (EPD) – US Retail 0.4 0.5
DOE BUILDINGS School 0.8 0.8
ENERGY DATABOOK (ALL Hospital 2.2 2.2
STATES) 2002.

TABLE 9: Building Type EPD (W/ft2)

Assembly 0.25
EQUIPMENT POWER
Office 0.75
DENSITY (EPD) – Retail 0.25
USERS MANUAL FOR Warehouse 0.10
ASHRAE STANDARD School 0.50
Hotel or Motel 0.25
90.1 – 2004 AVERAGE
Restaurant 0.10
RECEPTACLE POWER Health 1.00
DENSITIES (FOR Multi-family 0.75
COMPLIANCE

SIMULATIONS).

TABLE 10: Load Density Load Factor Description

2
of Office W/ft
ASHRAE HANDBOOK
Light 0.5 Assumes 167 ft2/workstation (6 workstations per 1000 ft2)
2005 FUNDAMENTALS, with computer and monitor at each plus printer and fax.
RECOMMENDED EPD Computer, monitor, and fax diversity 0.67, printer diversity
(NOTE THAT THESE
0.33.
Medium 1 Assumes 125 ft2/workstation (8 workstations per 1000 ft2)
VALUES ASSUME CRT with computer and monitor at each plus printer and fax.
MONITORS; THE USE OF Computer, monitor, and fax diversity 0.75, printer diversity
LCD MONITORS WOULD 0.50.
Medium/Heavy 1.5 Assumes 100 ft2/workstation (10 workstations per 1000 ft2)
RESULT IN
with computer and monitor at each plus printer and fax.
SIGNIFICANTLY LOWER Computer, monitor, and fax diversity 0.75, printer diversity
VALUES)
0.50.
Heavy 2 Assumes 83 ft2/workstation (12 workstations per 1000 ft2)
with computer and monitor at each plus printer and fax.
Computer, monitor, and fax diversity 1.0, printer diversity
0.50.

12
Paul Komor, ASHRAE Journal December 1997 paper “Space Cooling Demands From Office Plug Loads.”

X44X INTRODUCTION
As with lighting, ASHRAE research for plug loads has provided hourly diversity factors for
equipment power. Figure 14 shows equipment schedules for office buildings of different
sizes. The recommended diversity factors range from 70% to 95% for load calculations and
range between 40% and 90% for purposes of energy simulations.

Weekday Receptacle, Small Office Weekend Receptacle, Medium Office FIGURE 14:
1.0 1.0 MEASURED EQUIPMENT
0.8 0.8 SCHEDULES (90TH
0.6 0.6 PERCENTILE FOR DESIGN
0.4 0.4 LO A D C A LC U L AT I O N S
0.2 0.2 AND 50TH PERCENTILE
0.0 0.0
FOR ENERGY
1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23
S I M U L AT I O N S ) F O R
Hour Hour
load calc. simulation load calc. simulation SMALL, MEDIUM AND
LARGE OFFICE
Weekday Receptacle, Medium Office Weekday Receptacle, Large Office
BUILDINGS – ASHRAE
1.0 1.0
1093-RP
0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23
Hour Hour
load calc. simulation load calc. simulation

Weekend Receptacle, Small Office Weekend Receptacle, Large Office

1.0 1.0

0.8 0.8

0.6 0.6
0.4 0.4

0.2 0.2
0.0 0.0
1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23
Hour Hour
load calc. simulation load calc. simulation

INTRODUCTION X45X
 Monitored Plug Loads

Measured plug loads (15 minute intervals from 9/14/2001 to 8/15/2002) for site 1 shows a
peak of 0.67 W/ft² for the third floor office area. The profiles are shown for weekdays and
weekends in Figure 15 and Figure 16

0.7
0.6
0.5

Plug W/sf
0.4
0.3
0.2
0.1
0.0
1 3 5 7 9 11 13 15 17 19 21 23
Hour

Figure 15:
Measured Weekday Profile of Plug Power Density – Site 1 Office Area Showing Average (line) and Min/Max
(dashes)

0.7
0.6
0.5
Plug W/sf

0.4
0.3
0.2
0.1
0.0
1 3 5 7 9 11 13 15 17 19 21 23
Hour

Figure 16:
Measured Weekend Profile of Plug Power Density – Site 1 Office Area Showing Average (line) and
Min/Max (dashes)

At Site 5, the daytime (9AM – 6PM) average plug load density is 0.57 W/ft², and 0.35 W/ft2
in nighttime. The computer room has an average load of 2.4 W/ft², which causes the high
nighttime load.

0.7
0.6
Plug Load (W/ft2)

0.5
0.4
0.3
0.2
0.1
0
1 3 5 7 9 11 13 15 17 19 21 23

Figure 17:
Measured Weekday Plug Load Profile of Site 5 (November 1999 – September 2000) Source: Naoya Motegi
and Mary Ann Piette, “From Design Through Operations: Multi-Year Results from a New Construction
Performance Contract”, 2002 ACEEE Summer Study

X46X INTRODUCTION
Occupant Loads

Occupant load assumptions can have a large impact on equipment sizing because it affects
space loads as well as ventilation loads. For a typical office space, the sensible heat produced
by occupants can be as high as 0.75 W/ft2 (equal to 250 Btu/person at 100 ft2/person
density), which is comparable in magnitude to lighting and plug loads. In a high-density
space like a conference room, the occupant heat load can reach 5 W/ft2 (at 15 ft2/person),
which dominates the peak load calculation.

Due to the impact of occupant density assumptions, it is important to make an estimate of

the likely numbers of occupants as well as peak numbers. With those two density estimates it
is possible to ensure that the zone airflow can meet reasonable peak loads while the system
can also operate efficiently under more likely conditions.

Simulation and Performance Targets

Simulation and performance targets can be useful tools to focus a design team and deliver
whole building performance. The most commonly used simulation targets for new
construction are building energy standards, which are referenced by programs such as LEED
and Savings By Design. These programs seek to encourage integrated design by rewarding
energy savings beyond minimum code requirements.

There are other approaches being used to set whole building performance targets. The
University of California is using the past performance of existing buildings to set targets for a
new campus (see sidebar). Other sources of potential targets include benchmarking
programs such as Energy Star or the CalArch database (see sidebar).

A third approach is the E-Benchmark system from the New Buildings Institute, which takes
a step beyond energy codes with a system of basic, prescriptive and “extra credit” design
criteria. This approach utilizes a combination of simulation targets for the design phase and
performance targets for building operations. All of these approaches can be documented
using the Design Intent Tool developed by Lawrence Berkeley National Laboratory (LBNL),
available online (http://ateam.lbl.gov/DesignIntent).

INTRODUCTION X47X
 Performance Targets at the University of California

The University of California took advantage of feedback from existing buildings in developing
a new campus in Merced. Actual peak cooling loads for similar buildings from other campuses
were used as a benchmark for design of new buildings. Design targets for total energy
consumption and peak electric demand were also based on existing buildings. These targets,
listed in Table 11, are based on savings compared to the existing campus average, increasing
from 20% to 35% to 50% between 2004 and 2008.

It should be noted that the targets in this table were adjusted for anticipated space usage and
climate. The reader is referred to the source paper for details on how this was done.

Table 11. UC Merced Building Energy Budgets for Classrooms, Office, and Library Buildings.

Maximum Maximum Annual Maximum Annual

Power Chilled Water Electricity Thermal Thermal
W/gft2 Tons/kgft2 (1) kWh/gft2/yr Th/hr/kgft2 Th/gft2/yr
Opening in 2.9 1.6 12 0.10 0.16
2004
2005 – 2007 2.4 1.3 10 0.08 0.13
2008+ 1.8 1.0 7.6 0.06 0.10
Source: Karl Brown, Setting Enhanced Performance Targets for a New University Campus:
Benchmarks vs. Energy Standards as a Reference?, 2002 ACEEE Summer Study.
Note 1: kgft2 is 1000 gross square feet.

X48X INTRODUCTION
Energy Benchmarking

Tools are available to compare a building’s energy consumption to other similar facilities and
can be helpful in setting performance targets. One such tool is CalArch, an Internet site
allowing a user to plot energy consumption distributions for different building types and
locations within California. See http://poet.lbl.gov/cal-arch/ for California information or see
http://poet.lbl.gov/arch/ for U.S. national data.

Figure 18:
CalArch Benchmarking Tool Results, Office Building Electricity Use Intensity, PG&E and SCE Data (indicated
by different colors) for Total of 236 Buildings

Figure 19:
CalArch Benchmarking Tool Results, Office Building Gas Use Intensity, PG&E Data for Total of 43 Buildings

INTRODUCTION X49X
 ZONE ISSUES

This section covers zone design issues such as thermal comfort, zoning, thermostats,
application of CO2 sensors for demand control ventilation, integration of occupancy
controls, and issues affecting the design of conference rooms.

Thermal Comfort

The placement of thermostats is both crucial to comfort and can greatly affect the
performance of an HVAC system. Numerous reports from the Building Owners and
Managers Association (BOMA) 13 and the University of California’s Center for the Built
Environment (CBE) 14 document that second only to access to elevators, HVAC comfort is a
top concern for tenants and often the reason that they change buildings. Since the
thermostat is the HVAC systems proxy for occupant comfort, it is critical to make sure that
it accurately represents the needs of the occupant.

Comfort is defined in ASHRAE Standard 55 15 as a “condition of mind that expresses

satisfaction with the thermal environment.” It is a complex sensation that reflects a heat

13
See for instance Building Owners and Managers Association (BOMA) International and ULI—the Urban Land
Institute. What Office Tenants Want: 1999 BOMA/ULI Office Tenant Survey Report. Washington, D.C.,
BOMA International and ULI—the Urban Land Institute, 1999. Results of a survey of 1800 office building
tenants across the U.S and Canada. Survey asks respondents to rank the importance of and their satisfaction with
key building features, amenities and services.

14
See the CBE website http://www.cbe.berkeley.edu/RESEARCH/briefs-feedback.htm.

15
ASHRAE Standard 55-2004, Thermal Environmental Conditions for Human Occupancy

X50X ZONE ISSUES

balance between the occupant and their environment, but tempered by personal preferences
and by many environmental and social factors including job satisfaction. There are six
primary factors that affect thermal comfort:
1. Metabolic rate.
2. Clothing insulation.
3. Air temperature.
4. Radiant temperature
5. Air speed.
6. Humidity.

With most commercial HVAC systems, space temperature is the only one of these six factors
that is directly controlled, typically with a wall-mounted thermostat. Humidity is indirectly
limited on the high side as part of the cooling process, and can be limited on the low side
with humidifiers. For the mild, dry climates of California, humidity is not a major factor in
comfort in most commercial buildings.

While temperature and humidity are relatively constant throughout most conditioned
spaces, the radiant temperature may vary significantly from surface to surface. This
variation, called radiant asymmetry, is seldom directly controlled by the HVAC system 16.
Radiant asymmetry can be significant in perimeter offices. An occupant in a west-facing
zone with floor to ceiling single pane glass may be hot in the summer and cold in the winter
almost regardless of the space temperature because of the asymmetric radiant environment.
Luckily, this is less of an issue since Title 24 now requires double pane low-e glass in all
climates. However when dealing with a highly asymmetric radiant environment, the best
strategies, in order of preference, are 1) provide better glazing, less glazing and/or external
shading; 2) use a mean radiant temperature sensor to reset the zone thermostat setpoint.

ASHRAE Standard 55 also recognizes that thermal stratification that results in the air
temperature at the head level being warmer than at the ankle level may cause thermal
discomfort. Therefore, to meet Standard 55 this vertical air temperature difference must be
less than 5.4oF. Poor diffuser selection often results in a space not meeting this requirement.
Supply air temperature and minimum flow rate also affect stratification. These are discussed
in more detail in the sections below on VAV Box Selection and Control.

16
It can be controlled using window shades (internal or external) or through thermostat setpoint adjustment using a
radiant globe sensor. Typically, window shades are provided on the interior of windows and are manually
operated. Radiant space sensors are expensive and rarely applied in the field. Another option is radiant heating
and/or cooling systems.

ZONE ISSUES X51X

 Zoning and Thermostats

Zoning of mechanical systems is determined through a delicate balance between first cost
and comfort. Ideally one zone would be provided per room or workspace, but the cost is
prohibitive for most building owners. The cost/comfort balance typically results in zones of
500 ft2 to 1,200 ft2, encompassing five to 10 workstations per zone. Given that zones cost
between $2,000 and $3,500 per installed VAV box with controls, it is hard to convince an
owner to add an additional $3/ft2 to $6/ft2 to the mechanical system costs to increase the
number of zones. The unfortunate reality is that personal space heaters and fans are often
brought in by tenants to fix zoning problems at a tremendous cost to the owner in energy
bills.

Before ganging rooms or workstations together into a single zone, make sure they have
similar load characteristics. Perimeter zones should only group offices with the same
orientation of glass, and interior spaces should not mix enclosed conference rooms or
equipment rooms with general office space.

Lower cost options to subzoning include the use of self-powered VAV diffusers and the
addition of multiple temperature sensors in a zone. VAV diffusers can individually modulate
the room airflow to provide some level of subzoning. They cost approximately $200 to $250
per diffuser more than a conventional ceiling diffuser. For a space with one diffuser, this
might be a reasonable option for subzoning. Multiple temperature sensors can be used with
a signal selector to allow the room farthest off setpoint to control the box. Most VAV DDC
zone controllers have at least one spare analog input which could be used for additional
temperature sensors. The cost of additional sensors using spare points will be on the order of
$150. If additional sensors require additional DDC zone controllers, the cost can increase to
as much as $1,000/point. If the zone control system is based on LonWorks technology, it is
possible to tie sensors directly to the network at approximately $200/point.

Thermostats should be located on the plans and specified to be mounted at 3’ to 4’ above the
finished floor with gasketing on the control wiring to prevent bias of sensors from air leaking
from the wall behind the thermostat. Avoid mounting thermostats on exterior walls where
exterior heat gains and losses and infiltration can result in false readings. If mounting
thermostats on exterior walls is unavoidable, specify rigid insulation between the thermostat
backplate and the wall. When placing thermostats in the space, review the furniture plans
and avoid locations near heat producing equipment like coffee pots, computers, or copy
machines. Avoid locations that can receive direct solar radiation and require a shield on the
thermostat if this cannot be arranged. A poorly located thermostat guarantees comfort
complaints (unhappy customer) and excessive energy bills (a really unhappy customer).

X52X ZONE ISSUES

 Demand Control Ventilation (DCV)

In the 2005 version of Title 24, demand ventilation controls using CO2 sensors are required
on all single zone systems serving dense occupancies (less than or equal to 40 ft2/person) that
have an airside economizer. This requirement was based on a detailed life-cycle cost
analysis 17. Although not required, almost any VAV reheat zone serving an expected occupant
load denser than about 40 ft2/person can potentially benefit from CO2 control. In fact, a
detailed lifecycle cost analysis has shown that CO2 controls are cost effective in densely
occupied rooms as small as 125 ft2 (see section below on Life Cycle Cost Analysis of DCV in
Conference Rooms). If a space has a CO2 sensor, the minimum ventilation setpoint is set to
the Table 121-A value from Title 24 (0.15 CFM/ft2 for most spaces). The outdoor air rate is
then modulated upward from this lower limit as required to maintain the CO2 concentration
at 1,000 ppm. 18.

With multiple-zone systems, the zone CO2 controls should first increase the airflow rate at
the space then increase the outdoor air rate at the air handler as described in the following
sequence:

At the zone: during Occupied Mode, a proportional-only control loop shall maintain CO2
concentration at 1,000 ppm. The output of this loop (0 to 100%) shall be mapped as
follows: The loop output from 0 to 50% shall reset the minimum airflow setpoint to the
zone from the design minimum up to the maximum cooling airflow setpoint. The loop
output from 50% to 100% will be used at the system level to reset outdoor air minimum.

At the air handler: The minimum outdoor air setpoint shall be reset based on the highest
zone CO2 loop signal from absolute minimum at 50% signal to design minimum at 100%
signal. Design minimum is the sum of the code minimums as if there were no CO2 control.
Absolute minimum is the sum of the following:
1. For zones with CO2 controls: Table 121-A value from Title 24 (e.g., 0.15
cfm/ft2) times the occupied area.
2. For zones without CO2 controls: the design minimum outdoor air rate
which is the greater of the Table 121-A value from Title 24 (e.g., 0.15
cfm/ft2) times the occupied area and 15 CFM per person times the design
number of occupants. See Table 6.

17
For detailed life-cycle cost study refer to the report: Part 1, Measure Analysis and Life-Cycle Cost, 2005
California Building Energy Efficiency Standards. California Energy Commission. P400-02-011. April 2002.

18
Although a setting of 1,100 ppm of CO2 or 700 ppm above ambient is closer to 15 cfm/person at typical
metabolic rates of 1.2 met, the setpoint was reduced to 1,000 ppm in response to concerns raised by CalOSHA.
This is also the historical setpoint established in ASHRAE Standard 1989 before it was updated in 1999.

ZONE ISSUES X53X

 The 2005 Title 24 requirements also include the following requirements where CO2 sensors
are used: One CO2 sensor must be provided for each room that has a design occupancy of
less than 40 ft2/person. The sensors must be mounted between one and six feet above
finished floor, which is the occupant breathing zone. (Locating sensors in return air ducts or
plenums is not allowed since the sensor reading will be skewed by outdoor air leakage into
the duct, mixing with return air from other zones, and possible short-cycling of supply air
from diffusers into return air inlets. See sidebar.) CO2 sensors must have an accuracy of no
less than 75ppm and be certified from the manufacturer to require calibration no more
frequently than every five years

DCV Data

Site 4, a Federal courthouse, has a 31,000-ft2 floor plate with ~23% of the area dedicated to
three courtrooms and/or meeting rooms on a typical floor. The HVAC system was designed
with a fixed minimum outdoor air damper sized for ~0.2 cfm/ft2. The designer put in a CO2
sensor in the return that would reset the economizer if the CO2 levels exceeded 1,100 ppm in
the return air. As part of our project, the team installed new calibrated CO2 sensors in the
return, the outdoor air intake, and in three high occupant density spaces (two meeting rooms
and a courtroom).

During a three-week period of monitoring at this site, we overrode the economizer, forcing the
outdoor air flow to its minimum setpoint, to examine the space CO2 levels and ventilation
system design. For the entire period, the courtroom and meeting rooms were always below
1,100 ppm CO2 even with the economizer closed. During that time there were several days
with the meeting rooms overflowing with people. Title 24 requires a minimum ventilation
rate of about 0.19 cfm/ft2 (80 ft2/person) for courtrooms and approximately 1.0 cfm/ft2 (15
ft2/person) for the meeting rooms. With CO2 controls this could be reduced to 0.15 cfm/ft2
minimum with modulation upward based on zone demand.

This study yields several interesting results:

1. There is enough dilution in this building to handle the peak courtroom and meeting room
occupant densities with an outdoor air intake of only 0.2 cfm/ft2.

2. The VAV box minimums being set at 50% helped the space dilution but at the cost of
large amounts of fan and reheat energy.

3. With demand ventilation (CO2) controls, the minimum outdoor air dampers can be
rebalanced to 0.15 cfm/ft2. This is approximately a 25% reduction in the present
ventilation load.

4. The CO2 sensor in the return was useless. It measured the building’s diluted

X54X ZONE ISSUES

 concentrations of CO2, which did not track the peaks in the individual densely occupied
spaces. Figure 20 shows data from February 7th, 2003, the highest space levels of CO2 in the
three-week monitoring period. The return air CO2 concentration is as low as half the
peak concentration. The 2005 Title 24 explicitly requires sensors to be located at the
breathing level in each space for this reason.

This site is presently retrofitting all of its floors with CO2 sensors in the high density spaces.

th
Figure 20. Measured CO2 Levels At Site #4 on February 7 , 2003.

Modeling DCV in DOE-2

eQuest, the user interface for DOE-2.2, has recently added new features that allow modeling
several ventilation control sequences, including demand control ventilation. Appendix 10
provides guidance on how to model three different ventilation control sequences for multi-
zone systems with densely occupied zones. The three sequences are:
ƒ Fixed Minimum – The ventilation rate at the zones and at the system is fixed at the
design ventilation rate (i.e. based on peak occupancy)
ƒ DCV Using Sum of Zones – The ventilation rate at the zone is the larger of the area
based ventilation requirement (e.g. 0.15 CFM/ft2 and the occupant-based requirement
(e.g. 15 CFM/person), based on the actual number of occupants in each hour. The
System ventilation is the sum of the zone requirements. This sequence complies with
Title 24 2005.

ZONE ISSUES X55X

 ƒ DCV Using Critical Zone – This sequence is similar to the DCV Using Sum of Zones
except that the system OA fraction is equal to X/(1+X-Z), where X is the sum of the
current zone minimum flow rates and Z is the minimum flow rate of the critical zone,
i.e. the zone with the highest ventilation ratio. Thus the system ventilation requirement
will be higher than the sum of the zone requirements and less than or equal to the
ventilation rate of the critical zone. This is similar to the ASHRAE Standard 62-2004
multizone approach.

Life Cycle Cost Analysis of DCV in Conference Rooms

Energy Savings

A detailed simulation analysis was performed to determine the energy savings of the two
DCV control sequences compared to a Fixed Minimum control sequence. As shown in
2
Appendix 11, DCV Using Sum of Zones can save approximately $4/ft on a lifecycle cost
basis for a typical interior conference room when compared to a fixed minimum control
sequence. Savings for DCV Using Critical Zone were approximately $3/ft2.

Installed Costs

In 2005 four controls contractors bidding on a large office building project in Los Angeles
provided unit prices for CO2 control in conference rooms. The contractors were asked to
provide a unit price for installation and commissioning per CO2 sensor, assuming an
approximate count of 80 conference rooms with CO2 control. They were given the
following specification:

CO2 Sensors/Transmitters
ƒ Telaire 8101 or 8102 with cover over display.
ƒ Vulcain 90DM3A
ƒ Equal

The following bids were received.

Contractor Installed cost per CO2 sensor
Contractor A $482
Contractor B $555
Contractor C $685
Contractor D $253
Average $494

Based on an approximate first cost of $500 and energy savings of $4/ft2, DCV control is cost
effective for densely occupied zones greater than 125 ft2.

X56X ZONE ISSUES

 Occupancy Controls

Occupant sensors have come of age. Due to their prevalence in lighting systems, they are
stable in design and reliability and relatively inexpensive. In addition to controlling the
lighting, they can be used to control the occupancy status of individual zones. By setting
back temperature and airflow setpoints when the space served is unoccupied, central fan
airflow is reduced and zone reheat is minimized. Where zones are provided with sub-zone
sensors, the occupant sensor can be used to eliminate the sub-zone sensor reading from the
signal selection controlling the VAV box.

Unfortunately Title 24 requires that zones provide the code-required minimum outdoor
airflow rate when spaces are “usually occupied.” To comply with this, VAV box minimum
airflow setpoints cannot be set to zero in response to an occupant sensor. The box minimum
can be reset to a minimum setpoint equal to the Table 121-A value from Title 24 (e.g., 0.15
cfm/ft2) times the occupied area, and the temperature setpoints can be widened. To allow
spaces to return to comfortable temperatures fairly quickly after they are reoccupied, the
setpoints should not be set more than a few degrees off of occupied setpoints.

Window Switches

Where VAV boxes serve rooms with operable windows, consider the use of position
indicating switches on the windows interlocked with the VAV box controls. This interlock
is similar to the one described under occupant sensors above, but in this case, when the
switch indicates the window is open, the VAV box can be shut off to a zero airflow
minimum (since ventilation is provided by the windows) and setpoints can be extended even
further from occupied setpoints (to ensure energy is not wasted if windows are left open or
opened in extreme weather). Position switches are typically reed switches that operate with a
magnet to indicate the status of a window. They are used extensively in security systems.
The reed switch is typically only a few dollars in cost, the largest cost of the reed switches is
the labor to mount and wire them to the control system. Window manufacturers will often
mount and wire them as part of the window assembly but this requires coordination with the
architect or general contractor who specifies the window assembly.

In 2004 window switches were bid as an alternate on a new university classroom and office
building in Merced California. The wording of the alternate was a follows: “Operable
Window Shut-Off Sensors. Provide low voltage operable window shut-off sensors at
perimeter offices including material and labor for wiring within window frames. Provide
connection of wires to the DDC system along with the appropriate programming for 122
operable aluminum windows.” The price for this alternate from the winning bidder was
$295 per window.

ZONE ISSUES X57X

 A good article on operable windows by Allan Daly was published in HPAC in December
2002 (http://www.hpac.com/GlobalSearch/Article/24213/). This article included a detailed
simulation analysis of operable windows with and without window switches. Some of the
conclusions from this article include the following:
ƒ Non-integrated windows - Buildings with operable windows can use 30-40% more
HVAC energy than buildings without operable windows if the windows do not have
window switches.
ƒ Integrated windows - Buildings with operable windows and window switches can use
30-40% less HVAC energy than a building without operable windows and less than half
the energy of a building with non-integrated operable windows.
ƒ Occupants prefer operable windows and have a wider acceptable temperature range
when they have control over the windows.

Design of Conference Rooms

Conference rooms, because of their variable occupancy and high occupant design densities,
present a challenge to the designer. Minimum ventilation rates at the design occupancy
represent a high percentage of the overall supply air rate, particularly for interior conference
rooms. At low occupancies and low loads, design minimum ventilation rates may be above
the required supply air flow, potentially causing the room to be overcooled. Maintaining
minimum air flow rates and temperature control simultaneously can be done using one of
the following options:
1. VAV Reheat: Set the minimum airflow setpoint on the zone VAV box to the
design occupancy ventilation rate. For interior conference rooms, this
minimum rate will equate to 75% to 100% of the design cooling maximum
supply rate. Clearly, this option wastes fan energy as well as cooling and
heating energy through reheat. It can also require the heating system to
operate even in warm weather to prevent over-cooling conference rooms that
are only partially occupied. If the minimum ventilation rate represents more
than about 40% of the design cooling supply rate, this option is not
recommended. This typically limits the application to perimeter zones with
high solar loads.

2. Occupancy Sensor: Use a VAV box with a high minimum as above, but
integrate it with the lighting system occupant sensor to reduce the box
minimum to the Title 24 Table 121-A level (e.g. 0.15 cfm/ft2) during
unoccupied times. This option is better than option 1 above but it still wastes
energy when the conference room is lightly loaded (less than the design
number of occupants are in the room).

X58X ZONE ISSUES

 3. CO2 Control: Use a VAV box with a CO2 sensor to reset the zone minimum
between the Title 24 Table 121-A level (e.g. 0.15 cfm/ft2) and the design
ventilation minimum. This option uses less mechanical system energy than
the occupant sensor solution because it is effective when the space is partially
occupied as well as unoccupied.

4. Series Fan Box: Use a series fan-powered VAV box with a zero minimum
airflow setpoint. Because Title 24 allows transfer air to be used to meet
ventilation requirements (see Code Ventilation Requirements), minimum
ventilation can be provided by the series-fan supplying only plenum air,
eliminating central air and reheat. This is the simplest option from a controls
perspective and it is one of the most efficient.

Energy Analysis of Conference Room Options

Option 1 (fixed minimum) and Option 3 (DCV control) were simulated to determine the
energy savings from DCV in conference rooms. This analysis is described in detain in
Appendix 11. Unfortunately DOE-2 cannot readily be used to compare the energy
performance of Option 4 (series fan powered boxes) with the other options because fan
powered boxes can only be modeled with a PIU system type and because DOE-2 cannot
model transfer air ventilation. As shown in Appendix 11, DCV control can save
approximately $4/ft2 on a lifecycle cost basis for a typical interior conference room when
compared to a fixed minimum control sequence.

ZONE ISSUES X59X

 VAV BOX SELECTION

Selecting and controlling VAV reheat boxes has a significant impact on HVAC energy use
and comfort control. The larger a VAV box is, the lower its pressure drop, and in turn, the
lower the fan energy. However, the larger VAV box will require a higher minimum airflow
setpoint, which in turn will increase the amount of reheat and fan energy. In addition to
these energy trade-offs, smaller boxes also generate more noise than larger boxes at the same
airflow but they can provide more stable control because they have a greater damper
“authority” or α-value (see ASHRAE Handbook of Fundamentals Chapter 15 for details).
However, within the selection range discussed below, damper authority is seldom a
significant selection consideration.

This section gives guidance on selecting and controlling VAV boxes with hot water reheat.
Other types of VAV boxes (e.g., electric reheat, dual duct, fan-powered) are covered in
sections that follow, but in less detail. This document only applies to VAV boxes with
pressure independent controls19.

VAV Box Selection Summary

The discussion that follows can be summarized as follows, with details in later sections:
1. Use a “dual maximum” control logic, which allows for a very low
minimum airflow rate during no- and low-load periods (see the section
below, “Recommended Approach (Dual Maximum)”.
2. Set the minimum airflow setpoint to the larger of the lowest controllable
airflow setpoint allowed by the box controller (see the section below,

19
Pressure independent controls include two “cascading” (also called master and sub-master) controllers, one
controlling space temperature and one controlling supply airflow rate. The output of the space temperature
controller resets the setpoint of the airflow controller within the maximum and minimum airflow setpoints.

X60X VAV BOX SELECTION

 “Determining the Box Minimum Airflow”) and the minimum ventilation
requirement (see the section below, “Recommended Approach (Dual
Maximum)”).
3. For all except very noise sensitive applications, select VAV boxes for a
total (static plus velocity) pressure drop of 0.5 in. H2O. For most
applications, this provides the optimum energy balance (see the section
below, “Sizing VAV Reheat Boxes”).

VAV Reheat Box Control

Common Practice (Single Maximum)

Common practice in VAV box control is to use the control logic depicted in Figure 21. In
cooling, airflow to the zone is modulated between a minimum airflow setpoint and the
design cooling maximum airflow setpoint based on the space cooling demand. In heating,
the airflow is fixed at the minimum rate and only the reheat source (hot water or electric
heater) is modulated. The VAV box minimum airflow setpoint is kept relatively high,
typically between 30% and 50% of the cooling maximum airflow setpoint (see Code
Limitations”).

Advocates of this approach argue that it:

ƒ Insures high ventilation rates.
ƒ Provides adequate space heating capacity.
ƒ Prevents short circuiting due to stratification in heating mode by keeping supply air
temperature relatively low (e.g., less than 90°F). (Meeting the 5.4oF restriction on
vertical stratification in ASHRAE Standard 55 (See Section on Thermal Comfort above)
may not be possible if supply air temperature is more than 15-20oF above room
temperature)
ƒ Prevents “dumping” by keeping air outlet velocities from getting too low.
ƒ Works for all box direct digital controller manufacturers and control types (i.e.,
pneumatic, analog electronic or digital).

VAV BOX SELECTION X61X

 FIGURE 21:
VAV H OT W AT E R
R E H E AT B O X CO N T R O L -
SINGLE MAXIMUM

Recommended Approach (Dual Maximum)

A more energy efficient VAV box control logic is the “dual maximum” strategy depicted in
Figure 22. In addition to a minimum airflow setpoint and a cooling maximum airflow
setpoint, there is also a heating maximum airflow setpoint; hence the name “dual
maximum”. The heating maximum airflow setpoint is generally equal to the minimum
airflow setpoint in the conventional approach described above; in both cases they would be
determined based on meeting heating load requirements. That allows the minimum airflow
setpoint to be much lower (see “Minimum Airflow Setpoints”).

The control logic of the dual maximum approach is described by the following sequence of
operation:
1. When the zone is in the cooling mode, the cooling loop output is mapped
to the airflow setpoint from the cooling maximum to the minimum
airflow setpoints. The hot water valve is closed.
2. When the zone is in the deadband mode, the airflow setpoint shall be the
minimum airflow setpoint. The hot water valve is closed.
3. When the zone is in the heating mode, the heating loop shall maintain
space temperature at the heating setpoint as follows:
a. From 0%-50% loop signal, the heating loop output shall reset the
discharge temperature from supply air temperature setpoint (e.g., 55°F) to
90°F. Note the upper temperature is limited to prevent stratification
during heating.

b. From 50%-100% loop signal, the heating loop output shall reset the zone
airflow setpoint from the minimum airflow setpoint to the maximum

X62X VAV BOX SELECTION

 heating airflow setpoint. The supply air discharge temperature remains at
90°F.

4. The hot water valve shall be modulated using a PI control loop to

maintain the discharge temperature at setpoint. Note that directly
controlling the hot water valve from the zone temperature PI loop is not
acceptable since it will not allow supply air temperature to be under
control and limited in temperature to prevent stratification.
5. The VAV damper shall be modulated to maintain the measured airflow at
setpoint.

FIGURE 22:
VAV H OT W AT E R
R E H E AT B O X – D U A L
MAXIMUM

While the shaded area (which is proportional to the magnitude of the reheat energy) in
Figure 21 and Figure 22 may not appear to be very different, the difference can be quite
significant on an annual basis since VAV boxes typically spend much of their time in the
deadband and mild heating modes. For example, suppose a zone has a cooling design
maximum of 1.5 CFM/ft2. With a single maximum VAV box control and a 30% minimum,
0.45 CFM/ft2 would be supplied in deadband. With a dual maximum VAV box control and
a properly selected minimum (see “Minimum Airflow Setpoints”), this rate could drop to
about 0.15 CFM/ft2. In this case, the single maximum results in three times more airflow
and three times more reheat energy than the dual maximum approach in all but the coldest
weather.

The arguments supporting the dual maximum approach include:

ƒ It allows for much lower airflow rates in the deadband and first stage of heating while
still maintaining code ventilation requirements. This reduces both reheat energy and
fan energy.

VAV BOX SELECTION X63X

 ƒ By reducing the deadband minimum airflow rate, spaces are not over-cooled when there
is no cooling load and “pushed” into the heating mode.
ƒ By controlling the reheat valve to maintain discharge supply temperature rather than
space temperature, supply air temperature can be limited so that stratification and short
circuiting of supply to return does not occur. This improves heating performance and
ventilation effectiveness (see Figure 22). It also keeps the HW valve under control at all
times, even during transients such as warm-up. With two-way valves, this makes the
system completely self-balancing, obviating the need for balancing valves and associated
labor. (See also Taylor, S.T. Balancing Variable Flow Hydronic Systems, "ASHRAE
Journal, October 2002.) 20

Disadvantages include:
ƒ Only a few direct digital control manufacturers that have “burned-in” programming in
their controllers (often called “preprogrammed” “configurable” controllers) offer dual
maximum logic as a standard option. However, there are many fully programmable
zone controllers on the market and all of them can be programmed to use this logic.
ƒ There is a greater airflow turndown and potential risk of dumping and poor air
distribution with improperly selected diffusers. See “Minimum Airflow Setpoints”.
ƒ While ventilation codes are met, airflow rates are reduced which results in higher
(although acceptable) concentrations of indoor contaminants.

Other Dual Maximum Sequences (Not recommended)

The recommended dual maximum control sequence is as a “Temperature First” sequence

because as heating loop output increases the supply air temperature setpoint is reset first and
then the airflow setpoint is reset. It is also referred to as a “VAV Heating Sequence” because
the air flow varies in heating mode and is not constant.

Dual Max with Constant Volume Heating

Figure 23 depicts a dual maximum sequence with constant volume heating. As cooling load
decreases the airflow is reduced from the maximum airflow (on the far right side of the
figure) down to the minimum flow. Then as heating is required the airflow is reset from the
minimum to the heating maximum and the reheat valve is modulated to maintain the space
temperature at setpoint. This sequence has the advantage over the recommended sequence
of not requiring a discharge air temperature sensor. The major drawback with this sequence
is that it results in considerably more reheat than the recommended sequence. It also tends
to make the zone ‘stick’ in the heating mode. If the zone is already in heating then the zone
will have a relatively a large minimum flow ratio, with the reheat coil modulating on top of
that. As you approach the very top of the heating throttling range (low heating load),

20
In a traditional control sequence, the maximum call for heating would open up the heating valve fully. During
warm-up, the coils closest to the pump would likely take more than their design share of the hot water flow,
partially starving the coils furthest from the pump. By controlling leaving air temperature instead of valve
position each reheat coil is limited to its design flow.

X64X VAV BOX SELECTION

reheating will approach zero, and that large volume of cool air will act to cool the zone,
keeping it in the heating band. A relatively large cooling load is required to “un-stick” the
zone from heating mode. See Appendix 9 for a detailed energy analysis comparing this
sequence to the recommended sequence.

Maximum FIGURE 23:

Airflow Setpoint
DUAL MAXIMUM ZONE
Reheat Valve Position CO N T R O L SEQ U E N CE
WITH CO N STAN T
VO L U M E H E AT I N G

50%
Airflow Setpoint

20%

Heating Loop Cooling Loop

Dead
Band

Dual Maximum with VAV Heating - Valve and Airflow Together

This sequence (see Figure 24) is a variation on the recommended sequence. The reheat valve
position and the airflow setpoint are reset simultaneously in heating. Since the valve is not
fully open until the airflow is fully reset it is assumed that the supply air temperature will not
be too high and thus a supply air temperature sensor is not required. The disadvantage of
this sequence over the recommended sequence is that it results in more slightly more reheat
energy and there is a risk that supply air temperature will be considerably higher or lower
than desired. With primary air supply air temperature reset, hot water temperature reset and
variations in hydronic system pressure (as valves open and close throughout the system) the
supply air temperature in heating could vary considerably. This sequence cannot be modeled
in DOE-2 but empirically it is obvious that it results in higher reheat energy than the
proposed sequence.

VAV BOX SELECTION X65X

 FIGURE 24:
DUAL MAXIMUM WITH Max Cooling
100% Airflow Setpoint
V A V H E AT I N G - V A LV E Reheat Valve Position
AND AIRFLOW TOGETHER

50%

Airflow Setpoint

20%

Heating Loop Cooling Loop

Dead
Band

Simulation Guidance: Simulating Single Max and Dual Max in eQUEST

Modeling dual maximum zone control correctly in eQuest is not trivial. It requires that
several keywords be set correctly at both the system level and the zone level. It also requires
careful checking of the simulation results to insure that the model is behaving appropriately.
Detailed guidance on how to set up a dual maximum model and how to troubleshoot it is
presented in Appendix 8.

Simulation Results

Using the methods described above, a detailed simulation analysis was performed to compare
the energy performance of the recommended dual maximum sequence (with 20% min flow)
to a conventional single maximum sequence (with 30% minimum flow) and to the dual
maximum with constant volume heating sequence (with 20% min flow and 50% heating
flow). Extensive parametric runs were performed to determine the effect of load profile,
equipment oversizing, supply air temperature reset, climate and other factors on the savings.
This analysis showed that the recommended dual maximum sequence saves about 10
cents/ft2-yr for a typical office building. The Dual Max-Constant Volume never saves as
much as the Dual Max-VAV and in many cases uses more energy than the Single Maximum.
On average, the Dual Maximum-CV is no more efficient than the Single Maximum control
sequence.

X66X VAV BOX SELECTION

 Minimum Airflow Setpoints

Code Limitations

Title 24 places limits on both the lowest and highest allowable VAV box minimum airflow
setpoints.

The lowest allowable setpoints are those required to meet ventilation requirements (see Code
Ventilation Requirements). Note that since Title 24 allows air transferred or returned from
other ventilated spaces to be used for ventilation, the minimum airflow setpoint need not be
adjusted for the fraction of “fresh” air that is in the supply air. In other words, if the
minimum ventilation rate is 0.15 cfm/ft2, then the minimum airflow setpoint may be set to
that value even if the supply air is not 100% outdoor air, provided the design minimum
outdoor air at the air handler is delivered to some other spaces served by the system (again,
see Code Ventilation Requirements).

Title 24 Section 144 limits the highest allowable minimum airflow setpoints in order to
minimize reheat energy. In Section 144, the minimum setpoint is mandated to be no greater
than the largest of the following:
1. 30% of the peak supply volume; or
2. The minimum required to meet the ventilation requirements of Section
121; or
3. 0.4 cubic feet per minute (cfm) per square foot of conditioned floor area
of the zone; or
4. 300 cfm.

In common practice, VAV box minimums are set much higher than even this code limit,
and much higher than they need to be. In the buildings surveyed for this document, the box
minimums ranged between 30% and 50% of design airflow (see Table 12). Unfortunately,
this common practice significantly increases reheat fan, and cooling energy usage.

Site Average Type TABLE 12:

#1 No data VAV BOX MINIMUMS
#2 28% +/- 19% VAV reheat with dual maximums FROM FIVE MEASURED
#3 30% VAV interior with parallel fan-powered boxes with electric reheat
SITES.
#4 50% VAV reheat with single maximum
#5 40% VAV reheat with single maximum

With the dual maximum strategy (see “Recommended Approach (Dual Maximum)”), the
minimum airflow setpoint need not be based on peak heating requirements. To minimize
energy usage while still complying with Title 24 ventilation requirements, the minimum
airflow setpoint should be set to the greater of:

VAV BOX SELECTION X67X

 1. The minimum airflow at which the box can stably control the flow (see
“Determining the Box Minimum Airflow”); and
2. Ventilation requirement (see “Code Ventilation Requirements”).

Although the dual maximum strategy saves energy, meets the Title 24 Section 144
requirements and maintains code required ventilation, some engineers remain concerned
about the following issues:
ƒ Minimum air movement and stuffiness
ƒ Diffuser dumping and poor distribution problems
ƒ Air change effectiveness

These concerns are largely anecdotal and unsupported by research, as shown in the following
paragraphs.

Minimum Air Movement and Stuffiness

According to ASHRAE Standard 55-2004 there is no minimum air speed necessary for
thermal comfort if the other factors that affect comfort (drybulb temperature, humidity,
mean radiant temperature, radiant and thermal asymmetry, clothing level, activity level, etc.)
are within comfort ranges. People routinely experience this at home: they can be perfectly
comfortable with no air movement (windows closed, furnace and AC unit off) yet for some
reason many HVAC engineers insist that these same people need air movement at work.
They use this to justify higher minimum airflow setpoints (e.g., 0.4 CFM/ft2, the maximum
allowed by Title 24).

There are virtually no studies that support this perception, however. Even if perceptible air
motion was associated with comfort, higher airflow rates out of a given diffuser are unlikely
to increase perceived air velocities in the occupied region simply because the velocities are
below perceptible levels even at full airflow by design − that is, after all, what diffusers are
designed and selected to do.

Simply put, studies to date show fairly conclusively that complaints of “stuffiness” and poor
air motion are not due to lack of air movement but instead indicate that spaces are too warm.
Lower the thermostat (e.g., to <72°F) and the complaints almost always go away.

Dumping and Poor Distribution

Another concern when using a relatively low box minimum is degradation of diffuser
performance. There are two potential issues with low minimums: stratification and short-
circuiting in heating mode (see discussion of air change effectiveness) and dumping in
cooling mode. A diffuser designed for good mixing at design cooling conditions may
“dump” at low flow. Dumping means that the air leaving the diffuser does not have
sufficient velocity to hug the ceiling (the so-called Coanda effect) and mix with the room air
before reaching the occupied portion of the room. Instead, a jet of cold air descends into the

X68X VAV BOX SELECTION

occupied space creating draft and cold temperatures which in turn creates discomfort. The
industry quantifies diffuser performance with the Air Diffusion Performance Index (ADPI).
Maintaining nearly uniform temperatures and low air velocities in a space results in an ADPI
of 100. An ADPI of 70 to 80 is considered acceptable. The ASHRAE Handbook of
Fundamentals gives ranges of T50/L for various diffuser types that result in various ADPI
goals. L is the characteristic room length (e.g., distance from the outlet to the wall or mid-
plane between outlets) and T50 is the 50 FPM throw, the distance from the outlet at which
the supply air velocity drops to 50 feet per minute. For a perforated ceiling diffuser, the
Handbook indicates that acceptable ADPI will result when T50/L ranges from 1.0 to 3.4.
This basically means that best turndown possible while still maintaining an acceptable ADPI
is 1/3.4 = 30% turndown. Other types of diffusers have greater turndown. A light troffer
diffuser, for example, can turndown almost to zero and still maintain acceptable ADPI.

Note that ADPI tests are always done under a cooling load. For all diffuser types, the lower
the load, the greater the turn-down percentage while still maintaining acceptable ADPI. The
lowest load catalogued in the ASHRAE Handbook of Fundamentals is 20 Btu/h/ft2, equal to
roughly 1 cfm/ft2 which is a fairly high load, well above that required for interior zones and
even well shaded or north-facing perimeter zones. To achieve good air distribution when the
load is substantial, maintaining diffuser throw is important. However, when the low airflow
rates occur with the dual maximum strategy, loads are by definition very low or zero. Under
these conditions, acceptable ADPI may occur with even zero airflow. Again, consider
experiences in the home: temperatures around the home can be very uniform with no air
circulation when AC and heating equipment is off at low or no loads.

Concern about dumping may be overblown (no pun intended). There are many buildings
operating comfortably with lower than 30% airflow minimums. Researchers at UC Berkeley
and Lawrence Berkeley National Laboratory performed several laboratory experiments with
two types of perforated diffusers and two types of linear slot diffusers (Fisk, 1997; Bauman,
1995). They measured air change effectiveness (using tracer gas) and thermal comfort (using
thermal mannequins) in heating and cooling mode and at various flow rates (100%, 50%,
and 25% turndown). They also measured throw and space temperature and velocity
distribution from which they calculated ADPI. They found that in cooling mode ADPI
depended more on the diffuser type than the flow rate. For example, the least expensive
perforated diffuser had an ADPI of 81 at 25% flow. They also found that in nearly all
cooling tests thermal comfort was within the acceptable range and air change effectiveness
was consistently at or above 1.0.

There is also a reasonable chance that if there is little or no load in a space then there are no
people in the space, in which case dumping at minimum flow is a non-issue. Conversely,
there may be relatively few hours when there are people in the space and the airflow setpoint
is at its minimum.

VAV BOX SELECTION X69X

 Air Change Effectiveness

Air change effectiveness measures the ability of an air distribution system to deliver
ventilation air to the occupied (breathing) zone of a space. A value of 1.0 indicates perfect
mixing; the concentration of pollutants is nearly uniform. A value under 1.0 implies some
short-circuiting of supply air to the return. Values greater than 1.0 are possible with
displacement ventilation systems where the concentration of pollutants in the breathing zone
is less than that at the return. Studies have shown that air change effectiveness is primarily a
function of supply air temperature, not diffuser design or airflow rates. Measurements by all
major research to-date (e.g. Persily and Dols 1991, Persily 1992, Offerman and Int-Hout,
1989) indicate that air change effectiveness is around 1.0 for virtually all ceiling
supply/return applications when supply air temperature is lower than room temperature.
Bauman et al 1993 concluded that “a ceiling mounted supply and return air distribution
system supplying air over the range 0.2 to 1.0 cfm/ft2 [1.0 to 5.0 L/s.m2] was able to provide
uniform ventilation rates into partitioned work stations. The range of tested supply volumes
represented rates that were below and above the [diffuser] manufacturer’s minimum levels
for acceptable performance.” Fisk et al 1995 concluded that “when the supply air was
cooled, the [air change effectiveness] ranged from 0.99 to 1.15, adding to existing evidence
that short-circuiting is rarely a problem when the building is being cooled.” This study was
based on air flow rates ranging from 0.2 to 0.5 cfm/ft2 (1.0 to 2.5 L/s.m2) using linear slot
diffusers as well as two types of inexpensive perforated diffusers.

These studies indicate that low air change effectiveness is only an issue in heating mode; the
higher the supply air temperature above the space temperature, the lower the air change
effectiveness. This suggests that the low minimum airflow setpoints we propose will result in
lower air change effectiveness for a given heating load since the supply air temperature must
be higher. But air change effectiveness will stay around 1.0 if the supply air temperature is
no higher than about 85°F 21. With the dual maximum approach with the hot water valve
controlled to maintain supply air temperature (rather than directly from room temperature),
the supply air temperature can be limited below 85°F, thus mitigating or even eliminating
this problem. Note that some zones may require higher supply air temperatures to meet
peak heating load requirements. If so, the problem will be the same for both the dual
maximum and conventional single maximum approach since at peak heating (the far left side
of the control diagram), both have the same airflow setpoint. For these spaces, fan-powered
mixing boxes can be used to increase heating airflow rates while at the same time limiting
supply air temperatures below 85°F and maintaining low minimum airflow setpoints to
minimize reheat losses.

Engineers and operators who may not be convinced by these arguments are encouraged to
experiment with low minimums to see for themselves if problems occur. Minimum airflow

See ASHRAE Standard 62, Addendum 62n, Table 6.2. 85°F limit assumes 70°F space temperature (15°F ΔT).
21

X70X VAV BOX SELECTION

setpoints are easily adjusted up to higher levels if comfort complaints do arise. There are
many buildings in operation with this form of control and high degrees of perceived
comfort.

As mentioned previously one limitation on the minimum for the VAV box is the
controllability of the box. This section discusses how the designer can determine this value.

Determining the Box Minimum Airflow

VAV box manufacturers typically list a minimum recommended airflow setpoint for each
box size and for each standard control options (e.g. pneumatic, analog electronic, and
digital). However, the actual controllable minimum setpoint is usually much lower than the
box manufacturer’s scheduled minimum when modern digital controls are used.

The controllable minimum is a function of the design of the flow probe (amplification and
accuracy) and the digital conversion of the flow signal at the controller (precision). These
issues are elaborated in the following paragraphs:

The flow probe is installed in the VAV box and provides an air pressure signal that is
proportional to the velocity pressure of the airflow through the box. Flow probes, which are
typically manufactured by and factory installed in the VAV box by the box manufacturer, are
designed to provide accurate signals even when inlet conditions are not ideal (e.g. an elbow
close to the inlet) and to amplify the velocity pressure signal to improve low airflow
measurement. The amplification factor varies significantly by VAV box manufacturer and
box size. The greater the amplification, the lower the controllable minimum. The VAV box
manufacturer must balance this benefit with other design goals such as minimizing cost,
pressure drop, and noise.

The accuracy of the box controller in converting the velocity pressure signal from the
probe to a control signal. To make this conversion, digital controls include a transducer to
convert the velocity pressure signal from the probe to an analog electronic signal (typically 4-
20 mA or 0-10 Vdc) and an analog-to-digital (A/D) converter to convert the analog signal to
“bits,” the digital information the controller can understand. To stably control around a
setpoint, the controller must be a able to sense changes to the velocity pressure that are not
too abrupt. One controller manufacturer recommends a setpoint that equates to at least 14
bits. For this manufacturer’s controller, which uses a 0-1.5 in. transducer and a 10 bit A/D
converter, 14 bits equates to about 0.004 in. pressure at the input of the transducer. With a
similar transducer and an 8-bit A/D converter, the pressure would be about 0.03 in.

The steps to calculate the controllable minimum for a particular combination of VAV box
and VAV box controller are as follows:
1. Determine the velocity pressure sensor controllable setpoint, VPm in inches
of water (in.w.c,) that equates to 14 bits. This will vary by manufacturer
but for lack of better information, assume 0.004 in. for a 10-bit (or

VAV BOX SELECTION X71X

 higher) A/D converter and 0.03 in. for an 8-bit A/D converter. Ask the
VAV box controller manufacturer for the specification of the transducer
and A/D converter. 22
2. Calculate the velocity pressure sensor amplification factor, F, from the
manufacturers measured CFM at 1 in. signal from the VP sensor as
follows:

2
⎛ 4005 A ⎞
F =⎜ ⎟
⎜ CFM @ 1" ⎟
⎝ ⎠

where A is the nominal duct area (ft2), equal to:

2
⎛D⎞
A =π⎜ ⎟
⎝ 24 ⎠

where D is the nominal duct diameter (inches).

See Figure 25 for an example of manufacturer’s velocity sensor data. The

data on the right size of the graph are the airflows at 1 in. for various neck
sizes (shown on the left). For example using this figure, this
manufacturer’s sensor has 916 cfm at 1 in. signal with an 8 in. neck.
3. Calculate the minimum velocity vm for each VAV box size as:

VPm
vm = 4005
F

Where VPm is the magnified velocity pressure from Step 1.

4. Calculate the minimum airflow setpoint allowed by the controls (Vm) for
each VAV box size as:

Vm = vm A

22
If basing box selection on the performance of a 10 or 12 bit A/D converter, be sure to specify this in the
specification section on control hardware. This will somewhat limit the manufacturers that can provide the box
controls. Guidance on manufacturers’ product offerings can be found on the Iowa Energy Centers, DDC Online
Site at http://www.ddc-online.org/.

X72X VAV BOX SELECTION

 FIGURE 25:
SAMPLE VAV BOX INLET
SENSOR PERFORMANCE
C H A RT, C F M VS .
VELOCITY PRESSURE
SIGNAL

We’ll illustrate these calculations with an example. Table 13 shows the minimum airflow
setpoint Vm for the VAV box probe depicted in Figure 25 with a controller capable of a
0.004 in. velocity pressure setpoint.

VAV BOX SELECTION X73X

 TABLE 13: Nominal Min VP
Inlet Sensor CFM @ Minimum Minimum
SAMPLE CALCULATION
Diameter, reading, 1 in. Amplification Velocity, Flow,
OF BOX MINIMUM in. Area, ft2 in. w.g. w.g. factor FPM CFM
FLOW. D A VPm F vm Vm
4 0.087 0.004 229 2.33 166.02 14
5 0.136 0.004 358 2.33 166.02 23
6 0.196 0.004 515 2.33 166.02 33
7 0.267 0.004 702 2.33 166.02 44
8 0.349 0.004 916 2.33 166.02 58
9 0.442 0.004 1,160 2.33 166.02 73
10 0.545 0.004 1,432 2.33 166.02 91
12 0.785 0.004 2,062 2.33 166.02 130
14 1.069 0.004 2,806 2.33 166.02 177
16 1.396 0.004 3,665 2.33 166.02 232
22 2.64 0.004 7000 2.28 167.71 443

Recent Research on Accuracy and Stability of VAV Box Controls at Low Flow

A recent research project by performed by Darryl Dickerhoff and Taylor Engineering has
corroborated that VAV box controls are both stable and reasonably accurate at flow rates
described herein (i.e. at flow probe signals of 0.004 in.). These researchers tested VAV boxes
from two manufacturers and box controllers from 4 manufacturers. Thus a total of 8
box/controller combinations were tested. The following is an extract from the executive
summary of that report:

The stability and accuracy of a VAV box depends on two main components: the flow probe
(provided by the box manufacturer), and the zone controller (typically provided by a separate
controls manufacturer). These components were tested separately and as an assembly to
determine the contribution of each component to any potential stability or accuracy issues.

8 inch VAV boxes from two manufacturers (Titus and Nailor) were tested under a variety of
conditions including flows ranging from 85 FPM (0.001 in. probe signal) to 2000 FPM (0.5
in. signal), inlet pressures ranging from 0.1 in. to 1.5 in. and damper positions ranging from
nearly closed to fully open. The flow probes alone were found to be stable and accurate
under all conditions with no loss of amplification or signal quality.

Controllers from four manufacturers (Siemens, Alerton, Johnson and ALC) were set up and
calibrated by technicians from the respective manufacturers. They were first tested under a
variety of conditions to determine how stably and accurately the controller could measure a
known velocity pressure signal that fluctuated over time. Each of the four controllers were
then tested on both of the VAV boxes to test how accurately and stably the controllers could
maintain a given flow setpoint while the inlet pressure was fluctuating.

Stability was not an issue for any of the controllers. All controllers were able to track
fluctuating inlet pressure signals and all had good filters for smoothing “noisy” pressure
signals. All controllers were also able to maintain very low flow setpoints without excessive
damper adjustments, even when faced with fluctuating inlet pressures.

X74X VAV BOX SELECTION

The two controllers with hot wire type flow sensors (Alerton and ALC) were both very
accurate at the calibration points (e.g. 0 CFM and 600 CFM for Alerton) but were found to
under estimate actual flow at flow rates above the lowest calibration point. Thus the
controller will always err on the side of supplying a little more than the desired minimum
flow at very low setpoints so there is little risk of undersupplying at minimum flow. The hot
wire controllers did not exhibit zero-drift and there does not appear to be a need to re-zero
these controllers on a regular basis.

The pressure-based sensors (Siemens and Johnson), were highly accurate immediately after
calibration (e.g. ±10% at 0.003 in. signal) but exhibited zero-drift issues. The Siemens
controller zero-drift appears to be directly correlated with ambient temperature: 0.002
in./oF. By default the Siemens controller re-zero’s the sensor twice a day (by shutting the
damper) and thus is highly accurate immediately after re-zeroing but can drift quickly if
ambient temperature drifts. Siemens offers an optional pressure shorting bypass valve to
measure the zero more frequently without disturbing the flow. This auto-zero bypass was
tested (set to re-zero every hour) and was found to result in very accurate control at signals to
0.003 in.. Without the auto-zero bypass, reasonable accuracy (±15% of reading) was
achieved at setpoints down to about 0.01 in. (300 FPM)

Unlike the Siemens controller, the Johnson (JCI) controller did not exhibit zero-drift even
when the ambient temperature drifted. It did however, exhibit a different problem: by
default the JCI controller re-zero’s the damper every two weeks. When it performed this
function (four times during testing) it did something wrong and incorrectly measured the
zero flow resulting in sensor errors up to approximately 0.003 in. (perhaps it measured zero
before or after it fully closed the damper?). The cause of this error is not clear (JCI’s product
manager was notified but did not respond as of the writing of this report) but it seems likely
that the error is due to either an installation error or a software bug and that with correct
installation and/or software fix it is expected that reasonable accuracy can be expected with
the JCI controller at 0.005 in. signal. JCI also offers an optional bypass valve. This bypass
valve was not tested but it is believe that it would result in good accuracy at setpoints down
to at least 0.003 in.

Sizing VAV Reheat Boxes

The key consideration in sizing VAV reheat boxes are determining the box minimum and
maximum airflows for each neck size for a given product line. The minimum airflows are
determined by the ventilation and controllability issues addressed in the previous section,
“Determining the Box Minimum Airflow.” The maximum airflow rate the box can supply is
determined from the total pressure drop and sound power levels as discussed below. For a
given design airflow rate, more than one box size can meet the load, so the question is which
size to use.

VAV BOX SELECTION X75X

 Design Maximum Airflow Rate

Before a selection can be made, the design airflow rate must be determined from load
calculations. Caution should be taken to determine these loads accurately as VAV box
oversizing can lead to significant energy penalties particularly if the conventional single
maximum logic (see “VAV Reheat Box Control”) is used. For example, assume a VAV box
is selected for 1000 cfm with a 30% (300 cfm) minimum. If the box is actually oversized by
a factor of 2, then the true design airflow rate is 500 cfm and the effective minimum setpoint
is not 30% but 60%, almost a constant volume reheat system. For most operating hours,
this box will operate at its minimum airflow rate and temperature will be controlled be
reheating the cold supply air.

Noise

VAV box manufacturers provide two types of sound data: discharge and radiated. Discharge
noise is rarely an issue if the box has hard duct on the inlet, a lined outlet plenum and flex
duct between the plenum and diffusers. As a general rule, VAV boxes located above standard
acoustical ceilings should have radiated Noise Criteria (NC) levels no more than ~5 NC
above the desired room NC rating. For example, a typical office application with a desired
NC level of 30, the VAV box should be selected for a 35 NC.

Note that the assumptions used by manufacturers in determining resulting NC levels should
be checked to make sure they apply (see catalog data and ARI rating assumptions). If not,
then a more complex calculation using radiated sound power data must be done.

It is important to base the selection on the latest sound power data for the particular box
being used. One of the most important contributors to box noise is the design of the flow
sensor, which differs from one manufacturer to the next. Since the manufacturers routinely
modify the design of their flow sensors, the latest catalog information from the
manufacturer’s website or local sales representative should be used.

Total Pressure Drop

The total pressure drop (ΔTP), which is equal to the static pressure drop (ΔSP) plus the
velocity pressure drop (ΔVP), is the true indicator of the fan energy required to deliver the
design airflow through the box. Unfortunately, manufacturers typically only list the static
pressure drop which is always lower than the total pressure drop since the velocity at the box
inlet is much higher than the outlet velocity, resulting in static pressure regain. Therefore, in
order to size boxes when ΔTP is not cataloged, the designer needs to calculate the velocity
pressure drop using the following equation:

X76X VAV BOX SELECTION

 ΔTP = ΔSP + ΔVP
⎡⎛ v ⎞ 2 ⎛ v ⎞ 2 ⎤
= ΔSP + ⎢⎜ in ⎟ − ⎜ out ⎟ ⎥
⎣⎝ 4005 ⎠ ⎝ 4005 ⎠ ⎦

The velocity (FPM) at the box inlet and outlet are calculated by dividing the airflow rate
(CFM) by the inlet and outlet area (ft2), which in turn is determined from dimensions listed
in catalogs) 23.

Total Pressure Drop Selection Criteria

As noted above, smaller VAV boxes will have a higher total pressure drop, increasing fan
energy, and higher sound power levels. On the other hand, larger boxes cost more and are
more limited in how low the minimum airflow setpoint can be set, which can increase fan
energy and reheat energy under low load conditions.

Simulations were made to determine the optimum balance from an energy perspective
between pressure drop and minimum setpoint limitations. For most applications, the
analysis (described in Appendix 6) indicates that boxes should be selected for a total pressure
drop of about 0.5 in. H2O.

Table 14 shows the maximum airflows and sound data for a particular box manufacturer
based on a total pressure drop of 0.5 in.. The maximum airflow for each box in this table
was developed by iterating on the VAV box selection with the manufacturer’s selection
software: for each box, the maximum CFM was sought to obtain both a total pressure drop
of less than 0.5 in. and a radiated NC rating of less than 35. For each iteration, the
calculation of total pressure was done in a spreadsheet using the box inlet and outlet size to
determine the velocity pressures. Table 14 demonstrates that noise is not an issue for this
particular line of VAV boxes. The radiated NC values are quite low at 0.5 in. total pressure
drop. For other manufacturers this may not be the case. Refer to Appendix 7 for calculated
minimum and maximum airflows for Titus and Krueger VAV reheat boxes.

23
Inlet dimensions are typically quite easy to calculate as they are just circular cross sections at the scheduled neck
size. Outlets areas can be more difficult since they are typically rectangular flange connections that are much
larger than the inlet connections but not always clearly marked in catalogs. VAV box submittal data should be
consulted for outlet dimensions.

VAV BOX SELECTION X77X

 TABLE 14: Static Velocity Total
Inlet Outlet Outlet pressure pressure pressure
VAV BOX MAXIMUM Nominal diameter width height drop (in. drop (in. drop (in. Max Radiated
AIRFLOWS. size (in.) (in.) (in.) w.g.)* w.g.) w.g.) CFM NC*
4 4 12 8 0.08 0.42 0.50 230 21
5 5 12 8 0.15 0.35 0.50 333 20
6 6 12 8 0.24 0.25 0.49 425 21
7 7 12 10 0.25 0.25 0.50 580 20
8 8 12 10 0.33 0.17 0.50 675 22
9 9 14 13 0.27 0.23 0.50 930 17
10 10 14 13 0.32 0.18 0.50 1100 19
12 12 16 15 0.32 0.17 0.49 1560 19
14 14 20 18 0.31 0.19 0.50 2130 18
16 16 24 18 0.32 0.18 0.50 2730 22
*From selection software

One might think that the 0.5 in. pressure criterion need only apply to the box with the
greatest need for static pressure. This will determine the fan static pressure and hence the fan
power. Arguably then, VAV boxes closest to the fan hydraulically (where excess pressure
may be available) could be sized for a greater pressure drop than the most remote boxes.
However, as described in the following paragraphs, the 0.5 in. criteria should be applied to
all boxes regardless of location.

As loads shift throughout the day and year the most demanding box will change. Figure 26,
Figure 27 and Figure 28 are images of VAV box zone demand at different times of day for an
office building in Sacramento, California (Site 4). All three images are taken on the same
day, August 5, 2002. At 7am, Zone 14 on the southeast corner of the building has the most
demand. Later that morning at 9am, Zone 36 in the interior of the building experiences the
most demand. At 5pm, the high demand has shifted to Zone 30 in the northwest corner.
Throughout the period monitored (the better part of a year), the peak zone changed
throughout the floor plate, including both interior and perimeter zones. Hence the zone
requiring the most static pressure could vary throughout the day. If fan static pressure is
reset to meet the requirements of only the zone requiring the most pressure (see Demand-
Based Static Pressure Reset), and if boxes close to the fan are undersized to dissipate excess
pressure that is available at design conditions, then fan pressure and fan energy would
increase when these boxes become the most demanding during off-design conditions.

Therefore, since the most demanding box changes throughout time, all boxes on a job
should be sized using a consistent rule for maximum total pressure drop at design conditions.
This is also much simpler and more repeatable.

X78X VAV BOX SELECTION

 7 6 5
North FIGURE 26:

4 SITE 3 VAV BO X
30 28
DEMAND, 7AM MONDAY
23
29 27 19 AUGUST 5, 2002

31 26 25 22
8 24 21 20
3

18
32
9
38 2
33 34 37A 37B 39 16 17

35 36
15
10 11 Zone 14, 85% Open 1
12 13

Damper % Open
0
10
20
30
40
50
60
70
80
90
100

7 6 5
North FIGURE 27:

4 SITE 3 VAV BO X
30 28
DEMAND, 9AM MONDAY
23
19 AUGUST 5, 2002
29 27

31 26 25 22
8 24 21 20
3

18
32
9
2
33 34 37A 38 37B 16 17
39
Zone 36, 71% Open
35 15
10 11 14 1
12 13

VAV BOX SELECTION X79X

 FIGURE 28: 7 6 5
North

SITE 3 VAV BO X Zone 30, 4

74% Open 28
DEMAND, 5PM MONDAY
23
AUGUST 5, 2002 29 27 19

31 26 25 22
8 24 21 20
3

18
32
9
2
33 34 37A 38 37B 16 17
39

35 36
15
10 11 14 1
12 13

Table 15 summarizes the turndowns for typical selection of VAV boxes using the minimum
box airflow setpoint (“min CFM”) calculated in Table 13 and the maximum design airflow
rates (“max CFM”) calculated in Table 14. The column, “best turndown” is the ratio of the
min CFM to the max CFM as if the box size is selected just at the maximum allowable flow
rate. Worst turndown is the ratio of the min CFM for that box size to the max CFM of the
next smaller box size as if the box had the smallest airflow in it’s size range. These values for
best and worst represent the range of potential selections within a given box neck size. They
are computed both for all sizes of boxes and, to the right of the table, just for even neck sizes
of boxes. Many local VAV box suppliers only stock even sized boxes in their warehouses and
thus the lead-time to get odd size boxes (e.g., 5 in., 7 in., or 9 in.) to job site can be much
longer. Using only even sizes results in less turndown, but Appendix 6 shows that the
penalty for using only even sizes is fairly small. Similarly, 4 in. boxes are not a commonly
stocked items and it is common for a supplier to substitute a 6 in. box, possibly with a
“pancake reducer”. For more discussion of the pitfalls of odd size boxes and 4 in. boxes see
Appendix 13.

X80X VAV BOX SELECTION

 Odd and Even Sizes Even Sizes Only TABLE 15:
Nominal Max Min Best Worst Best Worst
SUMMARY OF SAMPLE
size CFM CFM turndown turndown turndown turndown
4 230 14 6% n/a 6% n/a BOX MAXIMUM AND
5 333 23 7% 10% MINIMUM AIR FLOW
6 425 33 8% 10% 8% 14%
7 580 44 8% 10%
8 675 58 9% 10% 9% 14%
9 930 73 8% 11%
10 1,100 91 8% 10% 8% 13%
12 1,560 130 8% 12% 8% 12%
14 2,130 177 8% 11% 8% 11%
16 2,730 232 8% 11% 8% 11%
Average 9% 10%
Note: These values were developed using a controller/sensor accuracy of 0.004 in.
w.c.

Other Box Types

Dual Duct Boxes

Dual duct VAV boxes are traditionally purchased with flow sensors in both the hot and cold
inlet. However, boxes with flow sensors in the cold inlet, the hot inlet, and/or the outlet are
also available. Three controls are recommended: snap-acting with a single sensor in the
outlet; mixing control with a single sensor in the outlet; and mixing control with a sensor on
the outlet and the cold inlet. All of these configurations are readily available with the sensors
mounted from the factory as a standard option.

Snap Acting Controls with a Single Sensor on the Outlet (or Sensors on Both Inlets)

Figure 29 and Figure 30 show the snap acting dual-duct VAV control scheme, followed by a
sample control sequence:

FIGURE 29:
DUAL DUCT - FROM
CO O L I N G T O H E AT I N G

VAV BOX SELECTION X81X

 FIGURE 30:
DUAL DUCT - FROM
H E AT I N G T O CO O L I N G

Temperature Control
1. When the zone is in the Cooling Mode, the Cooling Loop output shall reset
the discharge supply airflow setpoint from the minimum to cooling maximum
setpoints. The cooling damper shall be modulated by a PI loop to maintain
the measured discharge airflow at setpoint. The heating damper shall be
closed.
2. When the zone is in the Heating Mode, the Heating Loop output shall reset
the discharge supply airflow setpoint from the minimum to heating maximum
setpoints. The heating damper shall be modulated by a PI loop to maintain
the measured discharge airflow at setpoint. The cooling damper shall be
closed.
3. In the Deadband Mode, the discharge airflow setpoint shall be the zone
minimum, maintained by the damper that was operative just before entering
the Deadband. The other damper shall remain closed. In other words, when
going from Cooling Mode to Deadband Mode, the cooling damper shall
maintain the discharge airflow at the zone minimum setpoint and the heating
damper shall be closed. When going from Heating Mode to Deadband Mode,
the heating damper shall maintain the discharge airflow at the zone minimum
setpoint and the cooling damper shall be closed. This results in a snap-action
switch in the damper setpoint as indicated in the figures above.

Mixing Controls with a Single Sensor on the Outlet (or with a Sensor on the Outlet
and the Cold Inlet)

Figure 31 and Figure 32 show the mixing dual-duct VAV control scheme, followed by a
sample control sequence:

X82X VAV BOX SELECTION

 FIGURE 31:
DUAL DUCT MIXING -
F R O M CO O L I N G TO
H E AT I N G

FIGURE 32:
DUAL DUCT MIXING -
F R O M H E AT I N G TO
CO O L I N G

Temperature Control
1. If the system is in the Heating Mode, the Heating Loop output shall be
mapped to the heating damper position.
2. If the system is in the Cooling Mode, the Cooling Loop output shall be
mapped to the cooling damper position.
3. In the Deadband Mode, the cooling and heating dampers are controlled to
maintain minimum airflow, as described below.

Minimum Volume Control

1. In the Heating Mode, the cooling damper is modulated to maintain measured
discharge airflow at the minimum airflow setpoint.
2. In the Cooling Mode, the heating damper is modulated to maintain measured
discharge airflow at the minimum airflow setpoint.

VAV BOX SELECTION X83X

 3. In Deadband Mode, the last damper that was used to maintain minimum
airflow continues to do so (e.g., in transitioning from Heating into Deadband
Mode, the cooling damper would continue to maintain minimum airflow).

Maximum Volume Control

1. This control takes precedence over Temperature Control command of outputs
so that supply air volume does not exceed the maximum regardless of the
temperature control logic.
2. In the Heating Mode, if the discharge supply airflow rises above the maximum
heating airflow setpoint, the heat temperature control loop shall no longer be
allowed to open the damper. If the discharge supply airflow rises above the
maximum heating airflow setpoint by 10%, the heating damper shall be closed
until the airflow falls below setpoint.
3. In the Cooling Mode, if the discharge supply airflow rises above the maximum
cooling airflow setpoint, the cool temperature control loop shall no longer be
allowed to open the damper. If the discharge supply airflow rises above the
maximum cooling airflow setpoint by 10%, the cooling damper shall be closed
until the airflow falls below setpoint.

Comparison of Dual Duct Control Logic

Table 16 below discusses the advantages and disadvantages for each of these controls.

TABLE 16: Issue Snap-Acting with a Mixing Control with Mixing Control with a
COMPARISON OF DUAL- Single Sensor in the a Single Sensor in Sensor on the Outlet
Outlet the Outlet and the Cold Inlet
DUCT VAV CONTROLS.
Pressure Independent Yes No Yes
Control
First Cost Low Low High
Works with Demand No Yes Yes
Ventilation (CO2
Reset) Controls
Reheat Energy None Yes Yes
Thermal Comfort Ok Better Better

As shown in Table 16, advantages and disadvantages exist for each scheme. The snap acting
control has both low cost and low reheat energy, but it experiences wider zone temperature
fluctuations and will not work with demand ventilation controls and other applications
where the minimum airflow setpoint is a large fraction of the design maximum setpoint.
The mixing controls have better thermal comfort and will work with demand ventilation
controls, but the designer has to either save money and sacrifice pressure independent
control or buy another sensor (and analog input point) to get the highest thermal
performance. In general, the recommended approach is the single outlet sensor with snap-
acting controls for zones without DCV and mixing control with a single discharge sensor for
zones with DCV.

X84X VAV BOX SELECTION

The reason that snap-acting controls cause higher temperature fluctuations is that they
change rapidly between minimum flow with hot air and with cold air, which also prevents
them from working with demand ventilation controls. The temperature fluctuations are
relatively imperceptible at minimum design airflow. Demand ventilation controls increase
this minimum as more people enter the space. At an extreme, this situation could cause the
box to fluctuate between full cooling and full heating with no dead band in between.

The loss of pressure independence with the single sensor mixing scheme is not significant
when coupled with a demand limit on cfm (see Demand-Based Static Pressure Reset).
Compared with the premium of $500 to $1,000 per zone for an extra sensor and analog
input, it usually makes sense to use this configuration unless cost is not a concern for the
client. 24

Sizing Dual Duct Boxes

Dual duct boxes should be sized in the same manner as the single duct: the maximum CFM
per box is based on a uniform rule for total pressure drop (e.g. <= 0.5 in. w.c.), provided
noise levels are acceptable. As with reheat boxes, the minimum controllable airflow setpoint
is a function of the amplification factor of the velocity sensor, the minimum velocity pressure
setpoint capability of the controller, and the duct area at the sensor location. It is important
to use the area of the outlet in this calculation if the sensor is in the outlet. Outlet sizes are
typically larger than inlet sizes but this varies by manufacturer).

The pressure drop across dual duct boxes differs widely depending on the style of box and
the placement of the velocity pressure sensors. Boxes that have mixing baffles to ensure
complete mixing of the hot and cold airstreams have the highest pressure drops. Complete
mixing is only a factor when mixing control logic is used (it is not an issue with snap-acting
since the hot and cold dampers are never open at the same time) and it is only an issue when
the VAV box is serving multiple rooms where inconsistent supply air temperature can upset
balance. When discharge velocity pressure sensors are used, the discharge outlet is often
reduced from the size used when dual inlet velocity pressure sensors are used. This is
intended to increase velocity and improve airflow measurement, but it also results in better
mixing of the two airstreams and it increases pressure drop. The pressure drop for this
design varies widely among manufacturers; the bid list should be limited to the best one or
two or require that boxes be increased in size to match the pressure drop performance of the
specified manufacturer. With a discharge airflow sensor, we have found mixing to be

24
One VAV box manufacturer cautions that, “…discharge flow sensors may be highly inaccurate, even with
multiple point center averaging "flowcrosses". One should ensure that a manufacturer has tested the accuracy of
the discharge sensor over a range of inlet conditions. ASHRAE Standard 130 will define a method of test for
determining temperature mixing for dual duct boxes, with a ratio of temperature variation at the discharge vs the
difference between inlet temperatures. A ratio of 1:20 is possible, and usually results in good sensor performance
as well. We find that poor temperature mixing (1:15 or less) often results in poor sensor response as well.”

VAV BOX SELECTION X85X

 sufficient from a comfort perspective for most applications. Mixing baffles, which add
significantly to both first costs and pressure drop, should only be used for the most
demanding applications (e.g. hospitals).

In calculating the velocity pressure loss from a dual duct VAV box, note that although the
outlet sensor is typically in a round duct, the connecting duct is typically a larger rectangular
duct connected to a flange on the discharge plate. The manufacturers use this larger
rectangular duct size in rating the duct static pressure loss so its area should be used to
determine outlet velocity pressure.

Series Fan Powered Boxes

Series fan powered boxes should be avoided, with the exception of a few specific applications,
because the small fans and motors in fan-powered boxes are highly inefficient (as low as 15%
combined efficiency compared to central fans with 60% or greater combined efficiencies).

Series fan powered boxes are recommended for the following zones within a VAV-reheat
system:
ƒ Series boxes are one of the recommended options for interior conference rooms; see
Design of Conference Rooms for an explanation and discussion of other options.
ƒ Series boxes should be used for any space that requires a high minimum flow rate in
order to maintain good mixing, to prevent dumping, or to meet the heating load at a
reasonable supply air temperature (e.g. <90°F). For example, a large two-story lobby or
atrium might have a sidewall diffuser at the height of the first story. Without a ceiling
above the diffuser to provide the Coanda effect, the diffuser might “dump” at low flows
and not be able to “throw” across the entire space. A series box maintains constant
velocity under all load conditions.

Controls on systems with series-style boxes should stage the boxes on before the central fans
are activated in order to prevent the box fans from running backwards. Single phase motors
will run backwards at reduced airflow rates if they are spinning in reverse when they are
started.

X86X VAV BOX SELECTION

ECM Motors

Series fan powered boxes are available with high efficiency electrically commutated motors
(ECM). While these cost more than conventional fixed speed motors, they generally pay for
themselves in energy savings. ECM motor efficiency is generally in the 65-72% range.
Standard PSC (permanent split capacitor) can just as efficient as ECM motors but due to
acoustical considerations the PSC motor is usually adjusted with an electronic SCR speed
controller to operate at a lower speed. Installed PSC motor efficiencies are typically in the
12-45% range. Acoustics are less of an issue for smaller series boxes and thus standard PSC
motors can be just as efficient as ECM motors for the smaller sizes but for larger units ECM
motors are more efficient. ECM motors also have the advantage of soft start and speed
ramps which are typically programmed into the motor.

In the 2005 version of the Title 24 Standard, ECM motors are required for all series style
boxes with motors under 1 HP.

Parallel Fan Powered Boxes

Parallel fan powered boxes can reduce or eliminate reheat, but the first cost and maintenance
cost are higher than reheat boxes. The cycling of parallel box fans also may be an acoustical
nuisance.

The efficiencies of the parallel fan and motor are not a significant issue as they are with series
boxes because the fan generally operates only in the heating mode. Since all the fan energy is
supplied to the space, it is simply a form of electric resistance heat and not “lost” or reheated.

If the dual maximum control strategy is used along with maximum and minimum airflow
setpoints determined as described above, VAV reheat boxes are almost always a better option
than parallel fan powered boxes on a life-cycle cost basis. The exception may be if fan-
powered boxes can be operated with zero minimum airflow setpoints (see Zero Minimums),
thus completely eliminating reheat losses and significantly reducing fan energy.

Unlike series-style boxes, parallel-style boxes do not need special controls to prevent them
from running backwards. They are provided with integral backdraft dampers that prevent
system air from escaping out of the plenum when the box fan is off.

Parallel fan powered boxes are often a good option when using electricity, rather than hot
water as the heating source (see section below on Electric Reheat).

VAV BOX SELECTION X87X

 Other Issues

Zero Minimums

Some will argue that VAV boxes can have zero minimum airflow setpoints because if there is
a need for ventilation, i.e., the space is occupied, there will also be a cooling load in the
space, so the thermostat will call for cooling and the VAV box will provide the necessary
ventilation. This might be true for interior zones but is not necessarily true for exterior
zones, particularly in the winter. Furthermore, this does not strictly meet Title 24 which
requires that the minimum ventilation rate be provide whenever the space is “expected” to be
occupied, including times during the day when the space may not be occupied and at low
load (see “Code Ventilation Requirements”). Nevertheless, zero minimum airflow setpoints
are acceptable under some circumstances:
ƒ Multiple zones serving open office plan. The code allows some VAV boxes serving a
space to go to zero airflow, provided other boxes serving that space are controlled to
provide sufficient minimum ventilation for the entire space. For example, a large open
office plan might be served by two boxes, one in the interior and one along the
perimeter. Suppose the perimeter were 1000 ft2 and designed for 2000 cfm (2 CFM/ft2)
while the interior was 1000 ft2 and designed for 500 cfm (0.5 CFM/ft2). The minimum
airflow rate required for ventilation is 0.15 cfm/ft2 or 300 cfm. Code could be met
using an cooling-only box in the interior with a zero minimum airflow setpoint, and a
reheat box serving the perimeter with a 15% (300 cfm) minimum setpoint. If the
interior box is controlled to maintain its ventilation rate alone (equal to 30% of its
maximum), then a reheat coil would need to be added to this box to prevent
overcooling the space at minimum flow. Therefore, combining interior cooling-only
boxes with perimeter reheat boxes in open office plans saves first cost and energy. (This
concept does not apply when the interior and perimeter are separated from each other
with full height partitions.)
ƒ Multiple zones serving a large zone. Another application where zero minimum airflow
setpoints re allowed is for large zones (e.g., large meeting rooms) where more than one
box may be needed to meet the load. In this case, one or more of the boxes could have a
zero minimum, as long as at least one box has a non-zero minimum that can meet the
minimum ventilation requirements for the entire zone.

X88X VAV BOX SELECTION

ƒ Fan-powered boxes. Zero minimum volume setpoints are an option on series and
parallel style fan-powered boxes since the box fan can be used to supply the minimum
ventilation rate using plenum air. Title 24 specifically allows transfer air to be used to
meet ventilation requirements provided the system as a whole is provided with the sum
of the outdoor air rate required for all spaces served by the system. This design will only
work, however, if there are always some zones served by the system that are supplying
sufficient air that the minimum outdoor air for the system can be maintained at the air
handler. For example consider a system serving a combination of interior and perimeter
zones with fan-powered boxes with zero minimum airflow setpoints at the perimeter. In
cold weather, all the perimeter boxes will be in the heating mode and shut off. The load
in interior spaces must always be equal to or greater than the minimum ventilation rate
to provide enough airflow for the entire system ventilation requirements. If this is not
the case, non-zero minimums must be used at the perimeter. (A heating coil may also
be needed at the air handler to prevent supply air temperature from falling too low since
the minimum outdoor air may be nearly 100% of the supply air under this cold weather
design condition.)

Cooling-Only Boxes

In times past when interior lighting and PC loads were substantially higher than they are
now, interior spaces did not need heat and therefore could be served by cooling-only VAV
boxes. The loads were sufficient to allow boxes to be set to minimum rates required for
ventilation without overcooling. But with the very low lighting and plug load power
densities now common, overcooling is very possible, even likely. Except where zero
minimums may be used (see discussion above in “Zero Minimums”), reheat is probably
required to ensure both comfortable temperatures and adequate ventilation for interior areas.
Reheat is also required for interior zones with floor heat loss, such as from slabs on grade or
over an unconditioned basement/garage.

Electric Reheat

Title 24 has a prescriptive requirement that significantly limits the use of electric resistance
heat. There are a few exceptions and electric heat can be used if compliance is shown using
the Performance Approach where additional source energy from the electric heat can be
offset by other energy conservation measures. The prohibition on electric heat has been part
of Title 24 since the 1970’s but there have been a number of significant changes since then
that change the economics of hot water reheat versus electric reheat, including:
1. Building envelope codes have improved dramatically, particularly for
windows, so heating loads have decreased significantly. Gas heat results in
lower heating bills but the savings are smaller now and make it harder to
payback the higher first cost of the hot water reheat system.
2. With time-of-use electric rates and seasonally adjusted gas rates, electricity
is relatively inexpensive when heating is required and gas is relatively
expensive.
3. Hydronic heating systems are relatively more expensive as material costs
(e.g. copper pipe) and pipe fitter labor costs increase.

VAV BOX SELECTION X89X

 4. DDC zone controls now allow dual maximum control sequences (as
opposed to single minimum pneumatic sequences) which reduces reheat.
5. DDC controls now permit cold deck supply air temperature reset which
further reduces reheat.
6. Some electric heaters now have fully modulating control (as opposed to
one or two steps of heating) with proof of flow switches that can detect
very low airflows (http://www.thermolec.com/). Thus minimum heating
flow rates can be just as low for electric reheat as for hot water reheat. In
the past, it was usually necessary to have a higher minimum for electric
reheat.
7. Some proportional electric heat suppliers can use the three point floating
outputs of the DDC controller, often saving the ~$100 premium for a
controller with an analog output.

For these reasons many designers and owners are electing to use electric heat (and use the
Performance Method for Title 24 compliance). On two recent underfloor design projects
lifecycle cost analyses showed that the $0.02-$0.05 per square foot per year energy cost
savings could not reasonably pay for the $0.50 to 1.00/ft2 cost premium for the hot water
system. On a recent overhead VAV system the savings were calculated as $0.20/ft2-yr but the
first cost premium (including utility rebate) was about $2.50/ft2 for a 12 year payback.

A note of caution when performing lifecycle cost analysis on hot water versus electric reheat:
simulation models are generally a “best case scenario”, i.e. it assumes that all systems are
operated per the sequences. However, if heating loads turn out to be higher than predicted
then the penalty for electric heat could be more significant. For example, zone minimums
may end up being set higher than intended because the proof of flow switches do not work
as expected. Other examples include higher than expected infiltration rates, poorly
calibrated minimum ventilation controls, zone minimum flow rates reset by building
engineers, etc. Therefore, it is important to evaluate some reasonable worst case scenarios
when considering electric reheat.

Another consideration with electric reheat is that it often makes sense to switch from single
duct reheat to parallel fan powered boxes when switching from hot water to electric heat.
Parallel fan powered boxes can eliminate reheat because Title 24 allows the primary air
damper to be closed in heating mode. Thus a parallel fan powered box will almost always be
more efficient than reheat boxes (note that the box fan typically only runs in heating and the
fan heat contributes to space heating so there is really no energy penalty for the box fan).
The problem with parallel fan boxes (aside from potential noise and maintenance issues) is
that they cost more than reheat boxes, but most of this cost is electrical contractor's cost of
running line voltage power (e.g. 277V) to the box. Hot water reheat boxes only require 24V
which is typically run by the controls contractor. Reheat boxes with electric heat, however,
require line voltage so the cost premium for going from electric reheat to parallel fan box
with electric heat is fairly small. Furthermore, since a parallel fan box does not reheat and

X90X VAV BOX SELECTION

since fan speed is usually fixed, there is little advantage to the more expensive modulating
electric heater controls, i.e. step control is sufficient for parallel fan electric heaters. Where
electric resistance heat is used, the National Electric Code (NEC) requires both airflow
switches and thermal switches on electric coils. The airflow switches provided with electric
coils are often low quality and require a relatively high airflow to prove flow. As a result, the
effective minimum airflow for electric coils is higher than that for hot-water coils. As a
general rule, a minimum VP sensor reading of 0.03 in. is recommended for electric reheat
with step control, as opposed to fully modulating control. Table 17 shows typical turndown
ratios for electric reheat.

All electric coils are required to have automatic reset thermal switches. On large coils a
second manual reset thermal switch is required. Where electric heat is used, the controls
should ensure that the fans run for several minutes before and after the heating coil has been
engaged to prevent tripping of the thermal switches. It only takes a few false trips to
convince a building operator to run the system continuously to prevent having to reset
thermal switches above the ceiling.

Odd and Even Sizes Even Sizes Only TABLE 17:

Nominal Max Min Best Worst Best Worst
VAV BOX TURNDOWN
size CFM CFM turndown turndown turndown turndown
17% n/a 17% n/a WITH ELECTRIC REHEAT.
5 333 62 19% 27%
6 425 89 21% 27% 21% 39%
7 580 122 21% 29%
8 675 159 24% 27% 24% 37%
9 930 201 22% 30%
10 1,100 248 23% 27% 23% 37%
12 1,560 357 23% 32% 23% 32%
14 2,130 486 23% 31% 23% 31%
16 2,730 635 23% 30% 23% 30%
Average 25% 28%
Note: These values were developed using a controller/sensor accuracy of 0.004 in. w.c.

DDC at the Zone Level

Pneumatic controls are extremely simple to maintain and inexpensive to install. Pneumatic
actuators are fast acting – a characteristic that keeps them in the market for lab exhaust
controls. However, in general, pneumatic controls are less precise than DDC controls and
do not easily provide the zone feedback that can make VAV systems truly efficient.

A number of the control sequences in this document rely on zone feedback, including the
supply pressure setpoint reset for air-handling units or central fans, and supply temperature
setpoint reset for central coils. These sequences can provide significant energy savings, but
savings are rarely large enough to justify the ~$700/zone cost premium of DDC over
pneumatic controls. But DDC offers other benefits that make the cost premium worthwhile

VAV BOX SELECTION X91X

 to most owners and builders; currently DDC is provided for the majority of zone controls
and nearly 100% of the market for new buildings.

Benefits of DDC at the zone level other than energy savings include:
ƒ Zone control problems can be remotely detected, alarmed, and diagnosed by building
engineers or service technicians
ƒ Elimination of compressed air system and associated maintenance
ƒ More precise zone temperature control
ƒ Ability to restrict thermostat setpoints in software to prevent occupant abuse
ƒ Reduced calibration frequency
ƒ Ability to intertie occupancy sensors, window switches, and CO2 sensors
ƒ Ability to allow occupants to view and/or adjust their controls from their computer
(requires a web-based DDC system)

Given the cumulative effect of energy savings and other benefits, we recommend DDC zone
controls for new systems. In existing buildings, we recommend upgrading central systems to
DDC and replacing the zones with DDC controllers only during future tenant build-outs
and remodels.

X92X VAV BOX SELECTION

 DUCT DESIGN

General Guidelines

Duct design is an art as much as it is a science. To design duct systems well requires
knowledge of both the principles of fluid flow and the cost of ducts and duct fittings. The
ideal system has the lowest life-cycle cost (LCC), perfectly balancing first costs (cost of the
duct system and appurtenances such as dampers, VAV boxes, etc.) with operating costs
(primarily fan energy costs). To rigorously optimize LCC is impractical even with a very
large engineering budget; there are simply too many variables and too many unknowns. For
instance, first costs are not simply proportional to duct size or weight. Fittings cost more
than straight duct and round ducts generally cost less than rectangular ducts. Some fittings
that serve the same purpose are more expensive than others, depending on duct size and the
capabilities of the sheet metal shop. It is therefore difficult for a designer to optimize the
design of the duct system absent knowledge of who will be building the system. Estimating
operating costs is also inexact to a large part because duct system pressure drops cannot be
accurately calculated (see additional discussion below).

Still, some rules of thumb and general guidelines can be developed to help designers develop
a good design that provides a reasonable, if not optimum, balance between first costs and
operating costs, including the following:
1. Go straight! This is the most important rule of all. The straighter the
duct system, the lower both energy and first costs will be. From an energy
perspective, air “wants” to go straight and will lose energy if you make it
bend. From a cost perspective, straight duct costs less than fittings.
Fittings are expensive because they must be hand assembled even if the
pieces are automatically cut by plasma cutters. So, when laying out a
system, try to reduce the number of bends and turns to an absolute
minimum.

DUCT DESIGN X93X

 2. Use standard length straight ducts and minimize both the number of
transitions and of joints since the sheet metal is not as expensive as the
labor to connect pieces together and seal the joints. Straight, standard
length ducts are relatively inexpensive since duct machines, such as coil
lines for rectangular ducts, automatically produce duct sections. Standard
sheet metal coils are typically 5 feet wide, so standard rectangular duct
lengths are 5 feet long (somewhat less for machines that bend flanges and
joints out of the coil metal). Any rectangular duct that is not a standard
length is technically a fitting since it cannot be made by the coil line.
While spiral round duct can be virtually any length, it is commonly cut to
20 feet to fit in standard trucks. Oval duct standard lengths vary
depending on the fabricator but manufactured ducts are typically 12 feet
long. It is not uncommon for an inexperienced design to include too
many duct size reductions with false impression that reducing duct sizes
will reduce costs. In Figure 33, four transitions are reduced to one with
each remaining duct section sized for multiples of the standard 5-foot
rectangular duct length, which reduces both first costs and energy costs.
3. Use round spiral duct wherever it can fit within space constraints. Round
duct is less expensive than oval and rectangular duct, especially when run
in long, straight sections. Round duct fittings are relatively expensive, so
this rule would not apply where there are many transitions, elbows, and
other fittings close together. Round spiral duct also leaks less than
rectangular duct due the lack of longitudinal joints and generally fewer
transverse joints when run in long straight duct sections. Round duct also
allows less low frequency noise to break-out since it is round and stiff.
The flat sections of rectangular duct and wide flat oval duct behave like a
drum, easily transmitting low frequency duct rumble. Flat oval duct is
often the next best option when space does not allow use of round duct.
The cost, however, will vary by contractor since some have the machines
to fabricate oval duct while others must purchase factory-made duct and
fittings. Rectangular duct should usually be limited to ducts that must be
acoustically lined (lining rectangular duct is least expensive since it can be
done automatically on coil lines), for duct sections containing many
fittings (rectangular duct fittings are usually easier to assemble than round
and oval fittings), and for large plenums.
4. Use radius elbows rather than square elbows with turning vanes whenever
space allows. Figure 34 and Figure 35 show the performance of elbows
and tees in various configurations. Except for very large ducts (those
whose shape cannot be cut on a 5-foot wide plasma cutter), full radius
elbows will cost less than square elbows with turning vanes, yet they have
similar pressure drop and much improved acoustic properties. Turning
vanes generate some turbulence, which can be noisy at high velocities. On

X94X DUCT DESIGN

medium and high velocity VAV systems, where a full radius elbow cannot
fit, a part-radius elbow with one or more splitters should be used. The
splitters essentially convert the duct into nested full-radius elbows. This
design will have the lowest pressure drop and produce the least noise.
Turning vanes should only be used on low velocity systems where radius
elbows will not fit. Turning vanes should be single width, not airfoil
shaped. Intuitively, airfoil vanes would seem to offer better performance
but SMACNA and ASHRAE test data show that they have higher
pressure drop as well as higher cost.

FIGURE 33:
EXAMPLES OF POOR
AND BETTER DUCT
DESIGN

DUCT DESIGN X95X

 FIGURE 34:
PRESSURE DROP
THROUGH ELBOWS Δ

X96X DUCT DESIGN

 FIGURE 35:
PRESSURE DROP

Δ THROUGH RECTANGULAR
TEES

5. Use either conical or 45° taps at VAV box connections to medium pressure
duct mains. Taps in low velocity mains to air outlets will have a low-pressure
drop no matter how they are designed. Use of conical taps in these situations
is not justified because the energy savings are small. Inexpensive straight 90°
taps (e.g., spin-ins) can be used for round ducts and 45° saddle taps are
appropriate for rectangular ducts. Taps with extractors or splitter dampers
should never be used. They are expensive; they generate noise; and most
importantly, they cause an increase in the pressure drop of the duct main.
Since fan energy is determined by the pressure drop of the longest run,
increasing the pressure drop of the main can increase fan energy. These
devices reduce the pressure drop in the branch only, which is not typically the
index path that determines fan energy. Also, the pressure drop through the
branch will be about the same as with conical or 45° saddle taps, both of which
are less expensive. So there are no redeeming qualities that would ever justify
the use of extractors or splitter dampers.

6. VAV box inlets should be all sheet metal; do not use flex duct. This will
reduce pressure drop because the friction rate of VAV inlet ducts is very high

DUCT DESIGN X97X

 when sized at the box inlet size. It also will ensure smooth inlets to the VAV
box velocity pressure sensor, improving airflow measurement accuracy, and
reduce breakout noise from the VAV damper (flex duct is virtually transparent
to noise) 25.

FIGURE 36:
PRESSURE DROP
T HRO UG H DUC T TAPS

Δ Δ

Δ Δ

STRAIGH T TAP
SPLITTER DAMPER
EXTRAC TOR

Δ Δ

7. Avoid consecutive fittings because they can dramatically increase pressure drop.
For instance, two consecutive elbows can have a 50% higher pressure drop than
two elbows separated by a long straight section. A tap near the throat of an
elbow can even result in air being induced backwards into the fitting −
essentially an infinite pressure drop.

8. Use 5 to 20 feet of flex duct between the box and the diffuser. Using flex
duct is generally less expensive than hard duct, particularly at the diffuser
connection and it allows duct-borne noise (e.g. box damper noise) to break out
above the ceiling and thus reduces the noise to the space. Flex duct use should
be limited because it has a higher pressure drop than hard duct and is more

25
Refer to ARI 885 Acoustical Application Standard for guidance on VAV box sound calculations.

X98X DUCT DESIGN

 susceptible to being squashed, kinked and otherwise poorly installed and
maintained.

Pressure loss data for duct fittings are available from ASHRAE and SMACNA publications
(see the SMACNA HVAC Systems Duct Design Manual, the Duct Design section of the
ASHRAE Handbook of Fundamentals or ASHRAE’s Duct Fitting Database).

Supply Duct Sizing

Ideally, duct-sizing techniques such as the T-method or the static regain method should be
used (ASHRAE Handbook of Fundamentals, 2001, Chapter 34), but they seldom are in
actual practice for two primary reasons. First, they are complex and require computer tools
to implement, which increases design time and costs. Second, and perhaps most important,
the methods are over-simplified because they do not account for duct system effects. System
effects include the added pressure drop resulting from consecutive fittings that cannot
accurately be estimated by either hand or computer calculations since each fitting
combination is unique. Fan system effects result from fans with fittings at their inlets or
discharges that result in large pressure drops or uncataloged reductions in performance.
System effects, both at the fan and in the duct system, can account for 50% or more of the
total system pressure drop. Therefore, using a complex computerized duct sizing method
may not be justified given that the accuracy may be not much better than simpler hand
methods.

Low pressure ducts (ducts downstream of terminal boxes, toilet exhaust ducts, etc.) are
typically sized using the equal friction method (ASHRAE Handbook of Fundamentals,
2001, Chapter 34) with friction rates in the range of 0.08 in. to 0.12 in. per 100 feet. This
design condition should be considered an overall average rather than a hard limit in each
duct section. For instance, rather than changing duct sizes to maintain a constant friction
rate in each duct section as air is dropped off to outlets, it can be less expensive, but result in
similar performance, if the duct near the fan has a somewhat higher rate (e.g., 0.15 in. per
100 feet) and the duct size remains the same for long lengths as air is dropped off. The lower
friction rate in the end sections offset the higher rate near the fan, but overall the system
costs less because reducers are avoided.

For medium pressure VAV supply ducts, a relatively simple duct sizing technique called the
friction rate reduction method is recommended. The procedure is as follows:

1. Starting at the fan discharge, choose the larger duct size for both of the following
design limits:

a. Maximum velocity (to limit noise). Velocity limits are commonly used
as a surrogate for limiting duct breakout noise. Many argue it is a poor

DUCT DESIGN X99X

 indicator since noise is more likely to result from turbulence than velocity;
e.g., a high velocity system with smooth fittings may make less noise than
a low velocity system with abrupt fittings. Nevertheless, limiting velocity
to limit noise is a common practice. It is important to consult with the
project’s acoustical engineer on this issue. Many rules-of-thumb for
velocity limits exist depending on the noise criteria of the spaces served
and the location of the duct. The typical guidelines for office buildings
are:

i. 3500 fpm in mechanical rooms or shafts (non-noise sensitive).

ii. 2000 fpm for ducts in ceiling plenums.

iii. 1500 fpm for exposed ducts.

b. Maximum friction rate (to limit fan power). A reasonable starting

friction rate for VAV systems is 0.25 in. to 0.30 in. per 100 feet. The
rationale for this range appears below.
2. At the end of the duct system, choose a minimum friction rate, which is typically
0.10 in. to 0.15 in. per 100 feet.

3. Decide how many transitions will occur along the hydraulically longest duct main
(the so-called “index run,” the run with the highest pressure drop that will
determine the design pressure drop and fan power) from the fan to the most remote
VAV box. Typically, a transition should not be made any more frequently than
every 20 feet since the cost of the transition will generally offset the cost of the sheet
metal savings. The design is more flexible to accommodate future changes and is
more energy efficient with fewer transitions. It is not uncommon to have only
three or four major transitions along the index run.

4. Take the difference between the maximum friction rate as determined in step 1
(whether determined by the friction limit or velocity limit) and the minimum
friction rate from step 2 (e.g., 0.3 in. less 0.1 in. = 0.2 in.) and divide it by the
number of transitions. The result is called the friction rate reduction factor. Size
duct along the index run starting with the maximum friction rate, then reduce the
friction rate at each transition by the friction rate reduction factor. By design, the
last section will be sized for the minimum friction rate selected in step 2.

The method is illustrated in Figure 37 that shows a riser diagram of a simple three-story
building:

X100X DUCT DESIGN

 FIGURE 37:
EXAMPLE OF DUCT
SIZING USING THE
F R I C T I O N R AT E
REDUCTION METHOD

In this example, we start with a maximum friction rate of 0.3 and end with a minimum rate
of 0.15 at the beginning of the last section. The index run connects to the first floor. Three
transitions exist so the friction rate reduction factor is (0.3 – 0.15)/3 = 0.05 in. Each section
of the run is sized for ever-decreasing friction rates. The other floors should be sized for the
same friction rate as the duct on the index floor – 0.2 in. per 100 feet in this example –
primarily for simplicity (typical floors will have the same size ducts).

This technique emulates the static regain method, resulting in somewhat constant static
pressure from one end of the duct section to the other, but without complex calculations. It
is not intended to be precise, but precision is not possible in most cases due to system effects
and the normal changes that occur as design progresses. It is also important to realize that
precise duct sizing is not necessary for proper operation because VAV boxes can adjust for a
wide range of inlet pressures, generally more than what occurs in medium pressure systems
designed using the friction rate reduction method.

DUCT DESIGN X101X

 Design Friction Rates for VAV Systems

Some may consider the 0.3 in. per 100 feet initial friction rate to be very high for an energy
conserving design. But this design condition represents a reasonable balance between first
costs (including cost of sheet metal ducts plus the space required to house them) and
energy costs, recognizing that VAV systems seldom operate at their design capacity.

The appropriateness of this friction rate as a design condition can be demonstrated by an

analysis of the life-cycle costs of a simple duct distribution system. Assume that the life-
cycle cost (LCC) of the duct system is the sum of first costs (FC) and life-cycle energy costs
(EC, equal to annual energy costs adjusted by a life cycle present worth factor), as shown in
the equation below:

LCC = FC + EC

First costs are roughly proportional to duct surface area (area of sheet metal). For round
ducts, costs would then be proportional to duct diameter D:

FC ∝ D

Assuming that energy costs for a given fan system are proportional to duct friction rate, the
friction rate in a standard duct system can be calculated from the following equation that is
used in friction rate nomographs like the Trane Ductilator:

f ∝ D −1.2V 1.9

where D represents the duct hydraulic diameter and V is the velocity.

For a round duct, the velocity for a given airflow rate is inversely proportional to the square
of the diameter, so the friction rate varies with diameter:

f ∝ D −5

Based on the equations above, the life cycle cost as a function of diameter would be:

LCC = FC + EC
= K1D + K 2 D − 5

and as a function of friction, the LCC would be calculated as:

LCC = C1 f −0.2 + C2 f

X102X DUCT DESIGN

where K and C are constants for a given system.

LCC is minimized for a given friction rate when the derivative of the LCC with respect to
friction rate is zero:

∂LCC
= 0 = −0.2C1 f −1.2 + C2
∂f

Now assume that a constant volume system has a minimum LCC when the friction rate is
0.1 in. per 100 feet. This is probably the most common design friction rate used for
constant volume and low velocity duct systems:

0 = −0.2C1 (0.1) −1.2 + C2

3.2C1 = C2

The LCC equation can be simplified to:

LCC = C1 f −0.2 + 3.2C1 f

If assuming that the system is variable volume, at an average annual airflow rate of 60%, a
VAV system with a variable speed drive will use about 30% of the energy used by a
constant volume system of the same design size. The LCC equation then becomes:

LCC = C1 f −0.2 + 0.3 * 3.2C1 f

= C1 f − 0.2 + 0.96C1 f

Taking the derivative with respect to friction rate and setting to zero, it is possible to solve
for the friction factor that results in the lowest LCC:

∂LCC
= 0 = −0.2C1 f −1.2 + 0.96C1
∂f
f = (0.21) − 0.83
= 0.27

While this analysis is fairly simplistic, it does demonstrate that sizing ducts for a higher
friction rate for VAV systems than for constant volume systems is technically justified based
on life-cycle cost. If 0.1 in. per 100 feet is the “right” friction rate for constant volume
systems, then 0.25 in. to 0.3 in. per 100 feet is “right” for VAV systems. Note that with
the friction rate reduction method, this rate is only used for the first section of duct, so
average friction rates will be less, but still greater than that for constant volume systems.

DUCT DESIGN X103X

 Return Air System Sizing

The return airflow rate is equal to the supply rate minus building exhaust and an amount
that will mildly pressurize the building to reduce infiltration. The amount of air required for
mild pressurization (between 0.03 in. to 0.08 in. above ambient) will vary with building
construction tightness. Rules of thumb for typical commercial systems are between 0.1 in.
and 0.15 cfm/ft2. The 0.15 cfm/ft2 rate matches the minimum outdoor air quantities for
ventilation required by Title 24 for most commercial buildings. If this air were returned
through the shaft, it would have to be exhausted anyway. By reducing the return airflow rate
by this amount, return air path space requirements and return/relief fan energy usage are
reduced.

Techniques for sizing ducted returns depend on the economizer relief system. For instance,
if relief fans are used, the pressure drop should be kept low so ducts are sized using low
friction rates much like constant volume systems. For systems with return fans, return air
ducts are typically sized using the same technique used to size supply air ducts.

Unducted return airshafts, as shown in Figure 8, are typically sized for low pressure drop,
using either a fixed friction rate, velocity, or both.

To size the shaft on friction rate basis, the hydraulic (or equivalent) diameter of the shaft
must be calculated using the formula:

4 Afree
HD =
Pwetted

where Afree is the free area and Pwetted is the “wetted” perimeter. The “wetted” perimeter is the
length of the duct surface that is touching the air stream.

Looking at the example in Figure 8, Afree is the plan area of the shaft minus the area of all
ducts in the shaft (including the take-off to the floor!). Pwetted is the length of the inside
perimeter of the shaft wall plus the outside perimeter of the ducts in the shaft. The friction
rate is then calculated using the hydraulic diameter and the standard SMACNA/ASHRAE
equations for losses (see either the SMACNA HVAC Systems Duct Design Manual or the
Duct Design section of the ASHRAE Handbook of Fundamentals).

Typically, shaft area is simply sized using velocity rather than friction rate. Maximum
velocities are generally in the 800 fpm to 1200 fpm range through the free area at the top of
the shaft (highest airflow rate).

X104X DUCT DESIGN

 Fan Outlet Conditions

Fan performance is rated using a test assembly with long straight sections of ductwork at the
fan discharge. However, in practice, these long duct runs are seldom possible. Fans typically
discharge very close to an elbow or other fitting. The result is that the fan will not operate as
cataloged, behaving instead as if it were operating against an additional pressure drop. To
achieve a given airflow, fan speed and energy use will be higher than what is indicated on
performance curves. The extent of this “fan system effect” depends on how close the fitting
is to the fan discharge and the orientation of the fitting with respect to the rotation of the
fan. SMACNA has catalogued the effect for various fan discharge arrangements (SMACNA
HVAC Systems Duct Design Manual), but the magnitude of the effect in real systems is
largely unknown.

To avoid system effect, fans should discharge into duct sections that remain straight for as
long as possible, up to 10 duct diameters from the fan discharge to allow flow to fully
develop. Where this is not possible, the effect can be minimized by:
ƒ Orienting the fan so that an elbow close to the discharge bends in the direction of the
fan rotation. Figure 38 shows how the opposite arrangement results in significant
system effect.
ƒ Discharging the fan into a large plenum then tap duct mains into the plenum with
conical taps in situations like Figure 38 where poor discharge arrangement is
unavoidable. Although this discharge will waste the fan’s velocity pressure, it will
typically have a net lower energy impact than a poor discharge, and the plenum will
reduce fan noise.

Figure 39 shows measured data for a system that suffers from both fan and duct system
effect. The fan discharges directly into a sound trap, which was cataloged at 0.25 in. pressure
drop at the rated airflow but actually creates a 1.2 in. pressure drop. The pressure drop
resulting from the velocity profile off the fan is not symmetrical and most of the airflow goes
through only one section of the sound trap. The air then goes directly into an elbow with a
tap just below the throat of the elbow. Because the streamlines at the exit of the elbow are all
bunched to the right side, the pressure drop through the tap and fire/smoke damper is over
0.5 in. compared to a pressure drop calculated from SMACNA data with less than half that
value. Removing the sound trap to separate the fan discharge further from the elbow, and
using a shorter radius elbow with splitters to separate the elbow discharge further from the
riser tap would have improved the energy performance of this system. Sound levels would
likely have been better as well since the system effect losses through the trap caused the fan to
operate at much higher speed and sound power levels than it would with the sound trap
removed. Another option would have been to discharge the fan into a large plenum then tap
the riser into the bottom of the plenum.

DUCT DESIGN X105X

 FIGURE 38:
POOR DISCHARGE
CO N F I G U R AT I O N
R E S U LT I N G I N
SIGNIFICANT FAN
SYSTEM EFFECT

FIGURE 39:
MEASURED PRESSURE
IN A SYSTEM WITH
SIGNIFICANT FAN AND
DUCT SYSTEM EFFECT

Noise Control

Air distribution system noise can be controlled by one or both of the following strategies:
1. Reduce sound power levels at the source (the fan and turbulence in
duct systems).

2. Attenuate sound generated by the noise sources.

Typically both issues must be addressed. Reducing source sound power is generally the most
efficient, and sometimes results in the lowest first costs.

X106X DUCT DESIGN

Sound power can be reduced by considering:
ƒ Fan selection. Different fan types have different acoustic performance and the selection
of the fan size (wheel diameter) will also affect performance. See “Fan Selection
Criteria”.
ƒ Pressure drop. Lowering the system pressure drop allows for lower fan speed which
lowers sound power levels. The largest pressure drops are from coils, filters, and
dampers, which can be easily reduced by reducing face velocities, although often at high
costs. Duct fittings are the next biggest cause of both pressure drop and of noise due to
turbulence. These effects can be reduced by minimizing the number of fittings and by
proper fitting design as discussed under “General Guidelines”.
ƒ Terminal selection. VAV boxes and air terminals can be noisy but it is relatively easy
to avoid problems by following selection procedures from manufacturers and in this
document (see “Sizing VAV Reheat Boxes”). Attenuation measures include locating
noisy equipment well away from noise-sensitive spaces which reduces noise levels,
usually at low cost. Duct liners and attenuator are other strategies for attenuation and
are discussed in more detail below.

Duct Liners

Fiberglass duct liner has been used for many years in HVAC duct distribution systems.
Until recently, most engineers would not think twice about using duct liner for sound
attenuation. But more and more the use of liner is being questioned by indoor air quality
specialists and IAQ-conscious design engineers because of some potential problems
associated with the product or its application:
ƒ Duct liner can retain both dirt and moisture and thus may be a breeding ground for
microbial growth. The problem occurs primarily where humidity is very high for long
periods of time or where liquid water is present, such as at cooling coils or humidifiers.
ƒ The binding and air-surface facing of duct liner has been found in some cases to break
down over time and ends up being blown into occupied spaces as a black dust.
ƒ Where facing and binding have broken down or been damaged, or at poorly constructed
liner joints, fiberglass strands can break free and transferred to occupied spaces. Studies
to date have shown that fiberglass used in duct liner is not carcinogenic, but it is still
irritating to the skin.
ƒ While dust can collect on any surface in a duct system, including sheet metal ducts,
cleaning duct liner can be more difficult than other surfaces because its rough surface
traps dirt in crevices and because it is more easily damaged by mechanically cleaning
equipment such as brushes.

The jury is still out as to how significant these problems truly are. Clearly, some buildings
have had major problems that have been attributed at least in part to duct liner, particularly
issues with microbial growth in humid climates. But many more buildings that have
considerable lengths of lined duct are apparently “healthy.” Still, publicity about potential
problems and concerns about litigation are leading design engineers to look for alternative
products and designs to avoid, or at least mitigate, the use of duct liner.

DUCT DESIGN X107X

 But for sound attenuation, there are no simple substitutions for the benefits of duct liner.
Alternative designs and products are almost always more expensive, take up more space, and
use more fan energy due to increased pressure drops 26.

Options to attenuate noise in lieu of fiberglass duct liner include:

ƒ Sound traps. While widely mentioned as an alternative, sound traps usually contain
fiberglass and can harbor dirt and moisture just as readily as liner. Using a foil or
plenum rated plastic facing can protect the fiberglass or packless traps can be used to
avoid the fiberglass entirely but at extra cost and reduced effectiveness. As mentioned
above in the section on Fan Outlet Conditions, sound traps can actually increase noise
levels because of the higher fan energy required to overcome the pressure drop of the
sound trap. For example, sound attenuators on the plenum inlet of a fan box are
seldom effective as the resulting pressure drop may result in an increase in required fan
rpm, increasing sound more than the attenuator will reduce it!
ƒ Plenums. Abrupt discharge and intake plenums are effective at attenuating sound even
when unlined. However, they can increase pressure drop.
ƒ Alternative liner materials. Materials other than fiberglass liner are available, such as
closed cell foam. While they may avoid some of the problems with fiberglass liner, they
are usually less effective at sound attenuation.

In many, perhaps most, buildings, there simply is not enough space or not enough budget
available for these options to be implemented. Fiberglass liner may still be best or the only
viable option. Fortunately, the potential problems of duct liner can be at least partially
mitigated by covering it with a protective material like:
ƒ Perforated metal facing. Like sound traps, perforated liner is commonly considered a
good way to mitigate the problems of duct liner, but it too can still trap dirt and
moisture and air is still exposed to fiberglass. Foil or other facings can be used inside the
perforated liner to protect the fiberglass.
ƒ Foil and non-metallic facing. The acoustical benefits of duct liner can be partially
retained with foil and non-metallic facing films. These are standard options on most
VAV boxes.
ƒ “Tough” facing. Most liner manufacturers are producing liner with much more
resilient facing/binding materials designed to resist breakdown and damage to
mechanical cleaning.
ƒ Biostats. Liner can be treated with biostats to resist microbial growth. However, once
the biostat is covered with a film of dirt, its effectiveness may be reduced.

Finally, problems with liner can be minimized by locating it where a problem is less likely to
occur:

26
Refer to ARI Standard 885 for guidance on acoustical calculations.

X108X DUCT DESIGN

ƒ “Wet” sections. Avoid locating liner where it will be in direct contact with liquid water
such as at cooling coils and downstream of humidifiers. Most air handlers and many
large rooftop AC units can be specified with solid double wall construction in the
cooling coil sections to avoid insulation having direct contact with coil frames and
condensate pans.
ƒ Filters. Placing filters upstream of duct liner minimizes the liner’s exposure to dirt,
keeping it cleaner longer. Filters can also be located downstream of liner to prevent
degrading facing, binding materials, or fiberglass from being supplied to the space.

All in all, designing HVAC systems without duct liner is a major challenge and often an
expensive one. The best designs may be those that use duct liner only where needed for
sound attenuation, that locate it in clean and dry areas, and that protect it as best as possible
from damage and erosion with protective facings.

The following guidelines are recommended for duct liner:

ƒ Liner should be limited to the amount required for adequate sound attenuation. Advice
from an acoustical engineer, and perhaps some time and experimentation, will be
needed to determine exactly how much lining is actually necessary. Typically, liner is
only needed in fan discharge and inlet plenums, in main duct risers for a story or two,
and in VAV boxes. Duct mains on floors up to and after VAV boxes (other than the
box’s discharge plenum) are generally unlined. (This assumes that a limited amount of
flex duct is used between the VAV box and the diffuser as discussed in the section on
Duct Design.)
ƒ Liner should not be located in “wet sections” of air handlers (coil sections, humidifier
sections) where the manufacturer has an option for solid double-wall construction in
these sections. (This is not yet a common option on smaller air handlers and fan-coils,
unfortunately.)
ƒ “Tough” liner facings should be specified to improve resistance to erosion and damage.
ƒ In large air handlers where insulation may be damaged by personnel working around it,
perforated double wall construction should be specified.
ƒ Liner must be required to be protected from weather during construction and replaced
if it becomes wet.

DUCT DESIGN X109X

 SUPPLY AIR TEMPERATURE CONTROL

In most buildings the optimal setting for supply air temperature varies over time, often from
one hour to the next, and supply air temperature reset controls can provide significant energy
savings. This section describes some of the important design issues related to supply air
temperature control and includes recommended control sequences.

Optimal Supply Air Temperature

The optimal supply air temperature minimizes the combined energy for fan, cooling, and
heating energy. But this is a fairly complex tradeoff, and the optimal setpoint at any point in
time is not obvious.

Simulation can provide some insight into an optimal control strategy. Figure 40 and Figure
41 illustrate results for the Sacramento climate on two different days, one hot and the other
mild. In both figures, the top three charts show snapshots in time with energy consumption
plotted as a function of supply air temperature. These show, as expected, that as supply air
temperature increases, the fan energy goes up and cooling energy drops. On the hot day
(Figure 40), the supply air temperature that minimizes the total HVAC electricity changes
from 60°F in the morning to 50°F in the afternoon. At midday, it’s nearly a toss-up where
55°F is optimal but results are very close to those at 50°F and 60°F. The lower three graphs
show hourly results over the course of the day.

On the mild day, illustrated in Figure 41, the best choice is 60°F throughout the day because
it significantly reduces the amount of cooling energy with only a small increase in fan energy.
The 60°F setpoint also results in lower reheat energy.

X110X SUPPLY AIR TEMPERATURE CONTROL

These results illustrate the following general guidelines:
ƒ Use supply air temperature reset controls to avoid turning on the chiller whenever
possible. The setpoint should be the highest temperature that can still satisfy the cooling
demand in the warmest zone. Ideally, no chiller operation will be required until outdoor
air reaches somewhere between 60°F and 65°F. The warmer supply air temperatures in
cool and cold weather also reduces reheat at the zone level.
ƒ Continue to use supply air reset during moderate conditions when outdoor air
temperature is lower than about 70°F. In this range, the outdoor air is still providing a
portion of the cooling and it is worth spending a little extra fan energy to offset part of
the chiller demand.
ƒ Reduce the supply air temperature to its design setpoint, typically 53°F to 55°F, when
outdoor air temperature exceeds 70°F. At these warmer temperatures, the outdoor air is
providing little or no cooling benefit, and it is unlikely that any zones will require
reheat.

8:00 - 9:00 11:00 - 12:00

FIGURE 40:
14:00 - 15:00 Total
2.5 25 HVAC, CO MPAR I S O N O F H OT
Elec.
2.0 20
Electricity, W/ft2

D AY S I M U L AT I O N

Gas, Btu/h/ft2
Cooling,
1.5 15 Elec
R E S U LT S F O R T H R E E
1.0 10
Fans, S U P P LY A I R
0.5 5 Elec
T E M P E R AT U R E
0.0 0
Heating, S E T P O I N T S : 5 0 ° F,
50 55 60 50 55 60 50 55 60
Gas
Supply Air Temperature, F 5 5 ° F, A N D 6 0 ° F.

2.5 100
AUGUST 18.
HVAC Electricity,

Temperature, F
2.0 80 SA C R A M E N T O C L I M AT E
Outdoor Air
1.5 60
W/sf

1.0 40
0.5 20
0.0 0

20 100
Temperature, F
Heat Energy,

15 80
Outside Air
Btu/sf

60
10
40
5 20
0 0

25 100
Source Energy,

Temperature, F

20 80
Outdoor Air

15 60
Btu/sf

10 40
5 20
0 0
10:00

11:00

12:00

13:00

14:00

15:00

16:00

17:00

18:00

19:00
8/18
6:00

8/18
7:00

8/18
8:00

8/18
9:00

8/18

8/18

8/18

8/18

8/18

8/18

8/18

8/18

8/18

8/18

50 55 60 OAT

The top three charts show HVAC electricity and gas consumption at three snapshots in time. The bottom three show hourly
profiles for electricity, gas and source energy consumption.

SUPPLY AIR TEMPERATURE CONTROL X111X

 FIGURE 41: 8:00 - 9:00 11:00 - 12:00 14:00 - 15:00 Total
2.5 25 HVAC,
CO MPAR I S O N O F MI L D Elec.
2.0 20

Electricity, W/ft2

Gas, Btu/h/ft2
D AY S I M U L AT I O N Cooling,
1.5 15 Elec
R E S U LT S F O R T H R E E
1.0 10
S U P P LY A I R Fans,
0.5 5 Elec
T E M P E R AT U R E 0.0 0
Heating,
S E T P O I N T S : 5 0 ° F, 50 55 60 50 55 60 50 55 60
Gas
Supply Air Temperature, F
5 5 ° F, A N D 6 0 ° F.
MARCH 4. SACRAMENTO 2.5 100
HVAC Electricity,

Temperature, F
2.0 80

Outdoor Air
C L I M AT E
1.5 60
W/sf

1.0 40
0.5 20
0.0 0

20 100

Temperature, F
Heat Energy,

15 80

Outside Air
Btu/sf

60
10
40
5 20
0 0

25 100
Source Energy,

20 80

Temperature, F
Outdoor Air
15 60
Btu/sf

10 40
5 20
0 0
6:00

7:00

8:00

9:00

10:00

11:00

12:00

13:00

14:00

15:00

16:00

17:00

18:00

19:00
3/4

3/4

3/4

3/4

3/4

3/4

3/4

3/4

3/4

3/4

3/4

3/4

3/4

3/4
50 55 60 OAT

The top three charts show HVAC electricity and gas consumption at three snapshots in time.
The bottom three show hourly profiles for electricity, gas and source energy consumption.
The assumptions in the simulation are detailed in Appendix 6.

Recommended Sequence of Operation

The recommended control sequence is to lead with supply temperature setpoint reset in cool
weather where reheat might dominate the equation and to keep the chillers off as long as
possible, then return to a fixed low setpoint in warmer weather when the chillers are likely to
be on. During reset, employ a demand-based control that uses the warmest supply air
temperature that satisfies all of the zones in cooling.

Supply air temperature setpoint:

During occupied mode, the setpoint is reset from T-min (53°F) when the outdoor air
temperature is 70°F and above, proportionally up to T-max when the outdoor air
temperature is 65°F and below. T-max shall be reset using trim and respond logic within the

X112X SUPPLY AIR TEMPERATURE CONTROL

range 55°F to 65°F. When fan is off, freeze T-max at the maximum value (65ºF). While
fan is proven on, every 2 minutes, increase the setpoint by 0.2ºF if there are no zone cooling
requests. If there are more than two (adjustable) cooling requests, decrease the setpoint by
0.3ºF. A cooling request is generated when the cooling loop of any zone served by the
system is >99%. All values adjustable from fan graphic. (Note that 99% is used rather than
100% because some control system have rounding errors that sometimes prevent a software
value from ever reaching exactly 100%)
Supply Air Temperature Setpoint

FIGURE 42:
65 SAT reset by demand R E C O M M E N D E D S U P P LY
within this range
A I R T E M P E R AT U R E
60 RESET METHOD

55

Tmin = 53F

55 65 70 75
Outdoor Air Temperature

System Design Issues

Supply air temperature reset is usually a good idea in all California climates, though there
are some conditions where there will be limited benefit. Table 18 lists some factors affecting
the potential for energy savings.

Conditions Favoring SAT Reset Conditions that Reduce the Savings TABLE 18:
Potential for SAT Reset
CONDITIONS AFFECTING
Mild climate with many hours below 70°F. Dehumidification is necessary (typically not an
issue for California office buildings). THE IMPACT OF SUPPLY
VAV box minimum air flow setpoints of 30% Hot climate with few hours below 60°F. AIR TEMPERATURE
or higher. RESET
Low pressure loss air-side design, meaning there Inefficient air-side system.
is less penalty from higher airflow.
Skilled operating staff to maintain controls. Constant cooling loads that cannot be isolated
with a separate system.
Time varying levels of occupancy and interior Efficient part load fan modulation such as that
heat gain. provided by variable speed drives.

Supply air temperature reset is more than just an operational issue. There are several
important system design issues to consider to ensure that temperature reset can be
implemented successfully.
ƒ Size interior zone air flows so that the likely peak loads can be met at air temperatures
about 5 to 7oF higher than the minimum design temperature. This allows reset to occur
during cool weather and reduces reheat necessary in perimeter zones while still satisfying
cooling needs of interior zones. Sizing interior zones for more than 7 to 10 degrees of
reset unnecessarily increases first costs and may result in poor diffuser performance.

SUPPLY AIR TEMPERATURE CONTROL X113X

 ƒ Provide DDC control to the zone level with feedback regarding temperature and
setpoint from each zone.
ƒ Use a separate cooling system for unique loads such as computer centers so that they do
not force the whole building system to operate at a low fixed temperature.
ƒ Maximize air distribution system efficiency through supply air pressure controls and
low-pressure loss design. This strategy reduces the energy penalty for increased air flow
when supply air temperature is reset upwards.
ƒ Integrate the sequence of operations with supply air pressure reset control. In the
method described below, supply air temperature is reset based on a combination of
outdoor temperature and zone cooling demand, while the recommended supply
pressure controls are based on VAV box damper position (see “Conclusions”).
ƒ Include a clear specification for the sequence of operations.
ƒ Include commissioning requirements in the specifications. Reset controls can be highly
unstable unless well tuned.

Code Requirements

In previous versions of Title 24 supply air temperature reset was required for one-fourth of
the difference between the supply design temperature and the design space temperature. For
example, if the system design leaving temperature was 54°F and the design space temperature
was 74°F, then 20°F/4=5°F of reset (from 54°F to 59°F) was required.

As of 2005, reset is no longer required by Title 24 for VAV systems with variable speed
driven fans. This is because variable speed drives are so effective at reducing fan energy at
low airflow rates that the fan energy savings resulting from low supply air temperatures offset
the savings in reheat energy resulting from higher supply air temperatures. Although no
longer required by the 2005 energy code, supply temperature setpoint reset as detailed in this
section is still cost effective even with variable speed drives. Furthermore, it is likely that the
SAT reset requirement will be reinstated into Title 24 in 2008.

X114X SUPPLY AIR TEMPERATURE CONTROL

 FAN TYPE, SIZE AND CONTROL

This section discusses how to select fans for typical large VAV applications. Information
includes the best way to control single and parallel fans, as well as presentation of two
detailed fan selection case studies.

Fan Selection Criteria

The factors to consider when selecting a fan include:

ƒ Redundancy – a single fan or multiple fans.
ƒ Duty – CFM and static pressure at design conditions.
ƒ First cost – more efficient fans are often more expensive.
ƒ Space constraints – a tight space may limit fan choices.
ƒ Efficiency – varies greatly by type and sizing.
ƒ Noise – different fan types have different acoustic performance.
ƒ Surge – some fan selections are more likely to operate in surge at part-load conditions.

These issues are elaborated on below and in the case studies that follow.

FAN TYPE, SIZE AND CONTROL X115X

 Redundancy

One of the first questions to answer when selecting a fan is whether to use a single fan or
parallel fans. The primary advantage of parallel fans is that they offer some redundancy in
case one of the fans fails or is down for servicing. Parallel fans are sometimes necessary
because a single fan large enough for the duty is not available or because a single fan would
be too tall. Of course, parallel fans can also create space problems (e.g., two parallel fans
side-by-side are wider than a single fan). Parallel fans are also more expensive and create
more complexity in terms of fan control and isolation (as discussed below).

Type

Fans are classified in terms of impeller type (centrifugal, axial, mixed flow), blade type, and
housing type. See Table 19 Fan Classification.

The first step when selecting a fan type is to limit the choices based on the application. For
example, for medium to large supply or return fans (e.g., >30,000 CFM), the top choices
include housed airfoil and plenum airfoil centrifugal fans, but may also include multiple
forward curved centrifugal fans or mixed flow fans. 27 For small systems (<15,000 CFM),
forward curved fans are generally the optimum choice due to low first costs. All these fan
types are possible in the middle size range.

27
Vane-axial fans were once a common option as well when variable speed drives were new and expensive because
they were very efficient at part load, but they are seldom used anymore due to high first costs, the need for sound
traps on inlet and outlet, and high maintenance costs for variable pitch fans. Vane-axial fans were therefore not
considered in our analysis.

X116X FAN TYPE, SIZE AND CONTROL

 CENTRIFUGAL (flow radial to fan shaft) TABLE 19:
Blade Type FAN CLASSIFICATION
Backward Inclined
Straight/Flat Blade (BI)
Air Foil (AF)
Radial – (typically only for industrial applications)
Forward Inclined
Straight/Flat Blade
Forward Curved
Housing Type
Scroll Type (housed fan)
Single Width (ducted inlet from one side)
Double Width (air enters from two sides)
Plug Type
In-line (tubular)
Roof-top (dome) – (used for low static exhaust)
Plenum
AXIAL (flow parallel to fan shaft)
Blade Type
Slanted Blades
Air Foil
Cambered Twist
Housing Type
Propeller – (common for relief, low pressure exhaust)
Tube-axial
Vane-axial
Fixed Pitch
Adjustable Pitch
Variable Pitch
MIXED FLOW (hybrid – part centrifugal and part axial)
Blade Type
Contoured Single Thickness
Air Foil
Housing Type
In-line (tubular)

With large built-up systems and custom units, the designer’s first choice should be a housed
(scroll type) airfoil centrifugal fans. This is the most efficient fan type and, for built-up
systems when the cost of the discharge plenum is included, a housed fan system will generally
be less expensive than a plenum (plug type) fan system. The major disadvantage of housed
airfoil fans is noise; they generate high sound power levels in the low frequency bands which
are very difficult to attenuate.

If the housed airfoil fan will not fit or meet the acoustic criteria, the next choice should be a
plenum or mixed flow fan. In terms of efficiency, the housed airfoil fan is the best followed
by the mixed flow fan. The plenum fan is the least efficient choice for medium pressure
systems unless space constraints would cause a housed fan to be installed in a manner that
would lead to high system effects.

FAN TYPE, SIZE AND CONTROL X117X

 The characteristic of the fan types now most commonly used for VAV supply fan
applications are summarized in the Table below.

TABLE 20: Typical Space

Fan Type Applications First Cost Constraints Efficiency Noise Other
COMPARISON OF Housed forward- CFM <25000 Lowest Requires more Less efficient Slower Small surge
curved (FC) 1 in. <SP<3.5 in. space than than housed speed than region; may be
COMMON VAV SUPPLY centrifugal plenum fans for airfoil for >3 housed unstable for
smooth in., better or airfoil so parallel fans at
FAN TYPES
discharge same as housed usually low airflow, high
airfoil, BI for quieter. static.
<2 in.
Housed CFM <70000 Medium Requires more Somewhat less Similar to Larger surge
backwardly 2 in.<SP <6 in. Low space than than housed housed region than
inclined (BI) plenum fans for airfoil. airfoil. housed airfoil.
centrifugal smooth
discharge
Housed airfoil CFM <100000 Medium Requires more Highest Noisier than Small surge
(AF) centrifugal 2 in.<SP <8 in. space than efficiency. plenum and region; high shut
(double width) plenum fans for forward- off pressure.
smooth curved
discharge
Mixed flow CFM <60000 Highest Good for inline Similar to Quieter Small surge
2 in.<SP <6 in. use. housed airfoil than other region.
but drops off in housed fans.
surge region.
Plenum airfoil CFM <80000 High. Requires least Lower Quietest Large surge
centrifugal 2 in.<SP <6 in. Cost of space, efficiency than when region.
discharge particularly for housed airfoil plenum
plenum multiple fans unless space is effects
also must constrained. included
be
included

Fan Pressure Ratings

There is a great deal of confusion regarding the issue of fan total pressure drop versus fan
static pressure drop. This section attempts to clarify the issue of total versus static pressure.
The work that a fan must do is proportional to the total pressure rise across the fan. Total
pressure consists of velocity pressure and static pressure. The total pressure rise across a fan
is:

ΔTP = TPLEAVING − TPENTERING

= SPLEAVING + VPLEAVING − SPENTERING − VPENTERING

Most vane-axial fans are rated based on total pressure drop. However, most other fans
types (e.g., centrifugal fans) are not due to historical standard rating practices. It is
important to find out from the manufacturer's catalogue under what conditions the fan
ratings were developed.

Centrifugal fans are typically rated using a combination of inlet and outlet static and
velocity pressures defined as follows:

X118X FAN TYPE, SIZE AND CONTROL

 SPrating = TPLEAVING − TPENTERING − VPLEAVING
= SPLEAVING − SPENTERING − VPENTERING

This very confusing rating criterion is usually called the "fan static pressure" because it is
equal to the static pressure rise across the fan when the inlet velocity pressure is zero, which
is the condition when the fan inlet is in an open plenum (velocity = 0) as is the case when
the fan rating test is performed. To confuse matters further, it is also often called the "total
static pressure" to differentiate it from the "external static pressure," which is the pressure
drop external to a packaged air handler or air conditioner (i.e., the total static pressure drop
less the pressure drop of components within the air handler).

As noted under Duct Design, it is difficult to accurately calculate the total pressure drop at
design conditions, so engineers typically estimate or “guesstimate” the design pressure drop.
Therefore, it does not really matter that the fans are rated in this confusing manner because
the drop calculation is an educated guess in most cases.

Where total versus static pressure becomes important is when comparing housed
centrifugal fans versus plenum fans or axial fans. Housed fans are nominally more efficient
because they use the housing to concentrate all the air coming off of the wheel into a small
area, which creates higher static pressure at the outlet (leading to higher efficiency) but also
higher velocity pressure. If this velocity pressure is dissipated by poorly designed elbows
and other fittings at the fan discharge, then a housed fan can actually be less efficient than a
plenum fan in the same application because it is operating against a higher total external
pressure. A plenum fan is primarily creating static pressure in a pressurized plenum and is
less vulnerable to system effects due to high velocities at the discharge.

Visualizing Fan Performance

Fan curves and selection software provided by the manufacturers give a lot of useful
information about fan performance. However it is hard to visualize the operation of a fan
across the full range of operating conditions using a typical manufacturer’s fan curve. In
particular, the challenge is in determining the fan efficiency at any point other than the
design condition. Figure 43 shows a fan selection for a 60 in. plenum fan. The data in the
upper left hand corner indicates that the fan has a 63% static efficiency at the design point.

FAN TYPE, SIZE AND CONTROL X119X

 FIGURE 43:
A TYPICAL
MANUFACTURER’S FAN
CURVE (60 IN. PLENUM
FAN)

While developing this Guidelines, the authors developed the Characteristic System Curve
Fan Model (see Appendix 12 and Hydeman and Stein, January, 2004 ), which can be used
to develop three dimensional fan curves. These curves add fan efficiency to the z-axis on top
of the pressure (y-axis) and volume (x-axis) of the manufacturer’s curve. Figure 44 shows a
66 in. plenum airfoil fans and Figure 45 shows a 49 in. housed airfoil fan. Looking at Figure
44 and Figure 45, it is easy to see the breadth of the high efficiency region for the airfoil fan
across a range of operating conditions.

FIGURE 44:
THREE-DIMENSIONAL
FAN CURVE FOR 66 IN.
PLENUM AIRFOIL FAN

X120X FAN TYPE, SIZE AND CONTROL

 FIGURE 45:
THREE-DIMENSIONAL
FAN CURVE FOR 49 IN.
HOUSED AIRFOIL FAN

Another way of evaluating and comparing fans is to look at “Gamma Curves”. Any point in
fan space (CFM, SP) is on a characteristic system curve (a parabola through that point and
through the origin). Each characteristic system curve is defined by a unique system curve
coefficient (SCC), which can be calculated from any point on that characteristic system
curve. Gamma (γ) is defined as the negative natural log of SCC. (Gamma is easier to view
on a linear scale than SCC.)

ΔP
SCC = γ ≡ − ln(SCC )
CFM 2

Figure 46 is the gamma curve for Cook 60 in. plenum airfoil fan (600CPL-A). One of the
useful features of a gamma curve is that it collapses all of the performance data for a fan into
a single curve that can be used to calculate fan efficiency at any possible operating condition.
For example, the point 89,000 CFM and 6 in. w.c. has a gamma value of 21 which
corresponds to a fan efficiency of 55%. Similarly, the point 63,000 CFM and 3 in. w.c. also
has a gamma value of 21 and a fan efficiency of 55%. Gamma curves can be developed
using a handful of manufacturer’s data points and then used to quickly compare several fan
types and sizes (see Figure 47 and Figure 48). Figure 48, for example, shows three sizes of
plenum fans. It also shows that the 49 in. housed airfoil is more efficient than any of these
plenum fans under any operating conditions. Gamma curves are also useful for seeing the
relationship between the peak efficiency and the surge region. For plenum fans, for example,
the peak efficiency is right on the border of the surge region (see Figure 47). For airfoil fans,
however, the peak efficiency is well away from the surge region (see Figure 47 and Figure
49).

FAN TYPE, SIZE AND CONTROL X121X

 FIGURE 46:
Efficiency versus Gamma for 60" Plenum Fan
GAMMA CURVE
Surge and Non-Surge Regions
60" Plenum Airfoil (Surge)
80%
60" Plenum Airfoil (Non-Surge)
70%

60%

50%
Efficiency

40%

30%

20%

10%

0%
17 19 21 23
Gamma = -ln(system curve coefficient)

FIGURE 47:
4 Fan Types - Surge and Non-Surge Regions
GAMMA CURVES FOR
FOUR FAN TYPES 80%

70%

60%

50%
Efficiency

40%

30% 60" Plenum Airfoil (Surge)

60" Plenum Airfoil (Non-Surge)
66" BI (Surge)
20% 66" BI (Non-Surge) The point
54" Mixed Flow (Surge) 54,000 CFM, 4"
10% 54" Mixed Flow (Non-Surge) has a gamma
49" AF (Surge) value of 20.4
49" AF (Non-Surge)
0%
17 19 21 23
Gamma

X122X FAN TYPE, SIZE AND CONTROL

 FIGURE 48:
Surge and Non-Surge Regions for 5 Fan Types, Several Sizes
GAMMA CURVES FOR
90%
54" Plenum Airfoil SEVERAL FAN TYPES
80% AND SIZES
60" Plenum Airfoil

70%
66" Plenum Airfoil

60% 49" Housed Airfoil

Fan Efficiency

50% 27" Forward Curve

40" Housed Airfoil

40%
49" Backward Inclined
30%
54" Mixed Flow Airfoil
20%

10%

0%
16 17 18 19 20 21 22 23 24 25
Gamma = -Ln(SCC)

Cook Housed Airfoil Fans 120 CADWDI FIGURE 49:

90% surge and non-surge 135 CADWDI GAMMA CURVES FOR
non-surge region only regions displayed for
displayed for this size and 150 CADWDI
this size and up A L L CO O K H O U SE D
80% down 165 CADWDI
180 CADWDI AIRFOIL FANS
70% 195 CADWDI
210 CADWDI
225 CADWDI
60%
Static Fan Efficiency

245 CADWDI
270 CADWDI
50%
300 CADWDI
330 CADWDI
40%
365 CADWDI
402 CADWDI
30% 445 CADWDI
490 CADWDI
20% 540 CADWDI
600 CADWDI
10% 660 CADWDI
730 CADWDI
0%
13 14 15 16 17 18 19 20 21 22 23 24 25 26
Gamma = -ln(SCC)

FAN TYPE, SIZE AND CONTROL X123X

 FIGURE 50:
18" to 73" Greenheck Airfoil Fans
GAMMA CURVES FOR Models 18AFDW to 73AFDW
Non-surge Region Only
ALL GREENHECK
HOUSED AIRFOIL FANS 90%
( N O N -S U R G E R E G I O N
80%
O N LY )
70%

Efficiency 60%

50%

40%

30%

20%

10%

0%
15 16 17 18 19 20 21 22 23 24 25
Gamma

FIGURE 51:
5 Sizes of Cook Backward Inclined - Non-Surge Region Only
GAMMA CURVES FOR (CF-SWSI sizes 12, 19.5, 30, 49, 73)
(intermediate sizes are also available)
SO ME CO O K B AC KWARD
80%
INCLINED FANS

70%

60%

50%
Efficiency

40%

30%

20%

10%

0%
12 14 16 18 20 22 24
Gamma

X124X FAN TYPE, SIZE AND CONTROL

 17 Sizes of Cook Mixed Flow Fans - Non-Surge Region Only
FIGURE 52:
(Model QMX-HP Sizes 9" to 54")
GAMMA CURVES FOR
80%
A L L CO O K A I R F O I L
70% MIXED FLOW FA NS

60%

50%
Efficiency

40%

30%

20%

10%

0%
12 14 16 18 20 22 24
Gamma

Fan System Component Models for Motors, Belts and Variable Speed Drives

In addition to the Characteristic System Curve Fan Model for representing fan performance
at any operating condition, the authors also developed component models for the part load
performance of the other components in a fan system: the motor, the belts and the variable
speed drive. These component models are described in more detail in Appendix 12. The
figures below summarize the part load performance of each component.

FIGURE 53:
M OTO R PA RT- LO A D
EFFICIENCY EXAMPLE
S H O W I N G C U R R E N T,
POWER FACTOR AND
EFFICIENCY

FAN TYPE, SIZE AND CONTROL X125X

 FIGURE 54:
1.2
F A N B E LT E F F I C I E N C Y

Percent of full load motor & VFD efficiency

VS. BRAKE
HORSEPOWER 1

0.8

Manuf #1
0.6
Manuf #2
0.4
From Saftronics/MotorMaster (assumes 100% speed
= 100% MHP)
0.2 from ASHRAE preprint of Transactions paper by
Gao

0
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percent MHP

FIGURE 55:
CO MB I N E D MOT O R AN D
VARIABLE SPEED DRIVE
E F F I C I E N C Y D ATA F R O M
FOUR SOURCES

Fan Selection Case Studies

This section walks through the process that an engineer is likely to go through when
selecting a fan for a typical built-up air handler. The issues are generally similar for large
packaged or custom units but the choices of fan types and sizes are likely to be limited by the

X126X FAN TYPE, SIZE AND CONTROL

air handler manufacturer. In this section, two supply fan case studies illustrate some of
issues:

Case Study Design Condition Num. Fans Fan Types Compared

A 54,000 CFM at 4 in. 1 Housed Airfoil, Plenum,
BI, Mixed Flow
B 145,000 CFM at 4 in. 2 Housed Airfoil, Plenum

Several conclusions can be drawn from these case studies:

1. Use housed airfoil fans where they meet space and noise constraints.
These fans are generally more efficient and less likely to operate in surge
than plenum fans. They are also generally less expensive than plenum or
mixed flow fans.
2. Control fans using static pressure setpoint reset (see discussion below).
This can save up to 50% of the fan energy compared to a fixed setpoint
static control. It will also greatly reduce the operation of fans in surge,
which can lead to accelerated bearing wear.
3. For multiple fan systems, stage fans based on the pressure control scheme
shown in Figure 92 and Figure 93.

Case Study A

The first case study is a hypothetical example with a relatively small fan for which four types
of fans are available. The first step is to use manufacturer’s software to compare the
efficiency at the design point, and to compare first cost, motor size, and acoustics. It is
important to look not just at the fan cost, noise, efficiency, and motor size, but also at the
fan curve and where the design point lies relative to the surge line, which is often labeled
“Do not select to the left of this line.” Different fan types have fundamentally different
relationships between peak efficiency and surge. Housed airfoil fans, for example, have their
peak efficiency well to the right of the surge line. Plenum fans, however, are at their highest
efficiency right at the surge line.

Figure 56 shows the Loren Cook choices for housed airfoil (Model CADWDI) and housed
backward inclined (Model CF). Figure 57 indicates the Cook choices for plenum airfoil
(CPL-A) and airfoil mixed flow (QMX-HP). Each of these figures has two separate tables.
The top table shows data for a number of fans that will meet the design criteria, including
the model number, the design airflow (cfm), the design static pressure, the brake horsepower,
the recommended motor horsepower, the fan speed (rpm), the static efficiency (SE), the
weight, the relative cost, a budgetary price, an estimation of the annual operating costs, and a
payback. The operating costs are based on assumptions built into the manufacturer’s
software that should be taken with a large grain of salt. Assumptions on static pressure
control alone can have up to a 50% decrease in annual energy usage. The bottom table
presents wheel size, construction class, and sound power data for the same fans. As reflected

FAN TYPE, SIZE AND CONTROL X127X

 in these figures, the plenum fans have considerably lower static efficiency (SE) than the other
types.

FIGURE 56:
CASE STUDY A -
SELECTION SOFTWARE -
HOUSED AIRFOIL AND
BI CHOICES

FIGURE 57:
CASE STUDY A -
SELECTION SOFTWARE -
PLENUM AND MIXED
FLOW CHOICES

In order to account for the fact that the plenum fans might have a lower total pressure drop
due to reduced system effects, we reselected the plenum fans at a design condition of 3.5 in.,

X128X FAN TYPE, SIZE AND CONTROL

rather than the 4 in. used for the other fan types (see Figure 58). This is a somewhat
arbitrary assumption and assumes that 0.5 in. of the 4 in. of external pressure is due to
system effects near the high velocity discharge of the airfoil, BI, and mixed flow fans. A
plenum fan would not be subject to these system effects because of the low velocity pressure
at the fan discharge. Figure 58 shows that the 66 in. plenum fan has the highest efficiency
but the fan curve shows that the design point is too close to the surge region (see Figure 59).
As this fan unloads, it is likely to operate in surge, particularly if it is controlled against a
fixed static pressure setpoint (see discussion under case study B). Therefore, 60 in. is the best
plenum choice (see Figure 60).

FIGURE 58:
CASE STUDY A -
SELECTION SOFTWARE -
P L E N U M C H O I C E S AT
LOWER DES IG N
PRESSURE

FIGURE 59:
CASE STUDY A - 66 IN.
PLENUM FAN DESIGN
POINT

FAN TYPE, SIZE AND CONTROL X129X

 FIGURE 60:
CASE STUDY A - 60 IN.
PLENUM FAN DESIGN
POINT

Part Load Performance

We selected one or two fans of each fan type for further analysis. Using a handful of
manufacturer’s data points, we developed Characteristic Fan Curve models for each fan.
Part load performance depends on the shape of the true system curve. If static pressure
setpoint reset is perfectly implemented, the true system curve runs from the design point
through 0 in. at 0 CFM and the fans are all constant efficiency since this is a characteristic
system curve. If however, static pressure setpoint reset cannot be perfectly implemented (as
is typically the case in real applications), the true system curve will run through some non-
zero static pressure at 0 CFM and fan efficiency will not be constant. In order to bound the
problem, we evaluated the fans using both perfect static pressure setpoint reset and no static
pressure setpoint reset (fixed SP of 1.5 in. at 0 CFM) (see Figure 61). With perfect reset, the
fan efficiency is constant throughout part load operation (see Figure 61). Figure 62 shows
the design efficiency of each of the fans that we simulated.

X130X FAN TYPE, SIZE AND CONTROL

 FIGURE 61:
Case Study A - System Curves
4.5
CASE STUDY A -
SYSTEM CURVES
4 Perfect SP Reset
No SP Reset
3.5 Perfect SP Reset (3.5" Design SP)
static pressure (in. w.g.)

No SP Reset (3.5" Design SP)

3

2.5

1.5

0.5

0
0 10000 20000 30000 40000 50000 60000
CFM

FIGURE 62:
Case Study A - Design Pt. Efficiency
CASE STUDY A -
78%
DESIGN POINT
76% EFFICIENCY
74%

72%

70%

68%

66%

64%

62%

60%
60" Plenum 60" BI 66" BI 49" AF 44.5" AF 54" Mixed
Fan (3.5" Flow
Design SP)

With no static pressure setpoint reset, the fan efficiency varies at part load. Figure 63 shows
that the efficiency actually increases slightly as the fan starts to ride down the system curve
and then decreases at very low load.

FAN TYPE, SIZE AND CONTROL X131X

 FIGURE 63:
Case Study A - Efficiency vs CFM for Several Fans Systems on
C A S E ST U DY A - PA RT System Curve through 54,000/4" and 0/1.5" (i.e. No SP Reset)
90%
LOAD FAN E F F I C I ENC Y
80%

70%

60%

Fan Efficiency 50%

40%
60" Plenum Fan (3.5" Design SP)
60" BI
30%
66" BI
20% 49" AF
44.5" AF
10%
54" Mixed Flow

0%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Pct Design CFM

Figure 64 is identical to Figure 63 except that it only shows the non-surge region, i.e., where
the fans go into surge. The 66 in. BI fan, for example, goes into surge at 85% flow on this
system curve. Interestingly, the efficiency of the 61 in. BI in the surge region is similar to
that of the other housed fans, indicating that the issue with surge is not efficiency but control
stability, vibration, and noise.

FIGURE 64:
Case Study A - Efficiency vs CFM - Non-Surge Region Only
C A S E ST U DY A - PA RT System Curve through 54,000/4" and 0/1.5" (i.e. No SP Reset)
90%
LOAD E F F I C I ENC Y (N O N-
SURGE R E G I O N O N LY ) 80%

70%

60%
Fan Efficiency

50%

40% 60" Plenum Fan (3.5" Design SP)

60" BI
30% 66" BI
49" AF
20%
44.5" AF

10%
54" Mixed Flow

0%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Pct Design CFM

X132X FAN TYPE, SIZE AND CONTROL

Figure 65 shows the fan power of the Case Study A fan systems as a function of CFM. It
includes the part load efficiency of the fan, belts, motor, and variable speed drive.

FIGURE 65:
Case Study A - KW vs CFM for Several Fans Systems on
System Curve through 54,000/4" and 0/1.5" (i.e. No SP Reset) CASE STUDY A - KW
50
VERSUS CFM
45
60" Plenum Fan
40
60" Plenum Fan (3.5" Design SP)
Fan System KW (includes Belts, Motor, VSD)

60" BI
35 66" BI
49" AF
30 44.5" AF
54" Mixed Flow
25

20

15

10

0
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Pct Design CFM

Figure 66 reflects another way to represent the part load efficiency of some fans evaluated in
Case Study A. It shows the design point for the case study and how the fan efficiency
changes when moving away from the system curve of the design point.

FIGURE 66:
4 Fan Types - Surge and Non-Surge Regions
CASE STUDY A - GAMMA
80% CURVES

70%

60%

50%
Efficiency

40%

30% 60" Plenum Airfoil (Surge)

60" Plenum Airfoil (Non-Surge)
66" BI (Surge)
20% 66" BI (Non-Surge) The point
54" Mixed Flow (Surge) 54,000 CFM, 4"
10% 54" Mixed Flow (Non-Surge) has a gamma
49" AF (Surge) value of 20.4
49" AF (Non-Surge)
0%
17 19 21 23
Gamma

FAN TYPE, SIZE AND CONTROL X133X

 Extrapolating from Part Load Performance to Annual Energy Cost

There are several ways to estimate annual energy cost for a fan system. One method involves
developing a hypothetical fan load profile using DOE-2 and then applying the part load kW
to each point in the load profile. Figure 67 shows histograms of three load profiles
developed using DOE-2 as part of the VAV box sizing simulation analysis (See “Appendix
6”). These profiles represent an office building in the California Climate zone 3 (a mild
coastal environment that includes San Francisco). The High Load Profile assumes that most
of the lights and equipment are left on during occupied hours. The 24/7 profile represents
continuous fan operation.

FIGURE 67:
Simulated Fan Load Profiles
C ASE ST UDY A - LOAD
PROFILES 2000

1800

1600 Basecase
High Internal Loads
1400
24/7 Fan Operation
1200
hours/yr

1000

800

600

400

200

0
from from from from from from from from from from from from from from from from from from from from
0 to 5 to 10 to 15 to 20 to 25 to 30 to 35 to 40 to 45 to 50 to 55 to 60 to 65 to 70 to 75 to 80 to 85 to 90 to 95 to
5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100
% of Design Flow

Figure 68 shows the annual energy cost with perfect static pressure setpoint reset for each
fans and load profiles evaluated. Notice that the plenum fan at 3.5 in. uses about as much
energy as the housed fans at 4 in..

X134X FAN TYPE, SIZE AND CONTROL

 FIGURE 68:
Est. Annual Fan Energy for Several Fan Types Basecase Load Profile
Perfect SP Reset High Load Profile C A S E S T U D Y A R E S U LT S
Design Condition of 54,000/4" 24/7 Load Profile
- P E R F E C T S TAT I C
$2,500 PRESSURE RESET
$2,400 $2,279
$2,258 $2,297 $2,257 $2,258 $2,276
$2,300
$2,200
$2,091 $2,098 $2,016 $2,017 $2,062 $2,055
$2,100
Annual Fan Energy Cost (at $0.12/kwh)

$2,000
$1,900
$1,800
$1,700
$1,600
$1,500
$1,400 $1,250
$1,254 $1,266 $1,231 $1,232 $1,248
$1,300
$1,200
$1,100
$1,000
$900
$800
$700
$600
$500
$400
$300
$200
$100
$-
60" Plenum Fan 60" BI (71.2%) 66" BI (75.5%) 49" AF (75.4%) 44.5" AF (73.0%) 54" Mixed Flow
(3.5" Design SP) (73.4%)
(62.4%)

Figure 69 shows the annual energy cost with no static pressure setpoint reset for each of the
fans and load profiles evaluated. Notice that the plenum fan has consistently higher energy
costs than the housed airfoil and BI fans. Also energy costs in Figure 69 are more than
double the costs in Figure 68, which clearly implies that the type of fan selected is not nearly
as important as how it is controlled.

Est. Annual Fan Energy for Several Fan Types Basecase Load Profile FIGURE 69:
Fixed SP Setpoint = 1.5" High Load Profile
C A S E S T U D Y A R E S U LT S
Design Condition of 54,000/4" 24/7 Load Profile

$8,000 – N O S TAT I C P R E S S U R E
$8,000
$7,580 RESET
$7,000 $6,708
$6,508
Annual Fan Energy Cost (at $0.12/kwh)

$6,257
$6,081
$6,000

$5,000
$4,437 $4,352
$4,116 $4,047
$3,929 $3,862
$4,000 $3,555
$3,602
$3,246 $3,171
$3,057 $2,998
$3,000

$2,000

$1,000

$-
60" Plenum Fan 60" BI (71.2%) 66" BI (75.5%) 49" AF (75.4%) 44.5" AF 54" Mixed Flow
(3.5" Design SP) (73.0%) (73.4%)
(62.4%)

FAN TYPE, SIZE AND CONTROL X135X

 Noise

Figure 70 summarizes acoustic data shown in Figure 56 and Figure 57 for some of the
evaluated fans. Because low frequency noise is much harder to attenuate than high
frequency noise, the critical octave bands are OB1 (63 Hz) and OB2 (125 Hz). While the
plenum fan appears to be considerably noisier than the other types, this figure does not
present a fair comparison since it does not include the effect of the discharge plenum. Figure
71, from the Carrier Air Handler Builder Program, shows air handler discharge acoustic data
for a housed airfoil and a plenum fan, and includes the attenuation of the discharge plenum.
The plenum fan has considerably better acoustic performance than the housed airfoil fan at
the low frequency octave bands.

FIGURE 70: Sound Power Data for 4 Fan Types (Design Condition: 54,000 CFM, 4")

CASE STUDY A –
120
CARRIER ACOUSTIC
D ATA ( W I T H C A S I N G ) 100 54" Mixed Flow (inlet) - 73% effic.
54" Mixed Flow (outlet) - 73% effic.
49" AF (inlet) - 76%
80 60" Plenum (outlet*) - 64%
66" BI (inlet) - 74%
sound power (dB)

60

40
*Plenum fan sound data does not
include the attenuation of the
plenum
20

0
OB1 OB2 OB3 OB4 OB5 OB6 OB7 OB8
octave band

X136X FAN TYPE, SIZE AND CONTROL

 Discharge Acoustic Data from Carrier Air Handler Builder Software
FIGURE 71:
(54,000 CFM / 4") CASE STUDY A -
120
Plenum A CO U S T I C D ATA ( N O
AF CASING)
100

80
Sound Power (dB)

60

40

20

0
OB1 OB2 OB3 OB4 OB5 OB6 OB7 OB8
Octave Band

Curiously, the McQuay air handler selection software showed plenums fans having little or
no sound advantage over housed airfoil (AF) and forward-curved (FC) fans (Table 21). The
differences could have to do with the way the discharge sound power is measured. For
example, the outlet for the Carrier discharge plenum is field cut so clearly the manufacturer
is making some assumption about the size and location of that outlet when rating the sound
power.

Another suspicious aspect of this table is the very high efficiency of the Carrier plenum fans.
In fact, the Carrier catalog shows these fans having lower efficiency that is more in line with
the other air handler and fan manufacturers’ data. All this information simply reinforces
Rule #1 of HVAC design: “Do not always believe manufacturers’ data.”

FAN TYPE, SIZE AND CONTROL X137X

 Discharge Sound Power TABLE 21.
Manuf. Unit CFM TP FanType Dia. BHP RPM SE OB1 OB2 OB3 OB4 OB5 OB6 OB7 OB8 MANUFACTURERS’ AIR
Carrier 92 54,000 4 Plenum 44.4 630 77% 81 82 79 73 75 72 67 67 HANDLER SELECTION
Carrier 92 54,000 4 AF 57.1 1074 60% 104 107 99 94 88 82 74 65 SOFTWARE FAN DATA
Carrier 92 50,000 4 Plenum 41.1 615 77% 81 82 78 73 74 72 67 66

Carrier 92 50,000 4 AF 50.3 1028 63% 103 106 98 94 87 82 73 65

Carrier 92 45,000 4 Plenum 37.8 598 75% 80 81 77 72 74 71 66 65

Carrier 92 45,000 4 AF 42.5 971 67% 102 105 97 92 86 80 72 63

Carrier 92 40,000 4 Plenum 34.6 585 73% 80 81 76 72 73 70 65 65

Carrier 92 40,000 4 AF 35.9 918 70% 101 104 96 91 85 79 71 62

Carrier 61 30,000 4 Plenum 24.6 944 77% 75 79 77 72 73 73 69 67

Carrier 61 30,000 4 AF 26.7 951 71% 99 102 95 90 84 78 69 61

Carrier 61 30,000 4 FC 29.4 539 64% 99 99 97 91 84 78 69 59

Carrier 39 20,000 4 Plenum 16.7 1196 75% 69 78 77 70 71 72 70 67

Carrier 39 20,000 4 AF 18.5 1389 68% 98 98 97 89 83 77 69 60

Carrier 39 20,000 4 FC 20.5 722 61% 98 98 93 94 84 78 70 61

McQuay 50 20,000 4 Plenum 40.25 18.5 954 68% 88 96 93 91 87 85 79 74

McQuay 50 20,000 4 AF 33 16.4 1021 77% 91 96 91 89 86 80 76 68

McQuay 50 20,000 4 FC 27.62 19.8 705 64% 90 88 85 85 79 78 74 68

McQuay 65 30,000 4 Plenum 49 27.7 787 68% 95 98 91 89 85 81 75 70

McQuay 65 30,000 4 AF 36.5 25.7 980 73% 93 98 93 91 88 82 78 70

McQuay 65 30,000 4 FC 33 29.6 590 64% 89 86 83 83 78 77 72 65

McQuay 90 40,000 4 Plenum 54.25 37.0 730 68% 96 99 92 90 86 82 76 71

McQuay 90 40,000 4 AF 40.25 37.8 923 67% 95 100 95 93 90 84 80 72

McQuay 90 40,000 4 FC 40.25 41.5 513 61% 98 97 95 93 91 86 81 84

Trane 66 30,000 4 FC 33 30.8 595 61% 100 97 97 95 90 86 80 74

Trane 66 30,000 4 AF 36.5 26.7 970 71% 96 100 96 91 86 79 72 68

Trane 66 30,000 4 Q 44.5 29.6 1191 64% 94 96 96 97 95 91 85 77

FAN TYPE, SIZE AND CONTROL X138X

First Cost

Figure 72 shows the budget prices from the Cook software, which only includes the fan
itself, not the cost of discharge or inlet plenums. This figure shows the housed airfoil fans to
be $100 to $1,000 more than the plenum fan, but the discharge plenum required for the
plenum fan is likely to cost considerably more than $1,000. This figure also does not
include motor and variable speed drive costs. In this case study, the plenum fan requires a
60 HP motor, while the other fan types only require 50 HP motors. The motor and VSD
be more expensive for the plenum fan, along with the associated electrical service.

FIGURE 72:
Cook Budget Price
C ASE ST UDY A - CO O K
BUDGET PRICES
$18,000 $16,200 $16,500
$16,000
$13,200
$14,000
$12,000 $9,700
$10,000 $8,700 $8,800
$8,000
$6,000
$4,000
$2,000
$0
60" 60" BI 66" BI 49" AF 44.5" AF 54"
Plenum Mixed
Flow

Case Study B

This case study is based on an actual installation: Site #1, an office building in San Jose,
California. The air handler consists of two 66 in. plenum fans in parallel with a design
condition for each of 72,500 CFM at 4 in. for a total air handler design condition of
145,000 CFM at 4 in. In this case study we have the benefit of hindsight, in the form of
about one years worth of air handler load profile data (CFM, SP). This data allowed us to
evaluate the actual selection and compare it to other plenum fan sizes and to several sizes of
housed airfoil fans.

We believe the project engineer selected plenum fans based on space constraints and
acoustical concerns in the building, and the fact that redundancy was necessary. Looking at
Figure 73, it is clear that the 73 in. fan has the highest efficiency, lowest noise, and highest
cost. However, based on the fan curve for this size, this point is probably too close to the
surge line (see Figure 74). The 66 in. fan that was selected by the project engineer has lower
efficiency and higher noise, but the design point is farther away from the surge line (see
Figure 75). Another advantage of the 66 in. size is that it requires a smaller motor size than
the smaller fan sizes. Be aware, however, that the fan brake horsepower does not include the
belts, which are likely to be about 97% efficient at this size. A fan BHP of 74.3 would have

FAN TYPE, SIZE AND CONTROL X139X

 a load on the motor of about 76.5 BHP. Of course, most motors and variable speed drives
have a “service factor” allowing them to operate at least 10% above their nominal capacity.
The 66 in. plenum fan’s good combination of efficiency, relative cost, acoustics, and motor
size were undoubtedly the reasons why the engineer for this project selected this fan.

FIGURE 73:
CASE STUDY B -
SELECTION SOFTWARE
AIRFOIL AND PLENUM
FANS

FIGURE 74:
CASE STUDY B - 73 IN.
PLENUM FAN CURVE

X140X FAN TYPE, SIZE AND CONTROL

 FIGURE 75:
CASE STUDY B – 66 IN.
PLENUM FAN CURVE

We evaluated this fan selection by simulating a range of potential selections against the
monitored fan load profile (total CFM and differential pressure across the fan). Figure 76
indicates the monitored data and the design point that the project engineer selected. Figure
77 shows that the system spends the majority of the time at very low flows and never comes
close to the design condition during the monitoring period. Figure 77 has the same X-axis
scale as Figure 76, and together they display the frequency of operation for each region. As
Figure 76 shows, the actual system curve appears to run through 1.5 in. at 0 CFM. A
consequence of a high fixed static pressure setpoint is that the fan operates in the surge
region at low loads. Based on the monitored data, we calculated that the fan(s) operate in
the surge region over 60% of the time. Using a smaller fan would have reduced the time in
surge. But a better way to reduce or eliminate this problem is to aggressively reset the static
pressure setpoint.

Site 1 Monitored CFM vs DP from 12/20/01 to 9/16/02

FIGURE 76:
(with perfect system curve and likely actual system curve) CASE STUDY B -
4
Actual Data M O N I TO R E D DATA
3.5 Perfect System Curve (i.e. SP reset)
Differential Pressure Across Fan (" H2O)

Design Point (145,000 CFM at 4.0")

3 Likely Actual System Curve (No SP Reset)

2.5

1.5

0.5

0
0 20,000 40,000 60,000 80,000 100,000 120,000 140,000
Total CFM (SF-1 + SF-2)

FAN TYPE, SIZE AND CONTROL X141X

 FIGURE 77:
Histogram of Site 1 Total CFM
CASE STUDY B -
7000
HISTOGRAM OF CFM The system spends many hours at
night between 20,000 and 25,000
6000 CFM

Records (15 minute data)

5000

4000

3000

2000

1000

0
0 20,000 40,000 60,000 80,000 100,000 120,000 140,000
Airflow Bin (CFM)

Figure 78 shows the efficiency of the Site 1 fan system (66 in. plenum fan) along the
apparent system curve that appears in Figure 76 as a dashed purple line. It also includes the
next smaller plenum fan size (60 in.). Figure 78 shows that the fan efficiency goes up and
down as CFM changes and as the system stages from single to dual fan operation. According
to our simulations, the average fan efficiency of the actual system during the monitoring
period was 57%.

FIGURE 78: Fan Efficiency vs CFM along System Curve Through 0/1.5" and 145,000/4"
CASE ST UDY B – PART 75%

LOAD FAN E F F I C I ENC Y 70%

65%
60%
55%
50%
Staging points from 1 to 2 fans
Fan Efficiency

45% operating (This is the optimal staging

40% point for this fan system on this
660 CPL-A system curve)
35%
600 CPL-A
30%
25%
20%
15%
10%
5%
0%
0 20,000 40,000 60,000 CFM80,000 100,000 120,000 140,000

Several other fan selections were simulated against the actual measured load, including other
sizes of plenum-airfoil fans and several sizes of housed-airfoil fans. Figure 79 shows

X142X FAN TYPE, SIZE AND CONTROL

simulation results for the base case and alternate fan selections. (Based on the monitored
data, the existing control sequence seems to be to run one fan almost all the time except for a
few hours per week, which turns out to be close to the optimal staging sequence because the
loads are so low relative to available fan capacity.)

It is interesting to note that the annual energy ranking from the simulation (Figure 79) does
not follow the efficiency ranking from the manufacturer’s selection program (Figure 73).
Several reasons exist for this discrepancy. One reason has to do with the valleys and peaks
(or “sweet spots”) in the efficiency profile of each fan (see for example Figure 78) compared
to the load profile. Different fan systems have peaks and valleys at different spots.

Figure 79 also reveals that housed-airfoil fans (the fans marked CADWDI) are consistently
more efficient than the plenum fans (the fans marked CPL-A). Of course, this is not
necessarily a fair comparison because of the space requirements and acoustic issues with
housed fans as previously noted.

Comparison of Fan Systems FIGURE 79:

Using Site 1 Monitored CFM and DP
CASE STUDY B
S I M U L AT I O N R E S U LT S -
Basecase (660CPLA, actual staging) $20,447.22
490CPLA-OptimalStaging $20,951.27 N O S TAT I C P R E S S U R E
540CPLA-OptimalStaging $19,599.46 RESET
600CPLA-OptimalStaging $19,597.08
660CPLA-OptimalStaging $20,133.37
730CPLA-OptimalStaging $19,905.23
402CADWDI-OptimalStaging $17,426.46
445CADWDI-OptimalStaging $17,191.26
490CADWDI-OptimalStaging $17,309.60
540CADWDI-OptimalStaging $17,125.13
600CADWDI-OptimalStaging $16,917.61
660CADWDI-OptimalStaging $17,142.94
730CADWDI-OptimalStaging $18,063.19
730CADWDI-Single Fan $18,010.99
$0.00 $5,000.00 $10,000.00 $15,000.00 $20,000.00 $25,000.00

Annualized Electricity Cost ($/yr)

The impact of static pressure setpoint reset on both the annual energy use and the fan
selection was also evaluated. To simulate reset, a new load profile was developed by
replacing the monitored pressure with the pressure from the system curve in Figure 76
(perfect reset line) for the same airflow as the monitored data. These reset data were used to
compare the performance of the same fans evaluated in Figure 79. The results are presented
in Figure 80. It shows that annual fan energy use can potentially be cut by as much as 50%
if static pressure setpoint reset is successfully implemented (Compare Figure 79 and Figure
80). This corroborates the results reported by Hartman (Hartman 1993) and others, as well
as the results of Case Study A.

FAN TYPE, SIZE AND CONTROL X143X

 Figure 80 also shows that annual energy ranking now follows the efficiency ranking shown in
Figure 73, because a fan operating on a perfect system reset curve has constant efficiency.
This is also one of the reasons that static pressure setpoint reset saves so much energy – not
only is the fan doing less work (maintaining lower static pressures), but it is doing it at
higher efficiency and staying out of surge longer. A perfect system curve with reset starting at
a point to the right of surge will never end up in the surge region.

The results in Figure 80 imply that bigger fans are better (in terms of energy cost) for systems
with supply pressure setpoint reset. Indeed the estimated $385 in annual energy savings
from selecting the 73 in. plenum fan rather than the 66 in. plenum fan pays for the $1,200
incremental cost increase (see Figure 73) with a simple payback of about 3 years. However
these results need to be tempered with special considerations. In addition to the first cost of
the fan, other first costs should be considered, including the impacts on space and the
electrical service. These results should also be weighed against the increased risk that the fan
will operate in surge should perfect reset not occur. (The most common cause of less-than-
perfect reset is a zone or zones that are undersized, have lower then design temperature
setpoints, or have consistently high loads, all of which can result in steady high demand for
static pressure, even when the rest of the system is at low load.) The bigger the fan, the
closer the design point is to the surge region and the greater the risk of operating in surge for
a less than perfect reset curve.

Figure 80 also shows that a single 73 in. airfoil fan can serve the load more efficiently than
almost any other option evaluated. A single housed airfoil fan is also likely to be less
expensive than any of the parallel fan options (no backdraft damper either) but of course,
redundancy is lost.

FIGURE 80: Comparison of Fan Systems

Using Site 1 Monitored CFM and Mapped to System Curve Through 0"
CASE STUDY B (i.e. Perfect Reset)
S I M U L AT I O N R E S U LT S -
490CPLA-OptimalStaging $11,950.55
P E R F E C T S TAT I C
540CPLA-OptimalStaging $10,964.61
PRESSURE RESET 600CPLA-OptimalStaging $9,251.96
660CPLA-OptimalStaging $8,517.89
730CPLA-OptimalStaging $8,086.56
402CADWDI-OptimalStaging $9,580.73
445CADWDI-OptimalStaging $8,474.97
490CADWDI-OptimalStaging $7,884.34
540CADWDI-OptimalStaging $7,502.37
600CADWDI-OptimalStaging $7,224.85
660CADWDI-OptimalStaging $6,851.64
730CADWDI-OptimalStaging $6,466.79
730CADWDI-Single Fan $6,554.18
$0.00 $5,000.00 $10,000.00 $15,000.00 $20,000.00 $25,000.00

Annualized Electricity Cost ($/yr)

X144X FAN TYPE, SIZE AND CONTROL

Space Constraints

As mentioned earlier, Figure 79 and Figure 80 are not really “apples-to-apples” comparisons
because the housed airfoil fans are likely to have higher total pressure drop due to discharge
system effects. In order to answer the question “How much extra pressure drop would make
the housed fan no longer worth using?” we simulated the housed airfoil fans with an
additional pressure drop at each fan discharge. We compared the 66 in. plenum and the 54
in. housed airfoil, assuming no static pressure setpoint reset. The breakpoint was 1.25 in.
extra inches of pressure drop. In other words, an airfoil fan with a design condition of
72,500 and 5.25 in. is just as efficient on a life-cycle cost basis as a plenum fan with a design
condition of 72,500 and 4 in.

Noise

Clearly noise was a major concern when the engineer selected the fans for Site 1. Not only
were plenum fans chosen, but sound traps were also employed. Figure 81 shows that sound
traps were inserted into the discharge plenum at each riser take-off.

FIGURE 81:
PLAN VIEW OF SITE 1
AIR HANDLER

Another acoustical advantage that plenum fans have over housed fans is that they are much
more amenable to sound traps. A sound trap can be placed relatively close to a plenum fan
because the velocity is fairly low and uniform in the discharge plenum. A sound trap cannot
be placed too close to a housed fan because of the uneven velocity profile at the fan
discharge. A sound trap in a large office building in San Francisco was placed too close to
the fan (shown schematically in Figure 82). In that building, the sound trap was selected for
0.25 in. pressure drop at the design airflow rate, but the actual pressure drop was measured
at 1.2 in.. In extreme cases such as this, a sound trap can actually increase the sound level
because the fan has to speed up to overcome the extra pressure drop.

FAN TYPE, SIZE AND CONTROL X145X

 FIGURE 82:
VELOCITY PROFILE OFF
OF HOUSED FAN

Comparing Manufacturers

We have compared fan performance from several manufacturers for a variety of fan types and
none of them stand out as consistently more-or-less efficient from one manufacturer to
another. Clearly there are some differences, but we suspect that the significant similarities
are due in large part to how the fans are tested and rated, not necessarily from true
differences in efficiency. And as mentioned in an earlier section, some obvious inaccuracy
exists with the manufacturers rating tests. Figure 83, for example, shows that the Temtrol 27
in. plenum fan is less efficient than the 24 in. and the 30 in. models. We suspect that this
may have more to do with the accuracy of the testing than with the true efficiency of the
fans.

FIGURE 83: PF01-24

Temtrol Fans
PF01-27
TEMTROL PLENUM FAN PF01-30
80.0%
D ATA PF01-33
75.0% PF02-30
PF02-33
70.0%
AF01-20
65.0% AF01-22
AF01-22-72
60.0%
AF01-24
Fan Efficiency

55.0%

50.0%

45.0%

40.0%

35.0%

30.0%

25.0%
13 14 15 16 17 18 19 20
Gamma = -ln(system curve coefficient)

This issue is further complicated by the fact that large parts of the manufacturers’ reported
fan data are extrapolated from actual factory test data. Data is calculated using the
assumption of fixed efficiency along a fan characteristic system curve. Data is also
extrapolated between fan sizes within a model line using other perfect fan laws. Under
ANSI/ASHRAE Standard 51-1999 (ANSI/AMCA Standard 210-99), manufacturers are not
required to test all fan sizes. According to the standard, test information on a single fan may

X146X FAN TYPE, SIZE AND CONTROL

be used to determine the performance of larger fans that are geometrically similar using the
so-called “fan laws,” which have many simplifying assumptions.

Figure 52 clearly reveals, for example, that Cook only tested three of their 17 mixed flow fan
sizes and then extrapolated that data to the other sizes.

Figure 84 shows the highest efficiency for all Cook and Greenheck housed airfoil fans as a
function of wheel diameter. By reviewing the step changes in the peak efficiency data as a
function of fan diameter, it is clear from this figure which fans the manufacturers tested and
which they extrapolated (see also Figure 49 and Figure 50). Both manufacturers tested their
30 in. fans. Cook then extrapolated the 30 in. data all the way up to 73 in. (The variability
in the peak efficiency of the Cook 30 in. to 73 in. fans is due to rounding and sampling
error.) Greenheck only extrapolated the 30 in. up to 36 in., then they tested the 40 in. and
extrapolated that all the way to 73 in. Cook’s 30 in. is more efficient than the Greenheck 30
in. but not more efficient than the Greenheck 40in. Had Cook tested a 40 in. (or larger)
fan, they might have found that it had higher efficiency than equally sized Greenheck fans.

FIGURE 84:
Housed Airfoil Fans: Peak Efficiency vs Diameter
PEAK EFFICIENCY OF
80%
CO O K VS G R E E N H EC K
79%
78% HOUSED AIRFOIL FANS
77%
76%
75%
74%
73%
Peak Efficiency

72%
71%
70%
69%
Cook Airfoils
68%
67% Greenheck Airfoils
66% Both
65% Poly. (Both)
64%
3 2
63% y = 1E-06x - 0.0002x + 0.0135x + 0.5078
2
62% R = 0.9112

61%
60%
0 10 20 30 40 50 60 70 80
Diameter (inches)

Fan Control

Fan Speed Control

By far the most common and most efficient way of controlling medium to large VAV fans is
with variable speed drives (VSDs). Riding the fan curve, discharge dampers, inlet vanes, and
variable pitch blades were all common in the past but are rarely a good option given the

FAN TYPE, SIZE AND CONTROL X147X

 relatively low cost and energy savings from VSDs. The 2005 version of Title 24 requires
fans of 10 HP and larger to have either variable speed drives or variable pitch blades for vane-
axial fans.

The location of the static pressure sensor can greatly affect the energy efficiency potential of a
system when a fixed static pressure setpoint is used. An old rule of thumb was to locate the
sensor “2/3rd of the way down the duct,” but this approach wastes energy and is not
recommended. Instead, the sensor should be as far out in the system as possible, with
multiple sensors used if there are branches in the duct main. The design condition SP
setpoint should be the minimum SP necessary to get the air from the sensor location through
the ductwork to the hydraulically most remote VAV box, through its discharge ductwork
and air outlets, and into the space. The further the sensor is located from the fan, the lower
the SP setpoint needs to be, and vice versa. The worst case is to locate the sensor at the fan
outlet. The setpoint would have to be high enough to deliver supply air to the most remote
space at the maximum airflow that will occur at design conditions. This setpoint would
cause the fan to operate against a constant discharge pressure and nearly constant total
pressure. The energy usage of the fan would then be linear with airflow while the fan would
use close to the cube of the airflow ratio if the sensor were located near the extreme end of
the system.

If the static pressure setpoint is reset (see Demand-Based Static Pressure Reset), the location
of the sensor theoretically makes no difference since its setpoint will always be only as high as
needed for the box requiring the highest pressure. However, it is recommended that it be
located as far out into the system as practical to ensure proper operation if the reset logic fails
(any setpoint reset logic must be well tuned to provide stable performance.)

Practical considerations include: limiting the number of sensors to as few as possible, usually
one; and locating the sensor upstream of fire/smoke dampers (FSD) or isolation zone
damper. For example, in a high-rise building with a central air handler on the roof that uses
FSDs for off-hour floor isolation, the SP sensor should be located at the bottom of the riser
(e.g., just before the ground floor damper). If the sensor were downstream of an isolation or
FSD damper, the system will not function properly when that damper is closed but other
parts of the building are in operation.

Figure 85 shows the energy impact of the minimum static pressure setpoint on total fan
system energy (fan, belts, motor, and VSD). At 50% flow, the fan on the 1.5 in. system
curve uses about twice as much energy as the fan on the 0 in. curve. At 20% flow, the 1.5 in.
fan uses about four times as much energy as the 0 in. fan.

X148X FAN TYPE, SIZE AND CONTROL

 FIGURE 85:
Part Load Performance versus Minimum Static Pressure Setpoint
Plenum airfoil fans on system curves running through several min SP points SP SETPOINT VS FAN
100% SYSTEM ENERGY

90%
PL-0"
80%

70% PL-0.5
Percent Power

60% PL-1.0
50%
'PL-1.5
40%

30%

20%

10%

0%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percent Airflow

As part of this research, we have developed DOE-2 fan curves for each of the curves shown
in Figure 85, as well as several other curves representing other fan types and minimum SP
setpoints. These curves appear in Appendix 5 – DOE-2 Fan Curves.

It is also important that the SP sensor input and the variable speed drive output speed signal
be located on the same DDC control panel. This control loop is too critical and to be
subject to the variations in network traffic and other vagaries of the building-wide DDC
control system.

Demand-Based Static Pressure Reset

As illustrated in the case studies above, demand based static pressure setpoint reset has
tremendous potential for saving energy and reducing noise, as well as reducing or eliminating
fan operation in surge.

Demand-based static pressure setpoint reset can only be effectively implemented on a system
with zone-level DDC controls and some signal from the VAV box controllers back to the
DDC system indicating VAV box damper position. This signal may be either the damper
signal if modulating actuators are used, or estimates of damper position based on timing
open/close signals if floating actuators are used. A full-open position switch on the actuator
may also be used, although with more zones not satisfied. There are at least two possible
control sequences that can be used to implement demand-based static pressure reset: “Trim
and Respond” and “PID Control”. Based on our experience with both of these sequences,
the authors recommend “Trim and Respond”. Sample language for both sequences is

FAN TYPE, SIZE AND CONTROL X149X

 provided below. Another method of supply pressure reset which is not discussed here is load
demand based control (Hartman, 2003).

Trim and Respond (Recommended)

Static pressure setpoint shall be reset using trim and respond logic within the range 0.2
inches to 1.5 inches for VAV AHUs and 0.1 inches to 0.75 inches for underfloor AHUs.
When fan is off, freeze setpoint at the minimum value (0.2 inches). While fan is proven on,
every 2 minutes, decrease the setpoint by 0.04 inches if there are no pressure requests. If
there are more than two (adjustable) pressure requests, increase the setpoint by 0.04. Where
VAV zone damper position is known, a pressure request is generated when any VAV or
underfloor damper served by the system is wide open. Where VAV zone damper position is
unknown, a pressure request is made when the ratio of the zone’s actual supply airflow to
supply airflow setpoint is less than 90%. All values adjustable from fan graphic. The control
logic shall be slow-acting to avoid hunting.

Variations on this sequence include:

ƒ A damper position (valve position, loop output, etc.) of 80% counts as one request,
while a position of 95% counts as two requests.
ƒ The response rate is a function of the number of requests, e.g. response = 0.04 in. times
the number of requests.

PID Control (Not Recommended)

1. SP setpoint shall be determined within the range of 0.5 in. to MaxP by a
direct-acting control loop whose control point is the damper position of
the most open VAV damper and whose setpoint is 90% open. In other
words, the static setpoint will be reset to maintain the VAV box requiring
the most static pressure at 90% open.
2. MaxP shall be determined by the air balancer in conjunction with the
control contractor as required to provide design airflow in all boxes
downstream of the duct static pressure sensor. (See Determining Static
Pressure Setpoint.)
3. Supply fan speed is controlled to maintain duct static pressure at setpoint
when the fan is proven on. Minimum speed is 10% for motor cooling.
Where the isolation areas served by the system are small, provide multiple
sets of PI gains that are used in the PI loop as a function of a load
indicator (e.g., supply fan airflow rate, the area of the isolation areas that
are occupied, etc.).

Why Use Trim and Respond

One advantage of Trim and Respond over PID Control is that it is easier to “ignore” or
“starve” rogue zones. The number of “ignores” (i.e. the number or pressure requests that

X150X FAN TYPE, SIZE AND CONTROL

must be received before the pressure is increased) is easily adjusted. In any VAV system,
particularly ones with very large numbers of zones, it is almost a given that there will be at
least a couple rogue zones. If these rogue zones are ignored, then all the other zones (and
often even the rogue zones) can be satisfied at fairly low pressure setpoints.

Field experience has also shown that the PID Control method is more difficult to tune
properly. The reset control loop must be very slow relative to zone airflow control loops
because a change in static pressure has an immediate effect on VAV airflow and hence
damper position. This control sequence only seems to work well if the PID loop is very slow
and almost entirely integral (low proportional gain).

Another advantage of Trim and Respond control is that it can have different response rates
for “trim” and for “respond”. Generally, the trim rate should be very slow but the respond
rate may need to be very quick depending on the consequences of temporarily underserving a
zone. In general, Trim and Respond is more flexible and easier to tune than PID and should
be used for all demand-based reset sequences (static pressure reset, supply air temperature
reset, chilled water temperature and pressure reset, etc.)

Rogue Zones

Static pressure reset, like all demand-based reset sequences, relies on reasonably good
agreement between equipment sizing and actual loads. If one particular box or branch duct
is significantly undersized, that box may always be wide open and the zone undercooled, in
which case no static pressure setpoint reset is possible. One possible solution in this situation
is to exclude that box from the logic used to determine the SP setpoint. That approach may
suffice if the zone is a storage room, but if it contains the boss’s office then a better solution
is to replace the box. A single “rogue zone” or undersized box on a large VAV system could
result in thousands of dollars of lost energy savings on an annual basis. One way to avoid the
rogue zone problem is to oversize questionable zones, especially cooling-only zones such as
server rooms. In general, of course, oversizing should be avoided because it leads to excessive
reheat, but a cooling-only zone with a zero minimum flow can be oversized because there is
no reheat penalty. A good example is a computer server room served by a VAV box. The
room does not need minimum ventilation or heating so an oversized cooling-only box with
zero minimum is appropriate. The room will operate continuously, even when most other
zones are unoccupied making it particularly important for this type of zone not to require a
high SP when the overall system is at low flow.

Another source of rogue zones is if one zone has a significantly different space temperature
setpoint. For example, a recreation center has a packaged VAV unit that serves an aerobics
studio and many other spaces (lounges, offices, locker rooms, etc.). The aerobic studio
occupants want the space to be 60-65 degrees so that zone is almost always calling for colder
air and limiting the SAT reset. It is important therefore to consider likely thermostat
setpoints when designing a VAV system. In the recreation center case it might have made
more sense to put the aerobics studio on a separate packaged single zone system. Another

FAN TYPE, SIZE AND CONTROL X151X

 example is that if a zone is unable to maintain the desired cooling setpoint then the
occupants may incorrectly assume that cranking the thermostat down to 60oF will help the
system achieve setpoint. All they are really accomplishing is creating a rogue zone. To avoid
this sort of problem it is generally a good idea to limit the range within which occupants can
adjust setpoints.

Tools for Identifying Rogue Zones

Almost any zone can be driving the reset sequence at some time but if there are rogue zones
that are consistently driving a reset sequence (due to undersizing, sensor error, or other
reasons) then the reset sequence will not be successful. Thus it is necessary to identify which
zones are consistently demanding more duct pressure, colder air, etc. If there are many zones
in a reset sequence (e.g. one air handler can serve hundreds of VAV boxes) then finding the
culprit(s) can be very challenging. Here are some methods that commissioning agents and
facilities staff can use to identify rogue zones:

Trend Review

The control system should be programmed to collect trends on the potential reset
drivers such as damper position for all VAV boxes. This trend data can then be
analyzed using tools such as the Universal Translator (www.utonline.org)

Custom Reports

All major DDC Control systems can be configured to generate custom reports.
Here is a sample specification for such a report:

A. Setpoint Reset History. For setpoints that are reset by zone or system demand
(such as supply air temperature setpoint, differential pressure setpoint), create
the following report from trend data with user-selectable report historical
period by date and time.

1. At the top of the report, show system name and description, name
of setpoint, and setpoint average, minimum, and maximum value
over the report range when the reset is active.

2. The body of the report shall list each zone/system driving the reset
including zone/system names, hours driving the reset, reset lock-
out status, and average value of variable driving reset (such as
damper or valve position) when the reset is active.

X152X FAN TYPE, SIZE AND CONTROL

Highlights on Summary Screens and/or Floor Plans

Zones that are currently driving the reset (requesting more pressure, colder air, etc.)
should be highlighted on graphics screens such as summary screens and/or floor
plans. Figure 86 is a screen capture of a VAV box summary graphic showing which
zones are requesting pressure (P-Req) and colder air (C-Req). Note that currently
all zones shown on this screen are satisfied.

FIGURE 86:
GRAPHIC SCREEN
SHOWING RESET
DRIVERS

Display of Current Reset Drivers

Most control systems can be set up to display the zone name or ID of the zone(s)
that are currently driving the reset sequence. Figure 87 is an example of a graphic
screen that displays the ID of the zone with highest “need more air” signal and the
zone with the highest cooling signal. Unfortunately if there is more than one zone
with a cooling loop output of 100 it will only display the first zone that reached
100. Additional programming would be required to get it to display the ID of all
zones with cooling loop outputs of 100.

FAN TYPE, SIZE AND CONTROL X153X

 FIGURE 87:
GRAPHICS SCREEN
SHOWING ZONE ID FOR
RESET DRIVERS

This indicates that VAV box

20 has the highest cooling
loop output

Field Experience with Static Pressure, Supply Air Temperature Reset and Other
Demand-Based Reset Sequences

Appendix 14 describes several buildings where trend data was collected to see how
successfully supply air temperature setpoint reset, static pressure setpoint reset and demand
controlled ventilation sequences had been implemented. Trend data was collected on
approximately 15 air handling systems. Figure 88 summarizes the static pressure reset results
presented in Appendix 14. As this figure shows the average pressure setpoint was
approximately halfway between the maximum and minimum setpoints, which means that
these systems are achieving considerable fan energy savings compared to a fixed setpoint
scheme.

X154X FAN TYPE, SIZE AND CONTROL

 Max, Min, Avg Pressure Setpoint (in. w.g.)
FIGURE 88:
0 0.5 1 1.5 2 2.5
S U M M A R Y O F S TAT I C
San Ramon Tech PRESSURE RESET CASE
Center / Main AC1
STUDIES

San Ramon Tech

Center / Main AC2

Sacramento
Courthouse / (Average
of 32 AHUs)

Sacramento
Courthouse / AH-7A1

Sacramento
Courthouse / AH-7A2

UC Merced Library /
AHU-1

UC Merced Library /
AHU-2

UC Merced Library /
AHU-3

UC Merced Library /
AHU-4

UC Merced Classroom
Bldg / AHU-1 Cold Deck

Morgan Hill Rec Center

AC-1

Santa Clara Service

Center Building #3

Case study savings for supply air temperature reset were not as consistent as for static
pressure reset. One reason is that most of the pressure reset case studies used a trim and
respond sequence while most of the temperature reset case studies used PID control. Several
of the temperature reset case studies had one or two rogue zones that inhibited the reset.
There were however, a couple successful temperature reset case studies. Figure 89 illustrates
one such example. The gray lines in this figure illustrate the reset sequence, i.e. setpoint is
fixed at 53oF when outside air temperature is at 60oF and is reset between 55 and 63 when
outside air temperature is below 55oF. The actual setpoint is consistently pegged at the
maximum temperature that the reset sequence will allow.

FAN TYPE, SIZE AND CONTROL X155X

 FIGURE 89: 64

Supply Air Temperature Setpoint (oF)

S A M P L E S U P P LY A I R
62
T E M P E R AT U R E R E S E T
T R E N D D ATA 60

58

56

54

52
52 54 56 58 60 62 64

Outdoor Air Temperature (oF)

Determining Static Pressure Setpoint

Even if DDC is available at the zone level and reset controls are to be used, the design static
pressure setpoint must be determined in the field in conjunction with the air balancer. The
setpoint determined below is the fixed static pressure setpoint for systems without reset and
it is MaxP when reset is used as described above.
1. Set all boxes downstream of the static pressure sensor to operate at
maximum airflow setpoints.
2. Set all boxes upstream of the static pressure sensor to full shut-off (zero
flow).
3. Manually lower fan speed slowly while observing VAV box airflow rates
downstream of the static pressure sensor. Stop lowering speed when one
or more VAV box airflow rates just drops 10% below maximum airflow
rate setpoint.
4. Once flow condition in previous step is achieved, note the DDC system
static pressure reading at the duct static pressure sensor. This reading
becomes the static pressure setpoint:
5. If there are multiple static pressure sensors, repeat steps above for each
sensor. Each should have its own static pressure setpoint and control loop
with the fan speed based on the largest loop output.

X156X FAN TYPE, SIZE AND CONTROL

Fan Staging

Multiple fans in parallel are typically staged based on fan speed signal, with some deadband
to prevent short cycling. All operating fans must be controlled to the same speed. The
optimal speed for staging up (e.g. from one to two fans) and staging down (e.g. from two to
one fan) is a function of the actual system curve, which of course is a function of the SP
setpoint and static pressure setpoint reset.

Figure 90 and Figure 91 present optimal staging speeds for two-fan system in the Case Study
B with and without supply pressure reset. In these figures, the solid lines represent the power
consumed by the fan systems (fan, belts, motor, VSD) as they run up and down the system
curve. The light dashed lines (read on the secondary y-axis) represent the speed for these
fans at each condition. The heavy dashed lines show the speed at the optimal staging point.
For the 66 plenum fan system and for a system curve that runs through 1.5 in. (i.e., fixed
static pressure setpoint), the optimal point to stage from one fan to two fans occurs when the
fan exceeds approximately 79% speed. Conversely, the optimal point to stage from two to
one fan is when the speed drops below about 63%.

Optimal Staging from 1 to 2 Fans for System Curve Through 1.5" (No Reset) FIGURE 90:
Based on Site 1 Fan System (660 CPL-A, 75HP, 635 design RPM)
OPTIMAL STAGING (NO
1.5" Curve - 1 Fan (KW)
1.5" Curve - 2 Fans (KW) S TAT I C P R E S S U R E
1.5" Curve - 1 Fan (speed) 100%
140 1.5" Curve - 2 Fans (speed) RESET)
90%

120 80%

70%
100
Fan System KW

60%
Pct Speed

80
50%

60 40%

30%
40
20%
20
10%

0 0%
0

10,000

20,000

30,000

40,000

50,000

60,000

70,000

80,000

90,000

100,000

110,000

120,000

130,000

140,000

CFM

FAN TYPE, SIZE AND CONTROL X157X

 FIGURE 91:
Optimal Staging from 1 to 2 Fans Given Perfect Reset
OPTIMAL STAGING Based on Site 1 Fan System (660 CPL-A, 75HP, 635 design RPM)
10% Min VSD Speed
( P E R F E C T S TAT I C
100%
PRESSURE RESET) 140 0" Curve - 1 Fan (KW)
0" Curve - 2 Fans (KW) 90%
120 0" Curve - 1 Fan (speed) 80%
0" Curve - 2 Fans (speed)
100 70%

Fan System KW
60%

Pct Speed
80
50%
60 40%

40 30%

20%
20
10%

0 0%
0

10,000

20,000

30,000

40,000

50,000

60,000

70,000

80,000

90,000

100,000

110,000

120,000

130,000

140,000
CFM

Figure 92 shows the optimal staging points from Figure 90 (1.5 in.), Figure 91 (0 in.), as
well as three intermediate points. While it may not be possible to know exactly what the
minimum duct static pressure setpoint will be, a designer can use something like Figure 91
and his/her best guess of the min SP setpoint when writing the initial control sequence.
That guess can then be refined in the field using in order to fine-tune the optimal staging
control sequence.

Figure 93 builds on the information in Figure 92 by including optimal staging for a 54 in.
airfoil fan.

FIGURE 92:
Optimal Fan Staging Point versus Mimimum Duct Static Pressure Setpoint
OPTIMAL STAGING (for 660 CPL-A fan system)

POINT VS. MINIMUM 78

From 1 to 2 fans
D U C T S TAT I C P R E S S U R E From 2 to 1 fans
73
SETPOINT
68

63
Percent Fan Speed

58

53

48

43

38

33

28
0 0.5 1 1.5
Duct Static Pressure Setpoint

X158X FAN TYPE, SIZE AND CONTROL

 FIGURE 93:
Optimal Fan Staging Point versus Mimimum Duct Static Pressure Setpoint
(for two fan types) OPTIMAL STAGING
From 1 to 2 fans - 66" Plenum POINT FOR TWO FAN
78
From 2 to 1 fans - 66" Plenum
From 1 to 2 fans - 54" AF TYPES
73
From 2 to 1 fans - 54" AF
68

63
Percent Fan Speed

58

53

48

43

38

33

28
0 0.5 1 1.5
Duct Static Pressure Setpoint

It may be tempting, after seeing how little difference there is between the kW lines in Figure
90 and Figure 91 at low loads, to simply operate two fans during all hours of operation.
Besides the obvious waste of energy that would result, a few other problems exist with this
strategy. One problem is operation in surge. Figure 94 shows that at low flow and relatively
high fixed static, both fans are likely to operate in surge if the flow is divided between two
fans, but if the load is carried by only one fan, then it is less likely to be operating in surge.

Site #1 SF-1 CFM vs SP from 5/02 through 7/02 FIGURE 94:

SF-1 ALONE PA R A L L E L F A N S I N
3 SF-1 (while SF-2 is running)
Surge Line SURGE
Ideal System Curve @ 145,000 cfm, 4"
2.5
Static Pressure (in H2O)

1.5

0.5

0
0 10,000 20,000 30,000 40,000 50,000 60,000 70,000
Airflow (cfm)

Another problem with operating parallel fans at low flows and high fixed static is
“paralleling.” This phenomenon occurs with fan types that have flat spots or dips in the fan

FAN TYPE, SIZE AND CONTROL X159X

 curves in the surge region, such as plenum fans, forward curved fans, or mixed flow fans.
For example, as Figure 95 and Figure 96 illustrate, a 66 in. plenum fan with a speed signal of
680 RPM operating against 5.5 in. of static pressure can produce as little as 45,000 CFM, as
much as 56,000 CFM, or any point in between, which can lead to unstable operation. If
two fans in parallel are operating in a flat spot on the curve, they can flip-flop back and
forth, resulting in further instability.

FIGURE 95:
" PA R A L L E L I N G " - H I G H
FLOW

FIGURE 96:
" PA R A L L E L I N G " - LO W
FLOW

How To Isolate Fans in Parallel

Fans in parallel must be isolated, either with inlet cones, barometric backdraft dampers, or
motorized backdraft dampers 28. Any type of isolation will add static pressure to the fan

28
Excessive leakage through a fan that is OFF not only causes the ON fan to run harder but it will cause the OFF
fan to spin backwards which can cause serious problems when the OFF fan is turned ON. If it is a single phase

X160X FAN TYPE, SIZE AND CONTROL

system, although to widely varying degrees, and will have some leakage rate that will also
increase fan energy. Motorized backdraft dampers also add controls complexity and should
generally be avoided.

Plenum fans are best isolated with inlet cones. These have very low-pressure drop when fully
open and do not leak as much as backdraft dampers. Backdraft dampers also impart flow
turbulence to the fan inlet that can reduce fan performance (system effect). Backdraft
dampers can also add significantly to system pressure drop.

Conclusions

Static Pressure Reset

ƒ How a fan is controlled is probably more important to fan energy than the type of fan
selected. More specifically, demand-based static pressure setpoint reset has the potential
to:

o Reduce fan energy up to 50%.

o Reduce fan operation in surge and thereby reduce noise,

vibration, and bearing wear.

o Improve control stability.

ƒ If demand-based static pressure setpoint reset is not feasible or possible (e.g., no DDC
controls at zone level), then the sensor should be located as far out in the system as
possible and the SP setpoint should be as low as possible.

Fan Type
ƒ Housed airfoil fans are usually more efficient than plenum fans, even if space constraints
result in a poor discharge arrangement and system effects. The extra pressure drop on a
housed airfoil fan has to be surprisingly high before it is less efficient than a plenum fan
for the same application. However, noise and space constraints may still result in
plenum fans as being the best choice.
ƒ An airfoil fan selected near its peak efficiency will stay out of surge longer than a plenum
fan selected near its peak efficiency, because housed airfoil fans peak well to the right of
the surge line but plenum fans peak right at the surge line.
ƒ In order to estimate annual fan energy, it is necessary to consider part load performance
and how the fan is likely to be controlled.

motor, the fan will start and spin in the wrong direction; it will move air in the right direction but very
inefficiently. If it is a three phase fan with a VFD, it will probably lock the motor and then start spinning in the
right direction, but it could trip the drive. If it is a direct drive fan with a three phase motor, an across-the-line
start can shear the fan shaft and shatter the fan wheel and fan housing.

FAN TYPE, SIZE AND CONTROL X161X

 o Fan efficiency is fairly constant at part load if static pressure
setpoint reset is successfully implemented. If not, part load fan
efficiency depends on the fan type and size. Manufacturers’ data
can be combined with estimated system curves to develop part
load fan performance curves.

o Some estimate of the annual fan load profile is necessary to

estimate annual fan energy (e.g., DOE-2 simulation).

Fan Sizing
ƒ If there is a good chance that static pressure setpoint reset will be successfully
implemented, fan sizing is fairly straightforward since fan efficiency remains fairly
constant. If static pressure setpoint reset is not likely to be implemented, consider using
a smaller fan (i.e., lower efficiency at design condition) because it will stay out of surge
longer and the efficiency will actually improve as it rides down the system curve.

First Cost
ƒ A fair comparison of fan types for built up systems should include the cost to construct
the discharge plenum for plenum fans.
ƒ Motor and VSD costs should also be considered since less efficient fans may require
larger motors and drives.

Noise
ƒ Not only are plenum fans inherently quieter than housed fans due to the attenuation of
the discharge plenum, but they also work better with sound traps. A sound trap can be
placed much closer to a plenum fan than to a housed fan.
ƒ Parallel plenum fans can be fitted with inlet cones for very low pressure drop backdraft
protection. This option is not generally available on housed fans, which must rely on
backdraft dampers which have higher pressure drop but lower cost.
ƒ Housed airfoil and other types of housed fans should not be ruled out on the basis of
noise. Locate the air handler as far away from noise-sensitive spaces as possible. Use
duct liner to attenuate noise. Use a sound trap, if necessary, but only if it can be located
at least three duct diameters downstream of the fan.

Fan Staging
ƒ While it may not be possible to know exactly what the minimum duct static pressure
setpoint will be, a designer can use Figure 93 and his/her best guess of the min SP
setpoint when writing the initial control sequence for staging parallel fans. That guess
can then be refined based on monitored data in order to fine tune the optimal staging
control sequence.
ƒ Operating fans in parallel at low flow should be avoided, particularly if SP is not
successfully reset. By dividing the flow in half, it pushes the fans into the surge region
and can cause them to operate in particularly unstable areas within the surge region.

X162X FAN TYPE, SIZE AND CONTROL

 COILS AND FILTERS

Construction Filters

If air handlers must be used during construction, filtration media with a Minimum
Efficiency Reporting Value (MERV) of 6, as determined by ASHRAE 52.2-1999 should be
used to protect coils and supply systems. Replace all filtration media immediately prior to
occupancy.

Pre-Filters

Aside from pressure drop and added maintenance costs, pre-filters add little to a system.
They are typically not effective in extending the life of the main filters as most dust passes
through them. This is particularly true if final filters are changed frequently as is
recommended below. Prefilters increase energy costs and labor costs (they generally have
minor dust-loading capability and must be changed each quarter) and thus should be
avoided.

Final Filter Selection

A reasonable selection for typical commercial applications is 80 percent to 85 percent dust

spot efficiency (ASHRAE 52.1), MERV 12 (ASHRAE 52.2). Maximum initial pressure
drop at 500 feet per minute should not exceed 0.60 inches water column. When selecting
the fan, the mean air pressure drop (midway from clean to maximum) should be used.
Filters should be changed long before they reach the maximum pressure drop as indicated by
the filter manufacturer. More frequent change intervals (e.g. once per year) are now being
recommended by IAQ experts based on recent studies that have shown that occupants
perceive lower air quality as filters become dirty over relatively short time periods. The cause
is likely from volatile organic compounds emitted from microbial growth on the dirt
collected in the filter.

COILS AND FILTERS X163X

 Filter Area

Filter banks in large built up air handlers as well as in custom or modular air handlers are
sometimes installed with a blank-off panel to make up the difference between the filter bank
area and the air handling unit casing area. If the entire cross sectional area of the air handler
is filled with filters then pressure drop will be reduced and filter life will be extended. The
energy and maintenance savings can pay for the added first cost in a reasonably short
payback period.

Extended Surface Area Filters

Extended surface area filters are a new class of filters that have higher dust-holding capacity,
longer life, and lower pressure drops. They are designed to fit conventional filter framing.
While extended surface area filters cost more than standard filters they too may pay for
themselves in energy and maintenance savings.

Monitoring Filters

Monitor pressure drop across filters via the DDC system so that an alarm can be triggered if
filter pressure drop becomes excessive. Magnehelic gauges, or digital gauge now available on
DDC differential pressure sensors, are also commonly used for visual indication of filter
pressure drop.

The alarm in the DDC system on VAV systems should vary with fan speed (or inlet guide
vane (IGV) signal) roughly as follows:

DPx = DP100 ( x )
1.4

where DP100 is the high limit pressure drop at design cfm and DPx is the high limit at speed
(IGV) signal x (expressed as a fraction of full signal). For instance, the setpoint at 50% of
full speed would be (.5)1.4 or 38% of the design high limit pressure drop.

While filters will provide adequate filtration up to their design pressure drop, odors can
become a problem well before a filter reaches its design pressure drop. For this reason and
for simplicity of maintenance, filters are typically replaced on a regular schedule (e.g. every
12 or 18 months).

X164X COILS AND FILTERS

 Coil Selection

Many designers select cooling coils for a face velocity of 550 fpm. However, it is well worth
looking at lower face velocity coil selection ranging from 400 fpm to 550 fpm and selecting
the largest coil that can reasonably fit in the allocated space. Table 22 shows a range of coil
selections for each of the five monitored sites. The design selections in this table are shown
with yellow highlights. The blue highlights indicate flat blade coils, and the rest of the
selections are wavy fin coils.

COILS AND FILTERS X165X

Site Coil Dimensions AIR SIDE TABLE 22. ALTERNATE
Capacity (kBH) Pressure drop WATER SIDE COIL SELECTIONS FOR
Height Length Total Sensible (in.) GPM DP (ft.)
ALL FIVE MONITORED
(in.) (in.) Area FPM (Face FPI (fins (Flow) (Pressure
(ft2) velocity) ROWS per in.) drop) SITES
#1 36 120 30.0 537 6 10 466 426 0.74 68 16
42 120 35.0 460 6 10 471 428 0.31 68 12.6
42 120 35.0 460 6 10 468 429 0.55 68 10.8
48 120 40.0 403 6 10 482 436 0.25 68 10.3
48 120 40.0 403 6 10 491 443 0.47 68 10.3
#2 42.25 146 42.9 544 5 12 781 670 0.91 133 12.4
48.5 146 49.0 476 6 9 796 679 0.72 133 8.3
57.5 146 58.3 400 5 10 790 671 0.5 133 6.4
#3 49.5 74 25.4 472 6 10 403 358 0.79 50 2.3
57.5 74 29.5 406 6 10 409 362 0.61 50 1.4
#4 39.25 140 38.2 524 4 10 492 410 0.6 81 2
42.25 140 41.1 486 4 10 496 413 0.54 81 1.9
51.5 140 50.1 399 4 10 500 418 0.4 81 0.9
#5 36 108 27.0 555 3 14 440 419 0.73 63.6 25.7
42 108 31.5 476 6 10 483 445 0.33 63.6 10.3
42 108 31.5 476 5 11 468 436 0.54 63.6 8.9
51 108 38.3 392 5 10 477 438 0.2 63.6 18.2
51 108 38.3 392 5 9 472 435 0.34 63.6 18.2

COILS AND FILTERS X166X

For Site #1, increasing the coil bank height from 36 in. to 42 in. reduced the airside pressure
drop from 0.74 in. w.c. to 0.31 in. w.c. and lowered the annual fan energy around 3% to
5% (3% savings with fixed static pressure setpoint and 5% savings with demand based static
pressure reset). This coil selection also reduced the waterside pressure drop by 3.4 ft, which
may or may not have an impact on pumping energy depending on the piping configuration.
In addition to reducing the fan energy, the drop in static pressure can also reduce fan noise.

Most air handler manufacturers offer multiple coil sizes for a given air handler casing size.
Selecting the largest coil for a particular casing can have a significant impact on fan energy
and a minimal impact on first cost. However, for a supplier in a competitive situation, it can
be the difference between winning or losing a job. Therefore, the designer needs to be
specific enough in the construction documents to force the larger coil selection.

It is important to read the messages from the manufacturer’s selection program when
selecting a coil. They will provide warnings if the velocity and fin design pose any risk of
condensate carryover.

ASHRAE Standard 62.1 recommends selecting coils for under 0.75 in. w.c. for ease of coil
cleaning, since both the number of rows and the fin spacing contribute to the difficulty of
accessing the fins.

Coils should also be selected for true counter flow arrangements. At low loads, interlaced
and multi-inlet coils can lead to a drop in the low differential coil temperatures. This so-
called “Degrading ΔT syndrome” causes central plants to run inefficiently due to increased
pumping and inefficient staging of chillers.

Coils should be selected for the same ΔT as the chilled water plant but all coils in a chilled
water system do not have to be selected for the exact same ΔT as long as the weighted
average ΔT matches the plant.

All coils should have access doors upstream and, for larger coils (>2 rows), downstream as
well. This allows the coils to be cleaned and inspected, both of which are critical for
performance and IAQ. It also ensures that control sensors can be located appropriately. For
example, the freezestat on a 100% OA system with a preheat coil must be located between
the preheat coil and the cooling coil.

Coil Bypass

For coil banks in large built-up VAV systems, consider placing a bypass damper between coil
sections where the intermediate coil headers are located. Since this space is already allocated
for piping, it provides a low-cost option to further reduce the fan pressure drop. The bypass
will open except when the cooling coil is active. Airfoil damper blades (rather than vee-
groove blades) should be used for velocities over 1500 fpm.

COILS AND FILTERS X167X

 OUTSIDE AIR/RETURN AIR/EXHAUST AIR CONTROL

This section describes the design of airside economizers, building pressurization controls, and
control for code-required ventilation in a VAV system.

Control of Minimum Outdoor Air for VAV Systems

Ventilation that meets Title 24 minimums is required for all spaces when they are normally
occupied (§121 (c) 1.). Furthermore, the 2005 version of the Standard mandates that VAV
systems be tested for code-required ventilation both at design supply airflow and with all
VAV boxes at minimum position. Although providing code-minimum ventilation
throughout the range of system operation is implied by the existing standard, systems are
rarely designed to achieve this, so this section provides guidance on designing VAV systems
to dynamically adjust outdoor airflow.

This section presents several methods used to dynamically control the minimum outdoor air
in VAV systems, which are summarized in Table 23 and described in detail below.

Fixed Minimum Damper Position

Figure 97 depicts a typical VAV system. In standard practice, the TAB contractor sets the
minimum position setting for the outdoor air damper during construction. It is set under
the conditions of design airflow for the system, and remains in the same position throughout
the full range of system operation.

Does this meet code? The answer is no. As the system airflow drops so will the pressure in
the mixed air plenum. A fixed position on the minimum outdoor air damper will produce a
varying outdoor airflow. As depicted in Figure 97, this effect will be approximately linear (in
other words outdoor air airflow will drop directly in proportion to the supply airflow).

X168X OUTSIDE AIR/RETURN AIR/EXHAUST AIR CONTROL

 FIGURE 97:
VAV R E H E AT S Y S T E M
WITH A FIXED MINIMUM
OUTDOOR AIR DAMPER
SETPOINT

100%
Percent Outdoor Air

50%

0%
0% 50% 100%
Percent Supply Air
TABLE 23:
SUMMARY OF MINIMUM
Method Figure Description
Fixed minimum Figure 97 This method does not comply with Title 24; the airflow at a OUTDOOR AIR CONTROL
damper setpoint fixed minimum damper position will vary with the pressure STRATEGIES
in the mixed air plenum.
Dual minimum − This method complies with the letter of Title 24 but is not
damper setpoint at accurate over the entire range of airflow rates and when there
maximum and are wind or stack effect pressure fluctuations.
minimum supply air
rates
Energy balance Figure 98 This method does not work for two reasons: 1) inherent
method inaccuracy of the mixed air temperature sensor, and 2) the
denominator of the calculation amplifies sensor inaccuracy as
the return air temperature approaches the outdoor air
temperature.
Return fan tracking Figure 99 This approach does not work because the cumulative error of
the two airflow measurements can be large, particularly at
low supply/return airflow rates.
Airflow measurement Figure 100 This method may or may not work depending on the airflow
of the entire outdoor measurement technology. Most airflow sensors will not be
air inlet accurate to a 5%-15% turndown (the normal commercial
ventilation range).
Injection fan with Figure 101 This approach works, but is expensive and may require
dedicated minimum additional space.
ventilation damper
Dedicated minimum Figure 102 This successful approach is the recommended method of
ventilation damper control.
with pressure control

OUTSIDE AIR/RETURN AIR/EXHAUST AIR CONTROL X169X

 Dual Minimum Damper Position

An inexpensive enhancement to the fixed damper setpoint design is the dual minimum
setpoint design, commonly used on some packaged AC units. The minimum damper
position is set proportionally based on fan speed or airflow between a setpoint determined
when the fan is at full speed (or airflow) and minimum speed (or airflow). This method
complies with the letter of Title 24 but is not accurate over the entire range of airflow rates
and when there are wind or stack effect pressure fluctuations. But with DDC, this design
has very low costs.

Energy Balance

The energy balance method (Figure 98) uses temperature sensors in the outside, as well as
return and mixed air plenums to determine the percentage of outdoor air in the supply air
stream. The outdoor airflow is then calculated using the equations shown in Figure 98.
This method requires an airflow monitoring station on the supply fan.

This approach does not generally work for several reasons:

1. The accuracy of the mixed air temperature sensor is critical to the
calculation but is very difficult to perform with any precision in real
applications. Even with an averaging type bulb, most mixing plenums
have some stratification or horizontal separation between the outside and
mixed airstreams. 29
2. Even with the best installation, high accuracy sensors, and field calibration
of the sensors, the equation for percent outdoor air will become inaccurate
as the return air temperature approaches the outdoor air temperature.
When they are equal, this equation predicts an infinite percentage outdoor
air.
3. The accuracy of the airflow monitoring station at low supply airflows is
likely to be low.

29
This was the subject of ASHRAE Research Project 1045-RP, “Verifying Mixed Air Damper Temperature and Air
Mixing Characteristics.” Unless the return is over the outdoor air there are significant problems with
stratification or airstream separation in mixing plenums.

X170X OUTSIDE AIR/RETURN AIR/EXHAUST AIR CONTROL

 FIGURE 98:
ENERGY BALANCE
METHOD OF
Outdoor MAT
CO N T R O L L I N G MI N I M U M
OAT
Air Intake OUTDOOR AIR

RAT

Return Air

DDC

%OA = (MAT - RAT) / (OAT - RAT)

CFM-OA = %OA * CFM-SA

Return Fan Tracking

Return fan tracking (Figure 99) uses airflow monitoring stations on both the supply and
return fans. The theory behind this is that the difference between the supply and return fans
has to be made up by outdoor air, and controlling the flow of return air forces more
ventilation into the building. Several problems occur with this method: 1) the relative
accuracy of airflow monitoring stations is poor, particularly at low airflows; 2) the cost of
airflow monitoring stations; 3) it will cause building pressurization problems unless the
ventilation air is equal to the desired building exfiltration plus the building exhaust.
ASHRAE research has also demonstrated that in some cases this arrangement can cause
outdoor air to be drawn into the system through the exhaust dampers due to negative
pressures at the return fan discharge.

OUTSIDE AIR/RETURN AIR/EXHAUST AIR CONTROL X171X

 FIGURE 99:
RETURN FAN TRACKING
Outdoor Air
Intake

Exhaust Air
Return Air

Signal from
SAT Controller

Outside Air Flow Monitoring Station

Controlling the outdoor air damper by direct measurement with an airflow monitoring
station (Figure 100) can be an unreliable method. Its success relies on the turndown
accuracy of the airflow monitoring station. Depending on the loads in a building, the
ventilation airflow can be between 5% and 15% of the design airflow. If the outdoor airflow
sensor is sized for the design flow for the airside economizer, this method has to have an
airflow monitoring station that can turn down to the minimum ventilation flow (between
5% and 15%). Of the different types available, only a hot-wire anemometer array is likely to
have this low-flow accuracy while traditional pitot arrays will not. (Refer to Section 3.5.3 of
the PECI Control System Design Guide for a comparison of air flow measurement
technologies.) One advantage of this approach is that it provides outdoor airflow readings
under all operating conditions, not just when on minimum outdoor air.

X172X OUTSIDE AIR/RETURN AIR/EXHAUST AIR CONTROL

 Air Flow Measuring FIGURE 100:

Station AIRFLOW ME AS URE MENT

OF 100% OUTDOOR AIR

Outdoor
Air Intake

Return Air

Signal From SAT Controller

Injection Fan

The injection fan method (Figure 101) uses a separate outdoor air inlet and fan sized for the
minimum ventilation airflow. This inlet contains an airflow monitoring station, and a fan
with capacity control (e.g. discharge damper; VFD) which is modulated as required to
achieve the desired ventilation rate. The discharge damper is recommended since a damper
must be provided anyway to shut off the intake when the AHU is off, and also to prevent
excess outdoor air intake when the mixed air plenum is very negative under peak conditions.
(The fan is operating against a negative differential pressure and thus cannot stop flow just
by slowing or stopping the fan.) This method works, but the cost is high and often requires
additional space for the injection fan assembly.

OUTSIDE AIR/RETURN AIR/EXHAUST AIR CONTROL X173X

 FIGURE 101:
Injection Fan w/
INJECTION FAN WITH
Discharge Damper
D E D I C AT E D M I N I M U M
OUTDOOR AIR DAMPER

Outdoor Air
Intake

Return Air

Signal from SAT Controller

Separate Minimum Ventilation Damper

An inexpensive but effective design uses a minimum ventilation damper with differential
pressure control (Figure 102). In this method, the economizer damper is broken into two
pieces: a small two position damper controlled for minimum ventilation air and a larger,
modulating, maximum outdoor air damper that is used in economizer mode. A differential
pressure transducer is placed across the economizer damper section measuring the pressure in
the mixing plenum with the outside as a reference. During start-up, the air balancer opens
the minimum OA damper and return air damper, closes the economizer OA damper, runs
the supply fan at design airflow, measures the OA airflow (using a hand-held velometer) and
adjusts the minimum OA damper position until the OA airflow equals the design minimum
OA airflow. The linkages on the minimum OA damper are then adjusted so that the current
position is the “full open” actuator position. At this point the DP across the minimum OA
damper is measured. This value becomes the DP setpoint. The principle used here is that
airflow is constant across a fixed orifice (the open damper) at fixed DP.

As the supply fan modulates when the economizer is off, the return air damper is controlled
to maintain the design pressure DP setpoint across the minimum ventilation damper. (Refer
to ASHRAE Guideline 16 for damper type and sizing in this scheme.)

X174X OUTSIDE AIR/RETURN AIR/EXHAUST AIR CONTROL

The main downside to this method is the complexity of controls. A control sequence for this
scheme follows:

Minimum outdoor air control

Open minimum outdoor air damper when the supply air fan is proven on and the system is
not in warm-up, cool-down, setup, or setback mode. Damper shall be closed otherwise.

The minimum differential pressure setpoint across the mixed air plenum (MinDP) is
determined by the air balancer as required to maintain the design minimum outdoor airflow
rate across the minimum outdoor air damper with the supply air fan at design airflow. See
below for return air damper control of mixed air plenum pressure.

Return air dampers

When the economizer is locked out from the economizer high limit control (see Economizer
High-Limit Switches), the return air damper signal is modulated to maintain differential
pressure across the outdoor air damper at setpoint (MinDPsp) determined above.

When the economizer is in control, the return air damper is sequenced with the outdoor air
economizer damper as described in the section, Economizer Temperature Control.

FIGURE 102:
MINIMUM OUTDOOR AIR
DAMPER WITH
PR ESS U R E CO N T R O L

Outdoor Air
Intake

Return Air

Signal from SAT Controller

OUTSIDE AIR/RETURN AIR/EXHAUST AIR CONTROL X175X

 Isolation Zones

Regardless of how the minimum ventilation is controlled, care should be taken to reduce the
amount of outdoor air provided when the system is operating during the weekend or after
hours with only a fraction of the zones active. Title 24, section 122(g) requires provision of
“isolation zones” of 25,000 ft2 or less. This can be provided by having the VAV boxes return
to fully closed when their associated zone is in unoccupied mode. When a space or group of
spaces is returned to occupied mode (e.g. through off-hour scheduling or a janitor’s override)
only the boxes serving those zones need to be active. During this partial occupancy the
ventilation air can be reduced to the requirements of those zones that are active. If all zones
are of the same occupancy type (e.g. private offices), simply assign a floor area to each
isolation zone and prorate the minimum ventilation area by the ratio of the sum of the floor
areas presently active divided by the sum of all the floor areas served by the HVAC system.

For our recommended control scheme with a separate minimum outdoor air damper, this
same area ratio can be used to reduce the design pressure drop setpoint MinDPsp across the
economizer section from the design setpoint MinDP:

2
⎡A ⎤
MinDPsp = MinDP⎢ active ⎥
⎣ Atotal ⎦

where Aactive is area of active Isolation Areas and Atotal is the overall floor area served by the
system. The Contractor shall calculate the floor area of Isolation Areas from drawings.

Design of Airside Economizer Systems

Title 24 has a prescriptive requirement for economizers on all air-conditioning systems with
cooling capacities greater than 6.5 tons. Although waterside economizers can be used to
meet this requirement, airside economizers are generally more cost effective and always more
energy efficient in California climates. For built-up VAV systems, an exception to this rule is
floor-by-floor air-handling units served by a central ventilation shaft where insufficient space
exists to provide 100% outdoor air for the units. In this case, either water-cooled units or
chilled water units with a water-side economizer is generally a better solution. Water-side
economizers may also be more effective for areas requiring high humidity levels (>30%) since
the increase in humidifier energy can offset the cooling savings.

This section deals with design, configuration, and control of airside economizer systems.
The ASHRAE Guideline 16-2003 “Selecting Outdoor, Return, and Relief Dampers for
Airside Economizer Systems,” available at http://www.ashrae.org, contains practical and
detailed information on damper selection and guidance on control of economizer dampers.
This guideline purposely does not cover many of the topics addressed by Guideline 16 (e.g.

X176X OUTSIDE AIR/RETURN AIR/EXHAUST AIR CONTROL

recommended damper configuration and sizing). Readers are encouraged to purchase a copy
from ASHRAE.

Configuration of dampers for adequate mixing of outside and return air streams is the
subject of the ASHRAE Research Project 1045-RP, “Verifying Mixed Air Damper
Temperature and Air Mixing Characteristics.” This study found somewhat improved
mixing when the return air was provided on the roof of the mixing plenum over the outdoor
air rather than side-by-side or opposite wall configurations. There were no strong trends or
generalizations observed among design options such as damper blade length, blade
orientation, and face velocity. Fortunately, in most mild California climates, mixing
effectiveness is not a significant issue.

Common to all airside economizer systems is the need to relieve up to 100% design airflow
minus anticipated exfiltration and building exhaust, due to the fact that the economizer
could be providing up to 100% outdoor air. Exfiltration to maintain a mild pressurization
(between 0.03 in. to 0.08 in. above ambient) in a typical commercial building can be
assumed to be approximately 0.05 to 0.15 cfm/ft2.

Economizers can be designed with barometric relief, relief fan(s), or return fan(s) (Figure
103, Figure 104 and Figure 105). The choice of system return/relief path configuration is
usually driven by a number of design issues including physical space constraints, the pressure
drop in the return path, the need for interspatial pressurization control, acoustics and others.
From an energy standpoint, the choices in order of preference (from most efficient to least
efficient) are as follows: barometric relief (Figure 103), relief fans (Figure 104) and return
fans (Figure 105). Each of these options are described below.

While always the most efficient choice, barometric relief (Figure 103) may not be the most
cost effective choice. To work effectively barometric dampers must be chosen for low-
pressure drop (typically a maximum of 0.08 in.w.c. from the space to ambient) at relatively
high flow rates. As a result, the barometric relief openings can be excessively large -- a
challenge to the architectural design. Where barometric relief is used, the relief may be
provided anywhere within the areas served by the central system.

In addition to energy savings, another advantage of barometric relief is the simplicity of

controls for building pressurization, since no automatic control is required. A distinct
disadvantage is that it only works for low-pressure returns, typically limiting it to low-rise
projects.

OUTSIDE AIR/RETURN AIR/EXHAUST AIR CONTROL X177X

 FIGURE 103:
AIRSIDE ECONOMIZER
CO N F I G U R AT I O N W I T H
BAROMETRIC RELIEF
FROM ASHRAE
GUIDELINE 16-2003

Where barometric relief is not an option, relief fans (Figure 104) are the best bet. Relief fans
always use less energy than return fans and can incorporate barometric relief as the first stage
of building pressure control (see sequence below). In addition to the energy benefits, relief
fans are relatively compact, reducing impact on space planning and architectural design. The
two largest limitations are acoustics and static pressure. Acoustical control can usually be
achieved by placing the relief fans out of the line of site from the return shaft. Systems with
high return pressures (e.g., ducted returns) will generally require return fans.

The following is an example control sequence for a system with two relief fans and an
automated damper at each:
ƒ Relief system shall only be enabled when the associated supply fan is proven on and the
minimum outdoor air damper is open.
ƒ Building static pressure shall be time averaged with a sliding five-minute window (to
reduce damper and fan control fluctuations). The averaged value shall be that displayed
and used for control 30.

30
A single building static pressure sensor is usually sufficient, or one per wing or tower for large, irregularly shaped
buildings. The high side should be in an interior space on the second floor (first floor is too variable due to lobby
doors). Do not tap into a single tube in multiple locations in order to get an average signal. The pressure
differences between the various taps creates a flow in the tube and a false reading.

X178X OUTSIDE AIR/RETURN AIR/EXHAUST AIR CONTROL

ƒ A PI loop maintains the building pressure at a setpoint of 0.05 in. with an output
ranging from 0% to 100%. The loop is disabled and output set to zero when the relief
system is disabled. When the relief system is enabled, open the motorized dampers to
both relief fans (this provides barometric relief for the building). When the PI loop is
above 25%, start one relief fan (lead fan) and assign the fan %-speed analog output to
the PI loop output; and close the discharge damper of the adjacent relief fan (to prevent
backflow). Lead fan shall shut off when PI loop falls below 15% for five minutes (do
not limit speed signal to the motor – operating below 15% speed for 5 minutes should
31
not overheat motor ). Start lag fan and open its discharge damper when PI output rises
above 50%. Stop lag fan and shut its damper after fan has operated for at least 5
minutes and PI loop output falls below 40%. Fan speed signal to all operating fans shall
be the same.

Note that this sequence first opens the relief dampers before staging the fans on, which saves
considerable energy since at low loads, barometric relief is all that is required.

FIGURE 104:
AIRSIDE ECONOMIZER
CO N F I G U R AT I O N W I T H
RELIEF FAN FROM
ASHRAE GUIDELINE
16-2003

Return fans (Figure 105) should only be used for projects with high static pressure
requirements (e.g., ducted returns or the need for sound traps). They will always use more
energy than relief fans, will generally cost more to install, and will add to the complexity of
the control system.

31
Minimum motor speed limitations to ensure proper motor cooling have not been well studied. ABB suggests a
minimum of 10% (6 Hz) for pump and fan applications where power drops nearly as the cube of airflow. Other
manufacturers suggest there is no minimum speed for these applications provided it is acceptable that motor
surface temperatures become hot enough to cause burns if touched. Still others suggest minimum speeds as high
as 20 Hz, particularly for TEFC motors commonly used for outdoor applications. Our own experience is that
10% (6 Hz) provides adequate cooling for long term operation and there is no minimum speed for short term
operation.

OUTSIDE AIR/RETURN AIR/EXHAUST AIR CONTROL X179X

 A sample sequence of control for return fans follows. In this sequence, the return fan speed
is modulated to control the pressure in the return/relief air plenum and the exhaust/relief
damper is controlled to maintain building static pressure. This sequence ensures that
adequate return airflow can be provided when the economizer is off, and that the system can
provide 100% outdoor air and maintain the desired building static pressure when the
economizer is on.

Example return fan sequence:

ƒ Return fan operates whenever associated supply fan is proven on.
ƒ Return fan speed shall be controlled to maintain return fan discharge static pressure at
setpoint. The setpoint shall be determined in conjunction with the air balancer as the
larger of the following:
ƒ That required to deliver the design return air volume across the return air damper when
the supply air fan is at design airflow and on minimum outdoor air.
ƒ That required to exhaust enough air to maintain space pressure at setpoint (0.05 in.)
when the supply air fan is at design airflow and on 100% outdoor air.

Relief/exhaust dampers shall only be enabled when the associated supply and return fan are
proven on and the minimum outdoor air damper is open. The relief/exhaust dampers shall
be closed when disabled.

Building static pressure shall be time averaged with a sliding five-minute window (to reduce
damper and fan control fluctuations). The averaged value shall be that displayed and used
for control.

When the relief/exhaust dampers are enabled, they shall be controlled by a PI loop that
maintains the building pressure at a setpoint of 0.05 in.. (Due the potential for interaction
between the building pressurization and return fan control loops, extra care must be taken in
selecting the PI gains. ASHRAE Guideline 16-2003 recommends that the closed loop
response time of the building pressurization loop should not exceed one-fifth the closed loop
response time of the return fan control loop to prevent excessive control loop interaction.
This can be accomplished by decreasing the gain of the building pressurization controller.)

X180X OUTSIDE AIR/RETURN AIR/EXHAUST AIR CONTROL

 FIGURE 105:
AIRSIDE ECONOMIZER
CO N F I G U R AT I O N W I T H
RETURN FAN FROM
ASHRAE GUIDELINE
16-2003

Economizer Temperature Control

Most economizer control sequences stage the outdoor and return dampers in tandem, with
the return dampers closing as the outdoor dampers open. Although this sequence works, fan
energy savings can be achieved by staging these dampers in series (see Figure 106). In this
staged sequence, the outdoor air damper is opened as the first stage of cooling, while the
return damper remains open (provided that the economizer is operating). This sequence
provides less than 100% outdoor air but a very low-pressure air path for the supply fan. If
this is sufficient to cool the building, energy savings will result from the reduced fan pressure.
If more cooling is needed, the return damper is modulated closed to ensure that the system
has 100% outdoor air. ASHRAE Guideline 16 recommends this sequence.

100%
FIGURE 106:
Damper/Valve Position, % Open

Return AIRSIDE ECONOMIZER

Air Damper CO N T R O L STAG I N G
Outdoor Air Damper
FROM ASHRAE
GUIDELINE 16-2003

CHW Valve
Minimum Position
(MinOA)
0%
Supply Air Temperature Control
Loop Output Signal

OUTSIDE AIR/RETURN AIR/EXHAUST AIR CONTROL X181X

 Economizer High-Limit Switches

Title 24 has requirements for economizer high-limit switches. The high-limit switch is the
control that disables the economizer when the outdoor air is warmer (or has higher enthalpy)
than the return air. This requirement was based on a detailed study on the energy
performance of high-limit switches done by ASHRAE’s Standard 90.1 committee in
development of the 1999 Standard.

Table 24 presents the requirements by climate zone from Title 24. This table has five
different high-limit controls (identified as devices) including fixed and differential dry-bulb
temperature, fixed and differential enthalpy, and electronic enthalpy. Fixed dry-bulb and
enthalpy controls use a fixed reference for return air temperature rather than a direct
measurement. Differential controls provide a measurement both outside and in the return
air stream.

TABLE 24: Required High Limit

(Economizer Off When):
HIGH LIMIT SWITCH
Device Type Climate Zones
REQUIREMENTS FROM Equation Description
Fixed Dry Bulb 01, 02, 03, 05, 11, TOA > 75°F Outside air temperature
TITLE 24
13, 14, 15, & 16 TOA > 70°F exceeds 75° F
04, 06, 07, 08, 09, Outside air temperature
10 &12 exceeds 70°
Differential Dry Bulb All TOA >TRA Outside air temperature
exceeds return air
temperature
Fixed Enthalpyc 04, 06, 07, 08, 09, hOA > 28 Btu/lbb Outside air enthalpy
10 &12 exceeds 28 Btu/lb of dry
airb
Electronic Enthalpy All (TOA, RHOA) > A Outside air
temperature/RH exceeds
the “A” set-point curvea
Differential Enthalpy All hOA > hRA > Outside air enthalpy
exceeds return air
enthalpy
a Set point “A” corresponds to a curve on the psychometric chart that goes through a point at
approximately 75°F and 40% relative
b At altitudes substantially different than sea level, the Fixed Enthalpy limit value shall be set to
the enthalpy value at 75°F and 50%
c Fixed Enthalpy Controls are prohibited in climate zones 01, 02, 03, 05, 11, 13, 14, 15 & 16

The electronic enthalpy device measured is a Honeywell controller that is used in packaged
equipment. As shown in Figure 107, it acts like a dry bulb controller at low humidity and
an enthalpy controller at high humidity. This device is only available as a fixed reference and
offers four switch selectable reference curves for the return.

X182X OUTSIDE AIR/RETURN AIR/EXHAUST AIR CONTROL

 Control Control Curve 85 90 95 100 105 110
Curve (approx. *F) FIGURE 107:

46
at 50% RH
A 73 ELECTRONIC ENTHALPY

44
B 70

42
HIGH LIMIT

0.90
C 67

0.80
40
D 63

0.70
CO N T R O L L E R

R)

38

0
AI

0. 6
36

0
0.5
RY

34
(D

32
b

0
30
u/l

0.4
Bt

Y
28

IT
ID
PY

26

M
AL

24

0. HU
TH

Lines are

30
2 2
EN

VE
20

roughly

TI
LA
18

RE parallel to
16

0
0.2
14

enthalpy lines
12

for humid air,

0
0.1
drybulb lines
D for dry air
A BC

35 40 45 50 55 60 65 70 75 80 85 90 95 100 105
o
DRY BULB TEMPERATURE (approximate) F

Of all of the options, dry bulb temperature controls prove the most robust as dry-bulb
temperature sensors are easy to calibrate and do not drift excessively over time. Differential
control is recommended throughout California and the sensors should be selected for a
through system resolution of 0.5°F. Dry-bulb sensors work well in all but humid climates,
which are not typical in California.

Differential enthalpy controls are theoretically the most energy efficient. The problem with
them is that the sensors are very hard to keep calibrated and should be recalibrated on an
annual or semi-annual basis. Contrary to common perception, enthalpy controls do not
work in all climates. In hot dry climates they can hunt and excessively cycle the economizer
dampers when the hot dry outdoor air has lower enthalpy than the space(s) at cooling
balance point. What happens is that the economizer opens up and the coil is dry, which in
turn dries out the space(s) until the return enthalpy goes below the outdoor enthalpy. As a
result, the economizer damper closes, the space humidity increases, and the cycle repeats.

OUTSIDE AIR/RETURN AIR/EXHAUST AIR CONTROL X183X

 APPENDIX 1 – MONITORING SITES

TABLE 25: Site 1 Site 2 Site 3 Site 4 Site 5

Type Office Office Office Courthouse Office
SUMMARY OF
Floor 105,000 307,000 955,000 570,000 390,000
MONITORING SITE Area (ft2)
CHARACTERISTICS Stories 3 12 25 16 6/8
Location San Jose, San Jose, Sacramento, Sacramento, Oakland,
CA CA CA CA CA
Owner Private Private Public Public Public
Occupant Owner- Tenant Public Public Public
occupied Tenant Tenant Tenant
Fan Type Two Four Centrifugal, Centrifugal, Six
Centrifugal Centrifugal, Plenum, Plenum, Centrifugal,
Plenum Housed Two per Two per Housed
Floor Floor
Cooling 500 800 2,300 1,700 1,000
Total
Tons

Site 1

ƒ Occupancy type: Office, owner occupied, with data center.

ƒ Location: San Jose, California.
ƒ Floor area: 105,000 ft2.
ƒ Number of stories: Three.
ƒ Occupancy date: October 1999.
ƒ Built-up air handling unit with two 66 in. plenum fans with airfoil blades (Cook 660-
CPLA) with barometric back draft dampers on the inlets. Each fan has a 75 HP motor
and was designed for approximately 70,000 cfm at 4 in. wc. These fans are operated
24/7 to serve several computer rooms.
ƒ There are six relief fans with 5 HP motors controlled in two groups of three, with a
variable speed drive for each group.
ƒ Two hot-water unit heaters (containing hot water coils and fans) are located in the
mixed air plenum to provide preheat. No additional heating coils are located in the air-
handler.
ƒ Chilled water plant: Two centrifugal, VSD, water-cooled 250 ton chillers. Model York
YTH3A2C1-CJH.
ƒ Primary/secondary chilled water pumping arrangement: Two 7.5 HP primary pumps in
series with each chiller. Two 7.5 HP secondary pumps in parallel. Variable speed drives
on secondary pumps.

X184X APPENDIX 1 – MONITORING SITES

ƒ Cooling tower: One cooling tower with a VFD. Cooling tower was designed for a 6°
approach temperature, a 9° range, and a 68°F wet-bulb temperature. There are two
cells, each with a capacity of 705 GPM at design conditions. Each cell has a 25 HP
axial fan with a VFD.
ƒ Condenser water pumps: There are six condenser water pumps, two that serve the
chillers, two that pump to the heat exchanger (on the open side), and two on the closed
loop side that serve auxiliary loads off of the condenser water riser. The two chiller
condenser water pumps are 20 HP each, constant speed. The other two open-loop side
condenser water pumps are 5 HP each, constant speed. The closed-loop side condenser
water pumps are 5 HP each, variable speed driven.
ƒ Interior zones are served by cooling-only VAV boxes. Perimeter zones are served by
VAV boxes with hot-water reheat coils.
ƒ Two natural-draft boilers provide hot water for building heating. Each boiler has an
output of 1,400,000 Btu/hr. Two 5 HP variable-speed driven pumps, in a primary only
arrangement, distribute hot water.
ƒ A condenser water loop serves process loads within the building, including computer
room AC units. This condenser water is piped through a heat exchanger. The heat
exchanger is served by the cooling tower with two constant volume flow pumps, 5 HP
each.
ƒ The building has scheduled lighting controls for the core and occupancy sensors in the
private offices.
ƒ Supply air temperature reset control is in use. The supply pressure is operated at a fixed
setpoint of 1.5in.
ƒ The building has an Automated Logic Corporation (ALC) control system, which does
not expose VAV box damper position for trending.

FIGURE 108:
60000
SITE 1, MONITORED
HVAC ELECTRIC ITY END
50000 USES

40000
Pumps
Tower Fans
kWh

30000
Chillers
Fans
20000

10000

0
Nov-01
Dec-01

Jan-02
Feb-02
Mar-02

Apr-02
May-02

Jun-02
Jul-02

Aug-02
Sep-02

Oct-02
Nov-02
Dec-02

Jan-03
Feb-03

APPENDIX 1 – MONITORING SITES X185X

 FIGURE 109:
SITE 1, MONITORED
HVAC ELECTRIC ITY END Pumps
USES 23%

Fans
47%

Tower Fans
12%

Chillers
18%

Site 2

ƒ Occupancy type: Private office, speculative building, with computer rooms, 100%
occupied.
ƒ Location: San Jose, California.
ƒ Floor area: 307,000 ft2.
ƒ Number of stories: 12.
ƒ Occupancy date: December 2000.
ƒ Number of air handlers: Two built-up air handlers located in the mechanical penthouse
that serve separate shafts. The shafts are connected on each floor via a loop duct. Each
air handler has two housed centrifugal supply fans with airfoil blades, each with 100 HP
motor. Each of the four supply fans is sized for 70,000 cfm at 5.0 in. w.c.. Each air
handler also has six propeller-type vane-axial relief fans, each with 5 HP motors. All fans
have variable speed drives. The relief fans are controlled in two gangs of three fans for
each air handling system.
ƒ Chilled water plant: Two 400-ton water-cooled centrifugal chillers rated at 0.54
kW/ton, model Trane CVHF0500AIH. Chillers have inlet vanes to control capacity.
ƒ Chilled water is distributed by two constant-speed primary pumps, 25 HP each.
ƒ Two natural-draft boilers provide hot water for building heating. Each boiler has an
output of 2,400,000 Btu/hr. Two constant speed pumps, in parallel, distribute hot
water. Each pump is 7.5 HP.
ƒ Interior zones are served by cooling-only VAV boxes. Perimeter zones are served by
VAV boxes with hot water reheat coils.

X186X APPENDIX 1 – MONITORING SITES

ƒ Two condenser water pumps each at 40 HP serve a condenser water riser that is directly
connected (i.e., no heat exchanger) to the cooling tower. The cooling tower has two
cells and two VSD fans of 30 HP each. Design flow is 2,830 gpm, with a 10°F range
and 8.6°F approach. In addition, there is an auxiliary condenser water system with a
separate cooling tower for computer room air conditioners served by two pumps of 15
HP each.
ƒ The HVAC control system is by Siemens (Apogee).
ƒ The building has lighting controls.
ƒ Supply air temperature reset control is in use. Supply air static pressure is fixed.

FIGURE 110:
RELIEF FAN (ONE OF SIX
PER PENTHOUSE)

PHOTO COURTESY OF
TAY LO R E NGINEERING

FIGURE 111:
RELIEF FAN DISCHARGE
PHOTO COURTESY OF
TAY LO R E NGINEERING

APPENDIX 1 – MONITORING SITES X187X

 Site 3
ƒ Occupancy type: Public office with computer rooms on 8th floor and a gym.
ƒ Location: Sacramento, California.
ƒ Floor area: 955,000 ft2
ƒ Number of stories: 25.
ƒ Occupancy date: October 2000.
ƒ Number of air handlers: 58 packaged VAV units with centrifugal supply and exhaust
fans, as well as variable speed drives. The supply fans are plenum type installed in a
packaged air-handling unit. The exhaust fans are tubular centrifugal. Chilled water
cooling coil (draw-through). Hot water pre-heat coil. Typical arrangement is two air
handlers per floor connected through a loop duct.
ƒ Chilled water plant: Three Carrier electric centrifugal chillers with variable speed
compressor, 300 tons at 0.55 kW/ton, 800 and 1,200 tons at 0.50 kW/ton.
Primary/secondary pumping configuration with three constant speed primary pumps
totaling 110 HP, and three variable-speed secondary pumps totaling 150 HP.
ƒ Cooling tower: Four cells, each with variable speed axial fan, total 200 HP fan motor
and 2,875 tons heat rejection capacity. Three constant speed condenser water pumps:
30 HP, 75 HP, and 100 HP.
ƒ Two natural draft gas boilers that produce heating hot water (2,400 and 3,000 MBH).
ƒ Air handlers: The 16th floor was the focus of monitoring for this research. LBNL
recorded data from both 16 and 17 floors each 36,000 ft2, with 16 as the control floor
(i.e., no sealing or adding holes to the ducts). There are two air handlers on 16th floor
located in the northwest and northeast corners. Both are connected to a common supply
air duct loop. One branch of the loop runs between the two air handlers along the north
side. The other branch runs between the two AHUs and loops around the east, south,
and west sides.
ƒ Zone VAV boxes: There are 39 VAV boxes on the 16th floor. Fourteen are parallel-fan
powered boxes with electric reheat coils that serve perimeter zones; the remaining 25
boxes are cooling-only VAV boxes serving interior zones.
ƒ The building has a Johnson Metasys control system.
ƒ The supply fans on the 16th floor are controlled to a fixed 1 in. w.c. setpoint. It is
possible to reset the supply pressure by demand.

Data collected from interior zones provide examples of the range and diversity of zone
cooling loads. At Site 3, the data show that the most typical load is in the range of 1.0 to 1.5
W/ft2 and is seldom higher than 2.0 W/ft2. Figure 112 shows the distribution for a sample of
three interior zones.

X188X APPENDIX 1 – MONITORING SITES

 50%
FIGURE 112:
45%
MO N I TO R E D CO O L I N G
Fraction of Operating Hours
40%
LOADS FO R A SAMP LE
35%
OF THREE INTERIOR
30%
ZONES, SITE 3
25% Zone 636
20% (OFFICE)
Zone 623
15% Zone 630
10%
5%
0%
0.1- 0.5- 1.0- 1.5- 2.0- 2.5- 3.0- 3.5-4 >4
0.5 1.0 1.5 2.0 2.5 3.0 3.5
Zone Cooling Load, W/ft2

Site 4
ƒ Occupancy type: Federal Courthouse.
ƒ Location: Sacramento, California.
ƒ Floor area: 570,000 ft2.
ƒ Number of stories: 16 (7th floor monitored by this research).
ƒ Occupancy date: January 1999.
ƒ Number of air handlers: 30. Typical floor has two VAV air handlers with one serving
the core areas with courtrooms and the other serving perimeter areas that include public
and office areas. The majority of the air handlers are designed for 20,000 cfm airflow
and have 20 HP supply fans (centrifugal airfoil type) and 7.5 HP return fans (tubular
centrifugal in-line type). The return system is ducted. Chilled water and hot water coils
have two-way valves and include water flow sensors and supply and return temperature
sensors that allow coil cooling/heating load calculation. Each air handler also has airflow
measurement stations on the supply and return air. Minimum outdoor airflow is
specified as 3,200 Ccfm per air-handler. The systems were designed with a CO2 sensor
on the return ductwork of the interior unit. The control sequence for this sensor was
never uncovered.
ƒ Chilled water plant: Three Trane centrifugal chillers, two 675 tons at 0.545 kW/ton
and one 350 tons at 0.535 kW/ton. The smaller chiller has variable speed compressor
controls. Primary/secondary pump configuration with three constant speed primary
pumps totaling 50 HP and three variable-speed secondary pumps totaling 125 HP.
ƒ Cooling tower: Three cells, each with variable speed axial fan, total 180 HP fan motor
and 2,500 tons heat rejection capacity. Designed to produce 80°F water at 73°F outdoor
wetbulb temperature. Three constant speed condenser water pumps: 60 HP, 60 HP,
and 30 HP.
ƒ Controls: Johnson Metasys control system with extensive monitoring. Each floor has
CHW and heating HW btuh meters both at the take-off from the risers and for each
coil on the air-handling units. The base building also has supply and return airflow for
each air handler. The system also monitors chiller cooling load and electric demand.

APPENDIX 1 – MONITORING SITES X189X

 ƒ Perimeter zones are served by VAV boxes with hot-water reheat coils. Core zones,
including courtrooms, have VAV boxes without reheat.
ƒ Hot water plant: Two forced draft boilers of 10,420 kBtu/hr capacity each.
Primary/secondary hot water pumping configuration.
ƒ Supply air pressure is fixed but reset by demand is possible.
ƒ This site has floor-by-floor air-handling systems with separate units for the interior and
perimeter. The 7th floor was chosen for monitoring because it was both representative
of the building, but logistically easy to work with. This floor houses the bankruptcy
court and hearing rooms. The two hearing rooms vary from completely empty to fully
occupied throughout the day. Three CO2 sensors were installed, one in each of the two
hearing rooms and one in the courtrooms. The CO2 sensors were wired to the closest
VAV box and trended through the DDC controls system.

Site 5
ƒ Occupancy type: Municipal office, retail, computer room.
ƒ Location: Oakland, California.
ƒ Floor area: 173,000 ft2.
ƒ Number of stories: Eight.
ƒ Occupancy date: Summer 1998.
ƒ Number of air handlers: One. A central VAV system consists of two supply and two
return fans that serve floors two through eight (floor one is retail and is served by water
source heat pumps connected to a separate condenser water system with a fluid cooler).
ƒ Chilled water plant: The office building shares a chilled water plant with the adjacent
building (which was not studied) that consists of two 500-ton water-cooled centrifugal
chillers. Chilled water is distributed through a primary/secondary pump configuration
with a separate set of secondary pumps for each building. An air-cooled chiller serves
three computer room AC units.
ƒ The building has its own hot water boiler.
ƒ Perimeter zones are served by standard VAV boxes with hot water reheat coils. Core
zones have standard VAV boxes without reheat.
ƒ The building has lighting controls.
ƒ The control system is by Staeffa.

X190X APPENDIX 1 – MONITORING SITES

 FIGURE 113:
BUILDINGS SUMMARY
(SOURCE: NAOYA
MOTEGI, LBNL)

APPENDIX 1 – MONITORING SITES X191X

 APPENDIX 2 – MEASURED FAN PERFORMANCE

Energy Benchmark Data

Energy end use analysis can provide insights into what systems and equipment provide the
most potential for energy savings. Table 26 lists office building data from several surveys and
from monitoring of specific sites. In looking at this table, it is important to note that a
number factors influence the values like climate, hours of operation, energy sources for
heating and cooling, age of the building, building occupancy, and others. This is particularly
true for databases like CEUS, NRNC, and others. Annual electricity consumption varies
from 10.0 to 21.1 kWh/ft2. Of that total, HVAC electricity accounts for 15% to 51%, and
fans use 25% to 61% of the HVAC electricity. Annual fan energy consumption ranges from
1.5 to 4.0 kWh/ft2. In the three monitored buildings included in Table 26, fans consumed
between 47% to 61% of the total HVAC electricity. The design of airside systems clearly
deserves attention not only because fans are a significant end use but also because, as these
guidelines attempt to show, significant cost effective savings are possible.

CEUS, CEUS, NRNC, Bldgs Energy Site 1, Site 2, Site 5,

TABLE 26:
1997 1999 1999 Data book, 2002 2/02 - 8/99-
OFFICE BUILDING 2002 1/03 7/00
ENERGY END USE Fans (kWh/ft2/yr) 4.0 1.5 2.4 1.5 1.8 1.8 1.6
Cooling (kWh/ft2/yr) 3.2 4.5 2.9 2.7 2.1 1.1 1.3
CONSUMPTION FROM
Heating (kWh/ft2/yr) n.a. n.a. 0.4 n.a. n.a. n.a. n.a.
SEVERAL SOURCES Lighting (kWh/ft2/yr) 4.6 3.7 4.0 8.2 n.a.* n.a.* 3.6
Misc. (kWh/ft2/yr) 2.4 3.1 5.6 5.9 17.2 16.6 3.5
Total Electricity 14.2 12.7 15.3 18.4 21.1 19.5 10.0
(kWh/ft2/yr)
Heating Gas 22.4 20.6 n.a. 24.3 31.8 82.5 18.1
(kBtu/ft2/yr)
HVAC % of Total 51% 47% 37% 23% 19% 15% 30%
Electricity
Fans % of HVAC 56% 25% 45% 36% 47% 61% 56%
Electricity
* Lighting energy not monitored separately from other misc loads at sites 1 and 2.
Sources:
CEUS 1997. Commercial End-Use Survey, Pacific Gas & Electric Company.
CEUS 1999. Commercial End-Use Survey, Pacific Gas & Electric Company.
NRNC, 1999. Nonresidential New Construction Baseline Study, prepared by RLW Analytics for
Southern California Edison.
Buildings Energy Data book, 2002. U.S. Department of Energy, Office of Energy Efficiency and
Renewable Energy.
Site 1. Commercial office building, San Jose, CA. See Appendix for details.
Site 2. Commercial office building, San Jose, CA. See Appendix for details.
Site 5. Public office building, Oakland, CA. See Appendix for details.

X192X APPENDIX 2 – MEASURED FAN PERFORMANCE

Fans also contribute a significant amount to a building’s peak electricity demand. Figure 114
shows the fan-only peak day electric demand for three monitored sites, which reaches
between 0.5 and 1.0 W/ft2. Monitoring also shows that fans account for about 15% of the
peak day demand at Site 1 and 12% at Site 2, corresponding to 0.60 W/ft2 and 0.75 W/ft2,
respectively.

FIGURE 114:
1.0 PEAK DAY FAN ELECTRIC
DEMAND, THREE SITES
0.9

0.8 Site 1, 9/4/02

0.7 Site 2, 8/9/02

Total Fan Pow er, W/ft2

Site 5, 8/9/01
0.6

0.5

0.4

0.3

0.2

0.1

0.0
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00

FIGURE 115:
PEAK DAY ELECTRIC
Other
4.5 DEMAND, SITE 1,
C Tower
9/3/2002
4.0 Chiller
( C U M U L AT I V E G R A P H ;
Pumps
3.5 TOTAL PEAK IS 3.9
Electric Demand (W/ft2)

Fans W/FT )
2

3.0
2.5
2.0
1.5
1.0
0.5
0.0
0:00

3:00

6:00

9:00

12:00

15:00

18:00

21:00

0:00

APPENDIX 2 – MEASURED FAN PERFORMANCE X193X

 FIGURE 116:
PEAK DAY ELECTRIC
7.0
DEMAND, SITE 2,
Other
8/9/2002 6.0
Cooling Towers

Electric Demand (W/ft2)

( C U M U L AT I V E G R A P H ;
5.0 Chillers
TOTAL PEAK IS 6.4
Fans
2
W/FT ) 4.0

3.0

2.0

1.0

0.0
0:00

3:00

6:00

9:00

12:00

15:00

18:00

21:00
Fans vs. Chillers

Which is the bigger energy consumer? At two monitored sites, the fans account for more
electricity consumption than the chiller.

FIGURE 117:
0.35
CO MPAR I S O N O F FA N
AND C H I L L E R E N E R G Y AT 0.30
S I T E 1 ( C U M U L AT I V E
0.25
G R A P H , E .G . CO MB I N E D
2
TOTAL IS 0.30 KWH/FT 0.20
kWh/ft2

Chillers
IN J U LY ) Fans
0.15

0.10

0.05

0.00
Nov-01

Dec-01
Jan-02

Jul-02
Feb-02
Mar-02

Apr-02
May-02

Jun-02

Aug-02

Oct-02

Nov-02
Dec-02
Sep-02

Jan-03
Feb-03

Mar-03

X194X APPENDIX 2 – MEASURED FAN PERFORMANCE

 0.40 FIGURE 118:
0.35 CO MPAR I S O N O F FA N
AND C H I L L E R E N E R G Y AT
0.30
S I T E 2 ( C U M U L AT I V E
0.25
kWh/ft2

G R A P H , E .G . CO MB I N E D
Chillers
0.20 TOTAL IS 0.34 KWH/FT
2
Fans
0.15 IN J U LY ) .

0.10

0.05

0.00
Feb-02

Mar-02

Apr-02

May-02

Jun-02

Jul-02

Aug-02

Sep-02

Oct-02

Nov-02

Dec-02

Jan-03

APPENDIX 2 – MEASURED FAN PERFORMANCE X195X

 APPENDIX 3 – AIRFLOW IN THE REAL WORLD

Research shows, as should be expected, that VAV systems seldom, if ever, reach their design
airflow, usually getting by with significantly less. This fact is illustrated in the examples
below at the both the zone level and the air handler level. In addition, the zone level data
shows that many zones spend a majority of their time at minimum flow. In these cases, it’s
likely that even lower airflow would have provided comfort while also saving fan and reheat
energy.

Based on the real world dynamics of a VAV system, the designer should pay special attention
to system performance at typical conditions (where the system spends the most hours) as well
as at the minimum load conditions. The VAV Box Selection section provides relevant
guidance on VAV box selection and control. The Fan Type, Size and Control section
addresses design at the air handler level.

Figure 119 through Figure 124 illustrate several examples of zone airflow variations. Similar
data for total air handler airflow are shown in Figure 125 through Figure 128. Since airflow
requirements depend on many factors, these results should be considered illustrations of
VAV system dynamics and not be considered directly comparable to conditions in other
buildings.

Interior Zone Airflow

Interior zones are affected very little by building envelope cooling or heating loads, and
Figure 119 and Figure 120 show that airflow is nearly constant for a sample of three interior
zones at Site 3, an office building.

X196X APPENDIX 3 – AIRFLOW IN THE REAL WORLD

 Fraction of Operating Hours (%) 100% FIGURE 119:

90% SITE 3, SAMPLE OF

INTERIOR ZONES, WARM
80%
PERIOD (8/8/02 -
70%
9/7/02)
60%
50%
40%
30%
20%
10%
0%
0.0-0.1

0.1-0.2

0.2-0.3

0.3-0.4

0.4-0.5

0.5-0.6

0.6-0.7

0.7-0.8

0.8-0.9

0.9-1.0

1.0-1.1

1.1-1.2

1.2-1.3

1.3-1.4
Airflow (cfm/sf)

Zone A Zone B Zone C

FIGURE 120:
100%
Fraction of Operating Hours (%)

SITE 3, SAMPLE OF
90% I N T E R I O R ZO N E S , CO O L

80% PERIOD

70% (12/12/02-1/11/03)

60%
50%
40%
30%
20%
10%
0%
0.0-0.1

0.1-0.2

0.2-0.3

0.3-0.4

0.4-0.5

0.5-0.6

0.6-0.7

0.7-0.8

0.8-0.9

0.9-1.0

1.0-1.1

1.1-1.2

1.2-1.3

1.3-1.4

Airflow (cfm/sf)

Zone A Zone B Zone C

While some minor variation exists, airflow falls between 0.3 and 0.4 cfm/ft2 for 70% to 90%
of operating hours during both warm and cool times of the year. Figure 121 shows more
variation in interior zone airflow in a Site 4, a courthouse.

APPENDIX 3 – AIRFLOW IN THE REAL WORLD X197X

 FIGURE 121: 100%
SITE 4, SAMPLE OF 90%

Fraction of Operating Hours

INTERIOR ZONES 80%
(10/18/02-2/24/03) 70%
60%
50%
40%
30%
20%
10%
0%

0.0-0.15

0.15-0.2

0.2-0.3

0.3-0.4

0.4-0.5

0.5-0.6

0.6-0.7

0.7-0.8

0.8-0.9

0.9-1.0
Normalized Airflow (Fraction of Peak Flow)

Zone A Zone B Zone C

Perimeter Zone Airflow

Airflow variation can be more significant in perimeter zones than in interior spaces. The
following examples show this variation, but they also reveal that airflow is at low level
(probably the minimum flow set point entered in the control system) for a large part of the
time. As discussed earlier in this guideline, these minimum flow set points lead to lost savings
opportunities. See the chapter on VAV Box Selection.

X198X APPENDIX 3 – AIRFLOW IN THE REAL WORLD

 Fraction of Operating Hours (%) Fraction of Operating Hours (%)

0%
10%
20%
30%
40%
50%
60%
70%
80%
90%

0.0%
10.0%
20.0%
30.0%
40.0%
50.0%
60.0%
70.0%
80.0%
90.0%
100.0%
0.0-0.15
0.0-0.15
0.15-0.2
0.15-0.2
0.2-0.3
0.2-0.3
0.3-0.4
0.3-0.4
0.4-0.5
0.4-0.5
0.5-0.6
0.5-0.6
0.6-0.7
0.6-0.7

Zone D
0.7-0.8
0.7-0.8
0.8-0.9 0.8-0.9

Zone D
0.9-1.0 0.9-1.0

Zone E
1.0-1.1 1.0-1.1
1.1-1.2 1.1-1.2

Zone E
1.2-1.3 1.2-1.3

Airflow (cfm/sf)
Airflow (cfm/sf)

Zone F
1.3-1.4 1.3-1.4
1.4-1.5 Zone F 1.4-1.5
1.5-1.6 1.5-1.6
1.6-1.7 1.6-1.7
1.7-1.8 1.7-1.8
1.8-1.9 1.8-1.9
1.9-2.0 1.9-2.0
2.0+ 2.0+
FIGURE 123:
FIGURE 122:

CO O L PE R I O D
WARM PERIOD

PERIMETER ZONES,
PERIMETER ZONES,

SITE 3, SAMPLE OF
SITE 3, SAMPLE OF

(8/8/02 - 9/7/02)

(12/12/02-1/11/03)

APPENDIX 3 – AIRFLOW IN THE REAL WORLD X199X

 FIGURE 124: 80%
SITE 4, SAMPLE OF 70%

Fraction of Operating Hours

PERIMETER ZONES
60%
(10/18/02-2/24/03)
50%

40%

30%

20%

10%

0%

0.0-0.15

0.15-0.2

0.2-0.3

0.3-0.4

0.4-0.5

0.5-0.6

0.6-0.7

0.7-0.8

0.8-0.9

0.9-1.0
Normalized Airflow (Fraction of Peak Flow)

Zone D Zone E Zone F

System Level Airflow

Diversity of airflow at the zone level leads, of course, to diversity at the system level.
Monitored data from four sites presented in Figure 125 through Figure 128 provide a strong
argument for designing the system to work optimally at flows less that predicted by
traditional design methods.

Figure 125 shows that Site 1, which was designed to supply1.4 cfm/ft2, never exceeds 0.8
cfm/ft2 and usually operates between 0.4 and 0.6 cfm/ft2 during the day. This facility
operates 24 hours per day, which accounts for the large fraction of hours in the 0.2 to 0.3
cfm/ft2 range.

X200X APPENDIX 3 – AIRFLOW IN THE REAL WORLD

 60%

Fraction of Operating Hours (%)

FIGURE 125:
TOTAL SYSTEM
50%
A I R F LO W, S I T E 1
40%

30%
Design Airflow
20%
1.4 cfm/sf

10%

0%
0.0-0.1

0.1-0.2

0.2-0.3

0.3-0.4

0.4-0.5

0.5-0.6

0.6-0.7

0.7-0.8

0.8-0.9

0.9-1.0

1.0-1.1

1.1-1.2

1.2-1.3

1.3-1.4
Airflow (cfm/sf)

Warm (8/1/02-9-30/02) Cool (1/1/03-2/28/03)

At Site 2, illustrated in Figure 126, a clear seasonal variation exists in system airflow, which is
typically 0.6 to 0.8 cfm/ft2 during warm weather and 0.5 to 0.6 cfm/ft2 in cool conditions.
This building is located nearby Site 1 but has a significantly larger window area.

FIGURE 126:
60%
Fraction of Operating Hours (%)

TOTAL SYSTEM
50% A I R F LO W, S I T E 2

40%

30%
Design Airflow
20% 0.9 cfm/sf

10%

0%
0.0-0.1

0.1-0.2

0.2-0.3

0.3-0.4

0.4-0.5

0.5-0.6

0.6-0.7

0.7-0.8

0.8-0.9

0.9-1.0

1.0-1.1

1.1-1.2

1.2-1.3

1.3-1.4

Airflow (cfm/sf)

Warm (8/1/02-9-30/02) Cool (12/1/02-1/30/03)

Both Sites 3 (Figure 127) and 4 (Figure 128) experience seasonal airflow variation at the air
handler level, with the majority of operating hours occurring at or below one-half of the
design airflow.

APPENDIX 3 – AIRFLOW IN THE REAL WORLD X201X

 FIGURE 127:
80%

Fraction of Operating Hours (%)

TOTAL SYSTEM
70%
A I R F LO W, S I T E 3
60%
50%
40%
30% Design Airflow
20% 0.83 cfm/sf

10%
0%

0.0-0.1

0.1-0.2

0.2-0.3

0.3-0.4

0.4-0.5

0.5-0.6

0.6-0.7

0.7-0.8

0.8-0.9

0.9-1.0

1.0-1.1

1.1-1.2

1.2-1.3

1.3-1.4
Airflow (cfm/sf)

Warm (8/8/02 - 9/7/02) Cool (12/12/02-1/11/03)

FIGURE 128: 60%

Fraction of Operating Hours (%)

TOTAL SYSTEM
50%
A I R F LO W, S I T E 4
40%

30%
Design Airflow
20% 1.1 cfm/sf

10%

0% 0.0- 0.1- 0.2- 0.3- 0.4- 0.5- 0.6- 0.7- 0.8- 0.9- 1.0- 1.1- 1.2- 1.3-
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4

Airflow (cfm/sf)

Mild (10/18/02-11/6/02) Cool (12/1/02-1/31/03)

X202X APPENDIX 3 – AIRFLOW IN THE REAL WORLD

 APPENDIX 4 – COOLING LOADS IN THE REAL WORLD

For a different perspective on zone load profiles, this appendix discusses five examples of air
handler cooling output, including include both interior and perimeter zones. Load profile
are represented here as the number of hours that loads fall into different ranges.

At Sites 1 and 3, the cooling delivered by the air handler rarely exceeds an average of 2.0
W/ft2 and is often less than 1.0 W/ft2. The other two buildings show higher loads, with the
majority of hours at Sites 2 and 4 falling between 1.5 and 3.0 W/ft2.

Some of the sites show much more seasonal variation than others. Site 1 shows only slightly
higher loads in the warmer months, while Sites 2 and 3 have loads of about 1 W/ft2 higher in
the warm periods. (Warm weather data is not available for Site 4.)

With the exception of Site 2, none of these buildings require levels close to their peak air
handler cooling capacity.

These results can be useful in several ways:

ƒ Providing some insight in the range of operation typically required of an HVAC system.
ƒ Serving as a benchmark for evaluating simulation results.
ƒ Serving as a reminder that typical load calculations and sizing decisions are conservative.

Of course, judgment needs to used when applying monitored data from existing buildings.
Loads today may be lower than in the past. Advances in lighting technologies, glazing
performance, and office equipment have lead to decreases in loads over recent years. It is also
useful to know something about the controls and setpoints in the existing buildings.
Minimum zone airflow setpoints are especially important information because they affect the
air flow profile at both the zone level and the air handler level.

FIGURE 129:
60% SITE 1
Frequency (% hrs)

50%
(DARK BAR INCLUDES
40% Peak AHU Capacity
= 8.6 W/ft2 JAN-MAY 2002 AND
30%
20% NOV-DEC 2002, LIGHT

10% BAR COVERS

0% JUN-OCT 2002)
0.1-0.5

0.5-1.0

1.0-1.5

1.5-2.0

2.0-2.5

2.5-3.0

3.0-3.5

3.5-4.0

4.0-4.5

4.5-5.0

5.0-5.5

5.5-6.0

Cooling Load, W/ft2

APPENDIX 4 – COOLING LOADS IN THE REAL WORLD X203X

 FIGURE 130:
SITE 2 30%

Frequency (% hrs)
25% Peak AHU Capacity
(LIGHT BAR INCLUDES
20% = 6 W/ft2
JUN-OCT 2002, DARK
15%
BAR COVERS NOV 2002
10%
– JAN 2003) 5%
0%

0.1-0.5

0.5-1.0

1.0-1.5

1.5-2.0

2.0-2.5

2.5-3.0

3.0-3.5

3.5-4.0

4.0-4.5

4.5-5.0

5.0-5.5

5.5-6.0
Cooling Load, W/ft2

FIGURE 131:
SITE 3 60%
Frequency (% hrs)

50% Peak AHU Capacity

(LIGHT BAR INCLUDES
40% = 4.5 W/ft2
AUG-OCT 2002, DARK
30%
BAR COVERS NOV 2002
20%
– JAN 2003) 10%
0%
0.1-0.5

0.5-1.0

1.0-1.5

1.5-2.0

2.0-2.5

2.5-3.0

3.0-3.5

3.5-4.0

4.0-4.5

4.5-5.0

5.0-5.5

5.5-6.0
Cooling Load, W/ft2

FIGURE 132:
SITE 4 50%
Frequency (% hrs)

(DARK BAR INCLUDES 40%

Peak AHU Capacity
N O V. 2 5 , 2 0 0 2 - F E B . 30%
= 9+ W/ft2
24, 2003) 20%
10%
0%
0.1-0.5

0.5-1.0

1.0-1.5

1.5-2.0

2.0-2.5

2.5-3.0

3.0-3.5

3.5-4.0

4.0-4.5

4.5-5.0

5.0-5.5

5.5-6.0

Cooling Load, W/ft2

Figure 133 is similar to the previous graph except that it includes cooling delivered from an
air handler to a group of only interior zones. 32 An important thing to note here is how few

32
Since these zones have no connection to the outdoors, these data represent the sum of loads from lights, plug
loads and occupants.

X204X APPENDIX 4 – COOLING LOADS IN THE REAL WORLD

hours are spent at peak load compared to typical load. (This air handler is designed to deliver
up to 6 W/ft2 of cooling to these zones).

FIGURE 133:
25%
Fraction of Operating Hours

MONITORED SENSIBLE
CO O L I N G LOAD F O R AN
20%
AIR HANDLER SERVING
15% 19 INTERIOR ZONES,
SITE 4
10%

5%

0%
0 - 0.25
0.25 - 0.50

0.50 - 0.75
0.75 - 1.00
1.00 - 1.25
1.25 - 1.50
1.50 - 1.75
1.75 - 2.00

2.00 - 2.25
2.25 - 2.50

2.50 - 2.75
2.75 - 3.00

3.00 - 3.25

3.25 - 3.50
3.50 - 3.75

3.75 - 4.00
4.00 - 4.25

4.25 - 4.50

4.50 - 4.75
4.75 - 5.00
Cooling Load, W/ft2

APPENDIX 4 – COOLING LOADS IN THE REAL WORLD X205X

 APPENDIX 5 – DOE-2 FAN CURVES

Unlike supply air temperature reset, DOE-2 does not have keywords with which one can
explicitly model static pressure reset. However, static pressure reset can be reasonably
accurately modeled by using a DOE-2 fan curve that represents static pressure reset. Using
the Characteristic System Curve Fan Model described previously in Visualizing Fan
Performance and in Appendix 12, the authors have generated a library of DOE-2 fan curves
that can be imported into eQuest. These curves include the part load efficiency of the
motor, variable speed drive and standard belts.

How to use the library:

1. From within eQuest go to File, Import and browse to where you have
saved the “New DOE-2 Fan Curves.txt” text file. Be sure to change the
Files of type from BDL Input Files (*.inp) to All Files (*.*). Import the
text file.
2. Set Fan Control to Fan EIR FPLR and then select the desired fan curve
from the pulldown menu next to Fan EIR = f(PLR). See screen capture
below.

If you select “Variable Speed” under Fan Control in eQuest, the program by default uses a
quadratic equation with the following coefficients: ( 0.219762, -0.874784, 1.6526 ). This
default curve is graphed below along with two of the curves in New DOE-2 Fan Curves.txt.
These two curves cannot be fit with a quadratic curve but they can be fit with a cubic curve
and DOE-2 will accept a cubic curve for the FAN-EIR-FPLR Keyword

FIGURE 134:
DOE 2 Fan Curves
SA M P L E PA RT LO A D
PERFORMANCE FOR A 100%

25HP MOTOR (SOURCE: 90%

O A K R I D G E N AT I O N A L DOE2 default Variable Speed
80%
Perfect SP Reset
L A B O R AT O R Y ’ S
70% No SP Reset
PUMPING SYSTEM
Percent Power

60%
ASSESSMENT TOOL)
50%
Note that below 30% airflow the
power is constant if MIN-FAN-
40%
RATIO = 0.30
30%

20%

10%

0%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percent Airflow

X206X APPENDIX 5 – DOE-2 FAN CURVES

Using the Characteristic System Curve Fan Model the authors evaluated the factors that had
the biggest impact on the shape of the fan curve. They evaluated different fan types,
oversizing/undersizing fans, overestimating actual static, min VSD speed, and the shape of
the system curve. Some of these impacts are illustrated in the figures below. Clearly the
biggest factor is the shape of the system curve – specifically the intercept at 0 flow. (The
biggest factor affecting the shape of the system curve is the use of static pressure reset.)
Therefore the library only has fan curves that represent different system curves (i.e. how
successfully static pressure reset is implemented).

This is not to say that all fans are alike. Clearly some fan types and some fan selections are
more efficient than others but those difference are not captured by the fan part load
performance curve (FAN-EIR-FPLR), they are captured by the SUPPLY-STATIC and the
SUPPLY-EFF or SUPPLY-DELTA-T and SUPPLY-KW/FLOW keywords.

FIGURE 135:
Comparison of plenum, forward curve and housed airfoil fans on system curves
through 0.5" PA RT LO A D
PERFORMANCE BY FAN
100%
TYPE (0.5 IN.)
90%

80%
PL-0.5" FC-0.5
70%
Percent Power

60%

50% AF-0.5

40%

30%

20%

10%

0%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percent Airflow

APPENDIX 5 – DOE-2 FAN CURVES X207X

 FIGURE 136:
Comparison of plenum, forward curve and housed airfoil fans on system curves
PA RT LO A D through 0.5"
PERFORMANCE BY FAN
100%
TYPE (1.0 IN.)
90%

80%
PL-1.0 FC-1.0
70%
Percent Power
60%

50% AF-1.0

40%

30%

20%

10%

0%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percent Airflow

FIGURE 137: What if the engineer was more or less conservative in terms of how close the selection is to the Surge Line?
Plenum fans at design condition of 72,500 CFM at 4", on system curve through design point and through 0.5" at 0 flow
PA RT LO A D F A N Of course, what this does not show is that the conservative fan is higher
100% KW at design and the aggressive fan is lower KW at design

PERFORMANCE BY
90%
WHEEL SIZE
80% Basecase (66" fan)

70%

Conservative (60" fan)

60%
Percent Power

50% Aggressive (73" fan)

40%

30%

20%

10%

0%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percent Airflow

X208X APPENDIX 5 – DOE-2 FAN CURVES

 Plenum fans on system curves through 0.5" at 0 flow and through Different DP at Design Flow
FIGURE 138:
i.e. what if the engineer estimated 4" at design but it turned out to be 3"?
What this does not show is that the fan through 3" has a lower KW at design. PA RT LO A D F A N
100%
PERFORMANCE BASED
90%
ON T R U E D E S I G N S TAT I C
Plenun through 4" DP at Design Flow Plenum through 3" DP at Design Flow
80%

70% Air Foil through 4" Airfoil through 3"

60%
Percent Power

50%

40%

30%

20%

10%

0%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percent Airflow

FIGURE 139:
Part Load Performance versus Minimum Static Pressure Setpoint
Plenum airfoil fans on system curves running through several min SP points PA RT LO A D F A N
100% PERFORMANCE BASED
90% ON M I N I M U M S TAT I C
PL-0" PRESSURE
80%

70% PL-0.5
Percent Power

60% PL-1.0
50%
'PL-1.5
40%

30%

20%

10%

0%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percent Airflow

The curves in the library are described below:

"Typical VSD Fan" - This is based on a plenum airfoil fan on a system curve through 0.7 in.

"Perfect SP Reset VSD Fan" - This is based on a plenum airfoil fan on a system curve
through 0 in. Perfect reset is of course not possible because not all zones will modulate
exactly in unison, i.e. some zones will always want more pressure than others.

APPENDIX 5 – DOE-2 FAN CURVES X209X

 "Good SP Reset VSD Fan" - This is based on a plenum airfoil fan on a system curve through
0.5 in.

"No SP Reset VSD Fan" - This is based on a plenum airfoil fan on a system curve through
1.5 in.

The following curves are based on plenum airfoil fans on system curves ranging from 0” to
1.5 in.
Plenum 0 in.
Plenum 0.3 in.
Plenum 0.4 in.
Plenum 0.5 in.
Plenum 0.6 in.
Plenum 0.7 in.
Plenum 0.8 in.
Plenum 1.0 in.
Plenum 1.5 in.

The following fan curve coefficients were developed using the characteristic system curve fan
model. Curves include part load performance of the fan, belt, motor and VSD.
This is based on a plenum airfoil fan on a system curve through 0.7 in.
"Typical VSD Fan" = CURVE-FIT
TYPE = CUBIC
INPUT-TYPE = COEFFICIENTS
OUTPUT-MIN =0
OUTPUT-MAX =1
COEFFICIENTS = (0.047182815, 0.130541742, -0.117286942, 0.940313747)
This is based on a plenum airfoil fan on a system curve through 0 in.
"Perfect SP Reset VSD Fan" = CURVE-FIT
TYPE = CUBIC
INPUT-TYPE = COEFFICIENTS
OUTPUT-MIN =0
OUTPUT-MAX =1
COEFFICIENTS = (0.027827882, 0.026583195, -0.0870687, 1.03091975)
This is based on a plenum airfoil fan on a system curve through 0.5 in.
"Good SP Reset VSD Fan" = CURVE-FIT
TYPE = CUBIC
INPUT-TYPE = COEFFICIENTS
OUTPUT-MIN =0
OUTPUT-MAX =1
COEFFICIENTS = (0.040759894, 0.08804497, -0.07292612, 0.943739823)
This is based on a plenum airfoil fan on a system curve through 1.5 in.
"No SP Reset VSD Fan" = CURVE-FIT
TYPE = CUBIC
INPUT-TYPE = COEFFICIENTS
OUTPUT-MIN =0
OUTPUT-MAX =1
COEFFICIENTS = (0.070428852, 0.385330201, -0.460864118, 1.00920344)
"Plenum Fan 0.3 in. = CURVE-FIT

X210X APPENDIX 5 – DOE-2 FAN CURVES

 TYPE = CUBIC
INPUT-TYPE = COEFFICIENTS
OUTPUT-MIN =0
OUTPUT-MAX =1
COEFFICIENTS = (0.034171263, 0.059448041, -0.061049511, 0.966140782)
"Plenum Fan 0.4 in. = CURVE-FIT
TYPE = CUBIC
INPUT-TYPE = COEFFICIENTS
OUTPUT-MIN =0
OUTPUT-MAX =1
COEFFICIENTS = (0.037442571, 0.072000619, -0.062564426, 0.952238103)
"Plenum Fan 0.5 in. = CURVE-FIT
TYPE = CUBIC
INPUT-TYPE = COEFFICIENTS
OUTPUT-MIN =0
OUTPUT-MAX =1
COEFFICIENTS = (0.040759894, 0.08804497, -0.07292612, 0.943739823)
"Plenum Fan 0.6 in. = CURVE-FIT
TYPE = CUBIC
INPUT-TYPE = COEFFICIENTS
OUTPUT-MIN =0
OUTPUT-MAX =1
COEFFICIENTS = (0.044034586, 0.107518462, -0.091288825, 0.939910504)
"Plenum Fan 0.7 in. = CURVE-FIT
TYPE = CUBIC
INPUT-TYPE = COEFFICIENTS
OUTPUT-MIN =0
OUTPUT-MAX =1
COEFFICIENTS = (0.047182815, 0.130541742, -0.117286942, 0.940313747)
"Plenum Fan 0.8 in. = CURVE-FIT
TYPE = CUBIC
INPUT-TYPE = COEFFICIENTS
OUTPUT-MIN =0
OUTPUT-MAX =1
COEFFICIENTS = (0.050254136, 0.156227953, -0.148857337, 0.943697119)
Plenum 0 in. (0.027827882, 0.026583195, -0.0870687, 1.03091975)
Plenum 0.3 in. (0.034171263, 0.059448041, -0.061049511, 0.966140782)
Plenum 0.4 in. (0.037442571, 0.072000619, -0.062564426, 0.952238103)
Plenum 0.5 in. (0.040759894, 0.08804497, -0.07292612, 0.943739823)
Plenum 0.6 in. (0.044034586, 0.107518462, -0.091288825, 0.939910504)
Plenum 0.7 in. (0.047182815, 0.130541742, -0.117286942, 0.940313747)
Plenum 0.8 in. (0.050254136, 0.156227953, -0.148857337, 0.943697119)
Plenum 1.0 in. (0.056118534, 0.214726686, -0.226093052, 0.957646288)
Plenum 1.5 in. (0.070428852, 0.385330201, -0.460864118, 1.00920344)

APPENDIX 5 – DOE-2 FAN CURVES X211X

 APPENDIX 6 – SIMULATION MODEL DESCRIPTION

This appendix provides a brief description of the simulation model used to evaluate several of
the guideline recommendations, including the following:
ƒ Comparison of “Standard” to “Best Practice” design performance (Introduction).
ƒ VAV box sizing criteria (VAV Box Selection).
ƒ Optimal supply air temperature reset control methods (Supply Air Temperature
Control).

More details of the assumptions and results are described in Analysis Report that documents
guidelines-related research.

Assumptions

Building Envelope
1. Five story, 50,000 ft2 square building. Each floor is 100 feet by 100 feet. 5
zones per floor, total 25 zones. Floor to floor height is 13 feet, plenum
height is 4 feet.
2. Continuous strip of glazing, double pane, low-e glass (DOE-2 code 2637,
similar to Viracon VE1-2M. SC = 0.43, U-value = 0.31, Tvis = 0.44).
40% WWR (window height is 5.2 feet).
3. 12-foot deep perimeter zones.
4. Exterior wall construction U-value is 0.088 Btu/hr-ft2-°F.
5. No skylights, no daylighting controls.

Climate

California Zone 3 (San Francisco Bay Area) and Zone 12 (Sacramento).

Internal Loads
1. Lighting power density: 1.5 W/ft2.
2. Equipment power density: 2.0 W/ft2.
3. Occupancy density: 100 ft2/ person.

X212X APPENDIX 6 – SIMULATION MODEL DESCRIPTION

Load Schedules

In order to capture the effect of reheat at low load, we used schedules that went up and down
over the course of each day. We used three “high,” three “medium,” and three “low”
schedules. Each zone was randomly assigned one of the nine schedules. The schedules were
the same for all weekdays of the year. For simplicity, we used the same schedule for lights,
people, and equipment.

Fan Schedule

5 am – 7 pm, 14 hours of operation.

Thermostat setpoints

72°F cooling, 70°F heating.

Design Airflow

Loads calculations were run in Trace. For each climate zone, we determined a normalized
airflow (CFM/ft2) for each orientation. These airflows were then multiplied by the zone
areas used in DOE-2. DOE-2 Keyword: ASSIGNED-CFM.
Orientation Design Flow Rate (CFM/ft2) TABLE 27:
CZ03 CZ12
BASECASE DESIGN AIR
North Zones 1.06 1.23
South Zones 2.21 2.28 FLOWS
East Zones 1.96 2.01
West Zones 2.19 2.36
Interior Zones 0.77 0.77

Zone Properties
1. OA calculated based on 15 cfm/person.
2. THERMOSTAT-TYPE: Reverse Action. For VAV systems, this
Thermostat Type behaves like a dual maximum thermostat, it allows the
airflow rate to rise above the minimum design heating airflow rate (i.e.,
the Minimum Flow Ratio).
3. THROTTLING-RANGE: 0.5°F.
4. MIN-FLOW-RATIO: DOE-2 takes the maximum of MIN-FLOW-
RATIO and MIN-CFM/SQFT to determine the minimum airflow.

MIN-FLOW-RATIO: this is the box turndown. It varies from 8.4% to 13%.

MIN-CFM/SQFT is set to 0.15 cfm/ ft2 (for ventilation).

APPENDIX 6 – SIMULATION MODEL DESCRIPTION X213X

 System Properties
1. PIU system with standard VAV terminals. One air handler for the
building.

2. Supply Fan efficiency: 60% (this includes motor, belt and drive
efficiency). Note that this works more for FC and Plenum.

3. FAN-CONTROL: EIR-FPLR. (This references the following curve).

4. Fan EIRFPLR curve: AnyFanWithVSD (DOE2 default curve for

VSD fan).

5. MIN-FAN-RATIO: 30%. This basically means that fan energy is

fairly constant below about 30% flow, which appears to be reasonably
accurate from our fan modeling.

6. SUPPLY-STATIC: 3.2 + “BTP.”: BTP is the Box Total Pressure. It

varies, depending on the parametric run.

7. Motor efficiency: 100% (Motor and drive efficiency is modeled in

the fan curve and peak efficiency).

8. Coil and fan capacity: A TRACE load calculation determines these

parameters.

9. MIN-SUPPLY-T: 55

10. COOL-CONTROL: CONSTANT

11. Drybulb economizer: Use fixed dry-bulb with Title 24 High Limits
of 75°F.

12. No return fan.

13. REHEAT-DELTA-T: 43°F (i.e., the highest allowable diffuser air

temperature).

Plant Properties
1. Water-cooled chilled water plant – default efficiencies.

2. Default HW boiler.

Utility Rates
1. Electricity rate: PG&E E-20s.

2. Gas rate: PG&E GNR1.

X214X APPENDIX 6 – SIMULATION MODEL DESCRIPTION

DOE2 Version

eQuest version 3.21 build 1778 was used to perform these simulation runs. DOE-2.2-41m is
the calculation engine.

Results

This section provides a brief summary of simulation results related to guideline

recommendations. More details are included in the Analysis Report.

Standard Practice vs. Best Practice Results

Table 28 describes the simulation assumptions used to compare standard and best practice.
From the perspective of energy performance, the two most significant differences are the fan
curve that approximates the impact of supply air pressure controls and the VAV box
minimum airflow fraction. Both of these measures lead to reductions in fan energy, and the
minimum flow fraction also saves cooling and reheat energy. The end-use energy results are
listed in Table 29 for the San Francisco and Sacramento climates.

Item Standard Practice Best Practice TABLE 28:

Fan curve Standard VSD (DOE-2 VSD with perfect pressure reset, which reduces
AIRSIDE CONTROL
default) without pressure fan power at partial flow compared to a
reset standard VSD STRATEGIES FOR
VAV box minimum 30% 10% or 0.15 cfm/ft2, whichever is greater. SIMULATION OF
airflow fraction STANDARD PRACTICE
Thermostat type Proportional, which means Reverse acting, which means that airflow can
that airflow is fixed at increase in heating mode. This allows a lower AND BEST PRACTICE
minimum fraction in minimum without risk of airflow being too low
heating mode to provide enough heating.
Supply air 55ºF to 60ºF based on Reset in the range between 55ºF and 65ºF
temperature reset outdoor air temperature based on warmest zone (temperature first)

Lighting Equipment Cooling Heat Pumps Fans Total Heating

Rejection
TABLE 29:
kWh kWh kWh kWh kWh kWh kWh therms
San SIMULATION RESULTS
Francisco FOR COMPARISON OF
Standard 76,715 102,286 79,773 915 30,834 33,231 323,755 4,560
STANDARD PRACTICE
Practice
Best 76,706 102,282 63,871 770 24,788 12,613 281,111 2,374 AND BEST PRACTICE
Practice
Savings 0% 0% 20% 16% 20% 62% 13% 48%
Sacramento
Standard 76,715 102,286 99,370 1,574 30,844 38,158 348,947 5,288
Practice
Best 76,706 102,282 89,780 1,424 29,686 18,432 318,374 3,479
Practice
Savings 0% 0% 10% 10% 4% 52% 9% 34%

APPENDIX 6 – SIMULATION MODEL DESCRIPTION X215X

 The fan performance curves used in the comparison of standard and best practice are
illustrated in Figure 140.

FIGURE 140:
DOE 2 Fan Curves
FAN PERFORMANCE
CURVES FOR 100%

S I M U L AT I O N 90%
DOE2 default (ANYFAN-VSD)
80%
SP Perfect Reset
70% SP Good Reset
Percent Power

60%

50%
Note that below 30% airflow the
40% power is constant if MIN-FAN-
RATIO = 0.30
30%

20%

10%

0%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percent Airflow

VAV Box Sizing Results

Simulations were used to evaluate the energy impact of six different criteria or rules of thumb
for box sizing ranging from 0.3 in. to 0.8 in. total pressure drop across the box. The
simulation model described above was the baseline for analysis and then the model was
modified to test sensitivity as follows:
ƒ 8-bit analog-to-digital converter on airflow sensor.
ƒ Aggressive load calculation assumptions.
ƒ Highly conservative load calculation assumptions.
ƒ Low load operating schedules.
ƒ High load operating schedules.
ƒ 24/7 fan operation.
ƒ 50°F and 60°F supply air temperature.
ƒ Supply pressure reset.
ƒ Larger window area.
ƒ Higher utility rate.
ƒ Fixed zone minimum airflow fraction.
ƒ High outdoor air load.

X216X APPENDIX 6 – SIMULATION MODEL DESCRIPTION

Figure 141 shows the incremental energy cost results for an average of all parametric runs,
indicating that 0.5 inches provides the best performance. Descriptions and results for each
scenario can be found in the Analysis Report.

FIGURE 141:
Average Results Across All Simulation Runs A V E R A G E R E S U LT S
ACROSS ALL
Incr. $/1000sf-yr

$10.00 S I M U L AT I O N R U N S
$8.00
$6.00
$4.00
$2.00
$-
0.3 0.4 0.5 0.6 0.7 0.8
Box Total Pressure

Supply Air Temperature Control Results

The simulation model was used to compare several supply air temperature reset methods,
and the results are listed in Table 30. The first two options, constant 55°F and SAT reset by
warmest zone, are DOE-2 control options, and the results are available directly from eQuest.
Results for the other five control options are combinations of two different simulations. In
each of these five cases, the supply air temperature is reset by warmest zone up to the point
the chiller turns on or the outdoor air temperature exceeds a fixed setpoint. At that point, the
supply air temperature is reduced to the design setpoint, T-min (typically around 55°F).
1. Constant 55.
2. Reset by warmest zone. Supply air temperature is reset between 55°F and
65°F in order to meet loads in the warmest zone.
3. Switch to T-min when chiller runs. This scheme uses the same reset
method as in Case 2, but switches to low SAT whenever the chiller is
needed.
4. Similar to case 3, except that the SAT is reduced to it’s minimum setpoint
whenever outdoor air temperature exceeds 60°F.
5. Same as 4 except with 65°F changeover point.
6. Same as 5 except with 70°F changeover point.
7. Same as 6 except with 75°F changeover point.

These results show Case 5 or 6 to provide the lowest electricity consumption as well as the
lowest source energy consumption. Therefore, it appears that it is best to reset the supply air
temperature upwards until the outdoor air temperature exceeds 65°F or 70°F, then reduce
the supply air temperature to T-min in order to minimize fan energy and rely on the chiller
for cooling.

APPENDIX 6 – SIMULATION MODEL DESCRIPTION X217X

 TABLE 30:
Combined
SUPPLY AIR
Cooling & Total HVAC Source
TEMPERATURE CONTROL Pumps Fans HVAC Elec. Heating Energy
SIMULATION RESULTS SAT Control Method kWh/ft2 kWh/ft2 kWh/ft2 kBtu/ft2 kBtu/ft2
San Francisco Climate
1. Constant 55 2.43 0.38 2.81 5.23 33.9
2. Reset by zone demand 1.75 0.47 2.22 4.45 27.2
3. Switch to T-min when 1.82 0.40 2.22 4.64 27.3
chiller runs
4. Switch to T-min when 1.88 0.40 2.28 4.58 27.9
OAT > 60
5. Switch to T-min when 1.76 0.43 2.19 4.49 26.9
OAT > 65
6. Switch to T-min when 1.75 0.45 2.20 4.46 27.0
OAT > 70
7. Switch to T-min when 1.75 0.46 2.21 4.45 27.1
OAT > 75
Sacramento Climate
Constant 55 2.76 0.52 3.28 7.38 41.0
Reset by zone demand 2.30 0.63 2.93 6.55 36.5
Switch to T-min when 2.33 0.52 2.85 6.80 36.0
chiller runs
Switch to T-min when 2.39 0.52 2.91 6.79 36.6
OAT > 60
Switch to T-min when 2.30 0.54 2.84 6.60 35.7
OAT > 65
Switch to T-min when 2.29 0.55 2.84 6.56 35.7
OAT > 70
Switch to T-min when 2.29 0.57 2.86 6.55 35.9
OAT > 75

Typical vs. Best Practice Performance

Significant fan and reheat energy savings are possible through the design strategies promoted
in this Design Guide. The potential savings are illustrated in the graphs below which present
simulation results; in this example the “Standard” case is a reasonably efficient code-
complying system and the “Best” case includes a number of the improvements suggested in
this guideline. The result of this simulation show that fan energy drops by 50% to 60%, and
reheat energy reduces between 30% and 50%.

This example is by no means comprehensive. For example these savings do not include the
impact of reducing duct pressure drop through careful design, the impact of properly
designing 24/7 spaces and conference rooms, or the potential savings from demand based
ventilation controls in high density occupancies.

Most of the savings are due to the efficient “turndown” capability of the best practices design
and the fact that HVAC systems operate at partial load nearly all the time. The most
important measures are careful sizing of VAV boxes, minimizing VAV box supply airflow
setpoints, controlling VAV boxes using a “dual maximum” logic that allows lower airflows in

X218X APPENDIX 6 – SIMULATION MODEL DESCRIPTION

the deadband mode, and supply air pressure reset control. Together these provide substantial
fan and reheat savings because typical systems operate many hours at minimum (yet higher
than necessary) airflow. The importance of turndown capability is emphasized by examples
of monitored airflow profiles in Appendix 3 and cooling load profiles in Appendix 4.

FIGURE 142:
4 12
SAN FRANCISCO
10
3

kBtu/yr/f t2
8
kWh/yr/ft2

Cooling
2 6 Heat
Fan
4
1
2
0 0
Standard Best Standard Best

FIGURE 143:
4 12
SACRAMENTO
10
3
kWh/yr/ft2

kBtu/yr/ft2

8
Cooling
2 6 Heat
Fan
4
1
2
0 0
Standard Best Standard Best

APPENDIX 6 – SIMULATION MODEL DESCRIPTION X219X

 APPENDIX 7 – VAV BOX MINIMUM AND MAXIMUM FLOWS FOR 2 MANUFACTURERS
We iterated on Max CFM to achieve a total pressure drop (static + velocity) of 0.50"
Assumptions:
Inlet SP: 1.5"
downstream SP: 0.25"
Standard 1/2" liner
2 row hot water coil
ARI 885-95 noise attenuation

Titus DESV - 0.5" total pressure drop

Even and Odd Sizes Even Sizes Only
min VP Titus static velocity total Radia Outlet
Nomina sensor Amp. Min Min Max press. press. press. ted Best Worst Best Worst outlet Outlet Area Area velocity velocity
l size output Factor FPM CFM CFM Drop Drop Drop NC turndown Turndown turndown Turndown W H (ft2) Inlet in out VP in VP out
4 0.004 1.64 198 17 225 0.1 0.403 0.50 28 8% 17% 8% 17% 12 8.0 0.53 0.09 2,578 421 0.41 0.01
5 0.004 1.96 181 25 310 0.19 0.301 0.49 24 8% 11% 12 8.0 0.53 0.14 2,273 580 0.32 0.02
6 0.004 3.08 144 28 393 0.28 0.216 0.50 27 7% 9% 7% 13% 12 8.0 0.53 0.20 2,002 735 0.25 0.03
7 0.004 2.58 158 42 540 0.28 0.216 0.50 24 8% 11% 12 10.0 0.69 0.27 2,021 785 0.25 0.04
8 0.004 2.39 164 57 675 0.32 0.173 0.49 24 8% 11% 8% 15% 12 10.0 0.69 0.35 1,934 982 0.23 0.06
9 0.004 2.30 167 74 865 0.31 0.196 0.51 28 9% 11% 14 12.5 1.04 0.44 1,958 833 0.24 0.04
10 0.004 2.31 167 91 1045 0.33 0.166 0.50 32 9% 11% 9% 13% 14 12.5 1.04 0.55 1,916 1,007 0.23 0.06
12 0.004 2.77 152 120 1465 0.34 0.154 0.49 28 8% 11% 8% 11% 16 15.0 1.46 0.79 1,865 1,005 0.22 0.06
14 0.004 2.02 178 191 2020 0.33 0.169 0.50 37 9% 13% 9% 13% 20 17.5 2.18 1.07 1,890 928 0.22 0.05
16 0.004 2.12 174 243 2570 0.34 0.155 0.50 27 9% 12% 9% 12% 24 18.0 2.72 1.40 1,841 946 0.21 0.06
Average Turndown 10.0% 11.0%

Krueger LMHS - 0.5" total pressure drop

Even and Odd Sizes Even Sizes Only
min VP Krueger static velocity total Radia Outlet
Nomina sensor Amp. Min Min Max press. press. press. ted Best Worst Best Worst outlet Outlet Area Area velocity velocity
l size output Factor FPM CFM CFM Drop Drop Drop NC turndown Turndown turndown Turndown W H (ft2) Inlet in out VP in VP out
4 0.004 2.33 166 14 230 0.08 0.422 0.50 21 6% 14% 6% 14% 12 8.0 0.53 0.09 2,636 430 0.43 0.01
5 0.004 2.33 166 23 333 0.15 0.348 0.50 20 7% 10% 12 8.0 0.53 0.14 2,442 623 0.37 0.02
6 0.004 2.33 166 33 425 0.24 0.253 0.49 21 8% 10% 8% 14% 12 8.0 0.53 0.20 2,165 795 0.29 0.04
7 0.004 2.33 166 44 580 0.25 0.249 0.50 20 8% 10% 12 10.0 0.69 0.27 2,170 844 0.29 0.04
8 0.004 2.33 166 58 675 0.33 0.173 0.50 22 9% 10% 9% 14% 12 10.0 0.69 0.35 1,934 982 0.23 0.06
9 0.004 2.33 166 73 930 0.27 0.226 0.50 17 8% 11% 14 12.5 1.04 0.44 2,105 896 0.28 0.05
10 0.004 2.33 166 91 1100 0.32 0.184 0.50 19 8% 10% 8% 13% 14 12.5 1.04 0.55 2,017 1,060 0.25 0.07
12 0.004 2.33 166 130 1560 0.32 0.175 0.49 19 8% 12% 8% 12% 16 15.0 1.46 0.79 1,986 1,070 0.25 0.07
14 0.004 2.33 166 177 2130 0.31 0.188 0.50 18 8% 11% 8% 11% 20 17.5 2.18 1.07 1,992 978 0.25 0.06
16 0.004 2.33 166 232 2730 0.32 0.175 0.50 22 8% 11% 8% 11% 24 18.0 2.72 1.40 1,955 1,005 0.24 0.06
Average Turndown 9.4% 10.4%

Note: Amplification factor, static pressure drop, radiated noise criteria, and even inlet/outlet area are all subject to change as manufacturers make
modifications to their designs and to their selection software. Therefore the above data is for reference only. Please consult the latest manufacturers'
literature and software for more current information.

X220X APPENDIX 7 – VAV BOX MINIMUM AND MAXIMUM FLOWS FOR 2 MANUFACTURERS
 APPENDIX 8 – HOW TO MODEL DIFFERENT
VAV ZONE CONTROLS IN DOE-2.2

This section discusses how to model three different VAV zone controls: single maximum,
dual maximum with constant volume heating, and dual maximum with VAV heating, using
eQuest, the user interface for DOE-2.2.

Single Maximum

Control Sequence

The sequence of control is show in Figure 144. As cooling load decreases the airflow is
reduced from the maximum airflow (on the far right side of the figure) down to the
minimum flow. Then as heating is required the reheat valve is modulated to maintain the
space temperature at setpoint. With this sequence the minimum flow rate in deadband
(between heating and cooling) is also the flow rate in heating mode.

Maximum FIGURE 144:

Airflow Setpoint
SINGLE MAXIMUM ZONE

Reheat Valve Position CONTROL SEQUENCE

Airflow Setpoint

Minimum
Airflow
Setpoint

Heating Loop Cooling Loop

Dead
Band

eQuest Modeling Guidance

Figure 145 demonstrates key inputs in eQuest to model the single maximum control
sequence. Highlights of the inputs include the following:

ƒ Thermostat Type under each Zone command should be Proportional; this ensures that the
airflow rate in deadband mode and heating mode to be constant.

APPENDIX 8 – HOW TO MODEL DIFFERENT VAV ZONE CONTROLS IN DOE-2.2 X221X

 ƒ Minimum Flow Ratio is the airflow ratio in deadband and heating mode. Increasing this
value leads increasing of reheat energy consumption; decreasing this value may lead to
poor air mixing in space at deadband and heating mode. Title 24-2005 generally does
not allow the minimum flow ratio to be higher than 30% of the design maximum flow.
However, surveys of existing buildings has revealed that many systems actually operate
at 40% to 50% minimum flow ratio.

ƒ Throttle Range input describes how much the space temperature is allowed to float
around the temperature setpoint. Modern DDC controls generally allow much tighter
control (e.g. 0.1 degree throttling range) compared to pneumatic controls (e.g. 2-4
degree throttling range). Zone mode is determined by space temperature relative to
temperature setpoints and throttling range. For example, if the cooling setpoint is 75°F,
the heating setpoint is 70°F and the throttling range is 0.5°F then eQuest applies the
following rules based on zone temperature to determine the zone mode:

o 70.5°F to 74.5°F: deadband mode

o 70.5°F and below: heating mode (The zone reaches full heating when
the space temperature is at 69.5°F. The zone is under heated if
temperature is below 69.5 °F.)

o 74.5°F and above: cooling mode. (The zone reaches full cooling when
temperature reaches 75.5°F. The zone is under cooled if temperature is
above 75.5 °F.)

FIGURE 145:
EQUEST INPUTS FOR
MODELING SINGLE
MAXIMUM CONTROL

X222X APPENDIX 8 – HOW TO MODEL DIFFERENT VAV ZONE CONTROLS IN DOE-2.2

eQuest Simulation Validation

eQuest hourly report outputs are an excellent way to confirm that the model is operating as
expected. Figure 146 and Figure 147 are the screen captures of eQuest hourly report for the
system and each zone.

FIGURE 146:
EQUEST HOURLY
REPORT VARIABLES FOR
THE PACKAGED SYSTEM

FIGURE 147:
EQUEST HOURLY
REPORT VARIABLES FOR
EACH ZONE

APPENDIX 8 – HOW TO MODEL DIFFERENT VAV ZONE CONTROLS IN DOE-2.2 X223X

 Note that the heating extraction rate in Figure 147 is the actual energy extracted from or
added to the zone (cooling is positive). The difference between heating coil energy and
heating extraction energy is the energy used to reheat the supply air from the supply
temperature to the zone temperature.

Figure 148 shows the eQuest simulation results of the single maximum control sequence for
one zone of under a five zone packaged VAV system.

VAV flow ratio is calculated by dividing Zone supply air flow rate (cfm) hourly report value
with Zone design air flow rate from SV-A report (as shown in Figure 160).

Zone temperature is from Current hour zone temp (deg F) hourly report value.

Terminal coil output is from Zone coil heat (Btu/hr) hourly report value.

As shown in Figure 148a, in cooling mode, the VAV damper position is adjusted according
to zone load (the highest zone load is at a zone temperature of 75.5oF). In deadband and
heatingmode, the damper was kept at minimum position, i.e. 30%. Figure 148b confirms
that the zone airflow is always at minimum when the heating coil output is greater than zero.

X224X APPENDIX 8 – HOW TO MODEL DIFFERENT VAV ZONE CONTROLS IN DOE-2.2

 FIGURE 148:
(a)
SIMULATED SOUTH
90% ZONE SINGLE MAX

80% CONTROL
Single maximum
70%
VAV Flow Ratio [%]

60%

50%

40%

30%

20%

10%

0%
65 66 67 68 69 70 71 72 73 74 75 76
Zone Temperature [F]

(b)

90%

80% Single maximum

70%
VAV flow ratio [%]

60%

50%

40%

30%

20%

10%

0%
0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000
Heating coil output [Btu/hr]

APPENDIX 8 – HOW TO MODEL DIFFERENT VAV ZONE CONTROLS IN DOE-2.2 X225X

 Dual Maximum-VAV Heating

Control Sequence

Figure 149 illustrates a dual maximum zone control sequence with variable air volume in
heating. Airflow is reduced from cooling maximum airflow to minimum airflow as cooling
load reduces; as the zone enters into heating mode the discharge air temperature setpoint is
reset from minimum temperature (e.g. 55oF) to maximum temperature allowed by reheat
coil’s capacity, in this case, 95 °F. If more heating is required then the airflow is reset from
the minimum up to the heating maximum. With a dual maximum zone control sequence,
the airflow in deadband is lower than the airflow at full heating. The minimum flow needs
to only be high enough to satisfy the ventilation requirements, which are can be 10% or less
for perimeter zones.

FIGURE 149:
DUAL MAXIMUM WITH Max Cooling
90oF Airflow Setpoint
VAV HEATING

Supply Air Temperature Setpoint

(requires discharge temp. sensor)

50%

Airflow Setpoint

20%

Heating Loop Cooling Loop

Dead
Band

eQuest Modeling Guidance

Figure 150 and Figure 151 show key zone inputs in eQuest to model the dual maximum
control sequence. Highlights of the inputs including the following:

ƒ Thermostat type under each Zone command should be Reverse Action. This allows the
VAV damper to open from minimum position to increase air flow rate in heating mode.

ƒ Minimum Flow Ratio is the design zone flow ratio in deadband and low heating mode.
In dual maximum control sequence, this minimum flow ratio is less than the air flow
needed at full heating and needs only to be high enough to satisfy ventilation needs.

ƒ Throttling range should be set to a fairly small value to accurately represent DDC
control.

X226X APPENDIX 8 – HOW TO MODEL DIFFERENT VAV ZONE CONTROLS IN DOE-2.2

ƒ Reheat Delta T is the maximum increase in temperature for supply air passing through
the zone (or subzone) reheat coils. It is set at the system level even though it applies to
the zone reheat delta T. It must be greater than zero for terminal reheating VAV boxes
to be enabled. Allowed input is between 0 °F and 100 °F. Typical design T through
terminal reheat coils are 15 °F to 40 °F. CAUTION: When comparing single maximum
(proportional control) and dual maximum (reverse action) it is important to recognize
that if both runs have the same reheat delta T then the single maximum may have more
hours unsatisfied because the heating capacity is limited by the min flow ratio and the
reheat delta T. The dual maximum however, is not limited by the min flow ratio (it will
allow the airflow to go all the way up to the cooling maximum in heating mode).
Therefore it is important to check the hours unsatisfied of both runs and it may be
necessary to raise the min flow ratio or the reheat delta T for the single maximum case
in order to get an “apples to apples” comparison in terms of hours satisfied.

FIGURE 150:
EQUEST INPUTS TO
MODEL DUAL MAXIMUM
CONTROL

When using Reverse Action, the default zone control sequence in DOE-2.2 in heating mode
is to first reset the supply air temperature to the zone before increasing the airflow, i.e.
“temperature first” control. DOE-2.2 does not allow other control sequences such as
“airflow first” or “airflow and temperature” together. Note that there are keywords for air
flow first, temperature first or simultaneous, that define the control sequence between air
flow rate and temperature under the System command, but these keywords apply only to
Systems-level COOL-CONTROL keywork and not to Zones. Fortunately, DOE-2.2’s
default “temperature first” control strategy happens to be the same sequence in heating mode
as depicted in Figure 149, i.e. in heating mode increase discharging temperature first by

APPENDIX 8 – HOW TO MODEL DIFFERENT VAV ZONE CONTROLS IN DOE-2.2 X227X

 opening terminal heating hot water valve before increasing discharging air flowrate by
opening VAV damper.

Unfortunately DOE-2.2 does not have a keyword for heating maximum flow rate. With a
reverse acting thermostat it will allow the airflow rate in heating to go all the way up to the
cooling maximum. Therefore to determine how accurately DOE-2.2 is modeling the desired
control sequence it may be necessary to run hourly reports of zone flow versus heating coil
output to see if the model is reasonably accurate.

FIGURE 151:
EQUEST INPUT TO
ENABLE REHEAT COILS

eQuest Simulation Validation

Figure 152 shows eQuest model results of the dual maximum control sequence for one zone
of a five zone packaged VAV system.

X228X APPENDIX 8 – HOW TO MODEL DIFFERENT VAV ZONE CONTROLS IN DOE-2.2

 FIGURE 152:
(a)
SIMULATED DUAL

1 MAXIMUM CONTROL

0.9

0.8
Dual maxium with VAV
0.7
VAV Flow Ratio [%]

heating
0.6

0.5

0.4

0.3

0.2

0.1

0
67 68 69 70 71 72 73 74 75 76
Zone Temperature [F]

(b)

100% 120

90%
100
Calculated zone discharge air

80%

70%
VAV flow ratio [%]

80
temperature [F]

60%

50% 60

40%
VAV flow ratio
Calculated zone discharge air temperature [F] 40
30%

20%
20
10%

0% 0
0 20000 40000 60000 80000 100000 120000 140000
Reheat coil output [Btu/hr]

As shown by these figures, the eQuest simulation of the control sequence represents the
proposed dual max control sequence reasonably well.

APPENDIX 8 – HOW TO MODEL DIFFERENT VAV ZONE CONTROLS IN DOE-2.2 X229X

 Figure 152a shows that as the cooling load of the space decreased from the peak load the
VAV damper closes down, until it reached cooling minimum, in this case 20%. In
deadband, the damper stays at cooling minimum, i.e. 20%. In heating mode, the VAV
damper opens up and it appears that the airflow increases immediately but Figure 152b
confirms that it is in fact using a “temperature first” sequence. The damper does not begin
to open until the supply air temperature has been reset up to the maximum (system supply
air temperature + reheat delta T). Zone supply air temperature is not an hourly report that is
available from eQuest. It can be calculated based on the zone flow ratio, the reheat coil
output, the system supply air temperature, and the design zone flow (from the SV-A report).

Figure 152 also shows that airflow in heating goes all the way up to 100% but it is not clear
from this figure how many hours the flow is above 50% in heating. A histogram of this data
(Figure 153) confirms that there are relatively few hours where the flow rate is above 50% in
heating. For example, in this case, the number of hours in heating mode that the VAV
damper opens to above 50% (desired heating maximum) is 22 hours out of 3263 operating
hours (Note that the positive airflow in this figure represents times when the heating coil
output is zero and the negative flows represent positive airflow when the heating coil is
active).

FIGURE 153:
HISTOGRAM OF DAMPER
Mor
POSITION OF
1
SIMULATED DUAL 0.9
0.8 Dual maximum VAV heating
MAXIMUM CONTROL Cooling
0.7
SEQUENCE 0.6 mode
0.5
0.4
VAV flow ratio bin

0.3
0.2
0.1
0
-0.1
-0.2
-0.3 Heating
-0.4
-0.5 mode
-0.6
-0.7
-0.8
-0.9
-1
0 200 400 600 800 1000 1200 1400 1600 1800
# of hours

X230X APPENDIX 8 – HOW TO MODEL DIFFERENT VAV ZONE CONTROLS IN DOE-2.2

 FIGURE 154:
More HISTOGRAM OF FAN

1 FLOW RATIO IN

0.9 SIMULATION OF DUAL

Dual maximum VAV heating MAXIMUM CONTROL
0.8

0.7 SEQUENCE
Fan flow ratio bin

0.6

0.5

0.4

0.3

0.2

0.1

0
0 500 1000 1500 2000 2500
# of hours

Figure 155 shows the histogram of fan operation point. It can be seen that 1909 out of 3263
hours (59%) the fan operates at 30% or below range.

Dual Maximum with Constant Volume Heating

Control Sequence

Figure 156 shows the control sequence of dual maximum with constant volume heating. As
cooling load decreases the airflow is reduced from the maximum airflow (on the far right side
of the figure) down to the minimum flow; as heating is required the airflow is first reset from
the minimum to the heating maximum and then the reheat valve is modulated to maintain
the space temperature at setpoint.

Maximum FIGURE 155:

Airflow
Setpoint HISTOGRAM OF FAN
Reheat Valve Position OPERATING POINT
UNDER DUAL MAXIMUM
CONTROL

50%
Airflow
Setpoint

20%

Heating Loop Cooling Loop

Dead
Band

APPENDIX 8 – HOW TO MODEL DIFFERENT VAV ZONE CONTROLS IN DOE-2.2 X231X

 eQuest Modeling Inputs

Figure 157 demonstrates key inputs in eQuest to model the single maximum control
sequence. Highlights of the inputs including the following:

ƒ Thermostat type under each Zone command should be Proportional; this ensures the
VAV airflow rate in deadband mode and heating mode to be constant.

ƒ Throttling range input is small for digital thermostat

ƒ Under Zone command, Cooling Minimum Flow is the design VAV flow ratio at
deadband and low cooling mode; cooling minimum flow can be smaller and only needs
to satisfy the ventilation requirement; Heating Minimum Flow is the designed VAV flow
ratio in heating mode.

FIGURE 156:
DUAL MAXIMUM WITH
CV HEATING

X232X APPENDIX 8 – HOW TO MODEL DIFFERENT VAV ZONE CONTROLS IN DOE-2.2

 FIGURE 157:
EQUEST INPUTS FOR
MODELING DUAL
MAXIMUM WITH
CONSTANT HEATING
CONTROL

eQuest Simulation Validation

Figure 158 shows eQuest model results of the dual maximum with constant heating control
sequence for one zone of a five zone packaged VAV system.

APPENDIX 8 – HOW TO MODEL DIFFERENT VAV ZONE CONTROLS IN DOE-2.2 X233X

 FIGURE 158: (a)
SIMULATED DUAL
MAXIMUM WITH CV 90%
HEATING FOR SOUTH 80%
ZONE
70%

VAV Flow Ratio [%]

60%

50%

40%

30%

20%

10%

0%
67 68 69 70 71 72 73 74 75 76
Zone Temperature [F]

(b)

100%

90%

80%

70%
VAV flow ratio [%]

60%

50%

40%

30%

20%

10%

0%
0 20000 40000 60000 80000 100000 120000
Heating coil output [Btu/hr]

Tips for Accurate Modeling of Zone Controls

Realistic Schedules

In order to capture the true energy savings of lower zone minimum flows it is necessary to
have realistic hourly schedules for occupancy, lighting and equipment. A schedule that uses

X234X APPENDIX 8 – HOW TO MODEL DIFFERENT VAV ZONE CONTROLS IN DOE-2.2

averaged values for all hours will not be accurate. It must have at least some hours with low
loads to capture periods where the zone is in deadband. Figure 159 is a sample realistic day
schedule that could apply to people, lights or equipment. It is also important to have some
randomness across zones. If all zones have the same schedule at every hour then the system
flow and load will not necessarily be realistic. For example, if a system had supply air
temperature reset and all zones were at very low load then the supply air temperature could
be reset much higher than might be realistic. An unrealistically high supply air temperature
would incorrectly reduce potential savings of low minimum zone flow rates.

(a) FIGURE 159:

1.000 TYPICAL ZONE
0.900
0.800 SCHEDULE
0.700
0.600
0.500
0.400
0.300
0.200
0.100
0.000
1- 2AM
2-3am
3-4am
4-5am
5-6am
6-7am
7-8am
8-9am
9-10am
10-11am

noon-1pm
1-2pm
2-3pm
3-4pm
4-5pm
5-6pm
6-7pm
7-8pm
8-9pm
9-10pm
10-11pm
Midnight -1Am

11-noon

11-midnight

Autosizing in DOE-2.2

How oversized or undersized a zone is will have a significant impact on the number of hours
a zone is at minimum flow and therefore it is important to pay attention to zone sizing when
comparing zone control sequences.

eQuest can auto-size both the system design flow rate and the zone design flow rate.
However, DOE-2.2 is not particularly good at autosizing and it often calculates flow rates
that are too low such that many hours are unsatisfied or too high such that systems and zones
are oversized and the hourly flows never approach the design flows. Therefore, it is
important to check the SS-R and SS-F reports to see how well sized the zones are. If they are
not well sized then they can be manually sized (using DOE-2 estimated flows from the SV-A
report as a starting point) or the DOE-2 sizing ratio can be increased or decreased. The
sizing ratio (set at the system level) applies to both the system and zone and it applies not
only to auto-sized values but also manually sized values.

For example, Report SS-R in Figure 17 indicates that the maximum flow ratio in the South
Perim Zn is less than 70% of design airflow. Therefore, the supply flow to this zone can be
reduced accordingly.

APPENDIX 8 – HOW TO MODEL DIFFERENT VAV ZONE CONTROLS IN DOE-2.2 X235X

 FIGURE 160:
SAMPLE SV-A REPORT
FOR SYSTEM SIZING

FIGURE 161:
SAMPLE SS-R REPORT
TO CHECK SYSTEM
OVERSIZING

X236X APPENDIX 8 – HOW TO MODEL DIFFERENT VAV ZONE CONTROLS IN DOE-2.2

 APPENDIX 9: SIMULATED PERFORMANCE OF
THREE ZONE CONTROL SEQUENCES

Summary

DOE-2.2 was used to compare the energy performance of three zone control sequences:
Single Maximum, Dual Maximum with VAV Heating and Dual Maximum with Constant
Volume Heating. These three sequences are described in detail in Appendix 9.

The Single Maximum is modeled with a 30% minimum, which is generally the highest
minimum allowed by Title 24. 30% however, is often not high enough to meet peak
heating loads with a supply air that is low enough to prevent stratification and allow good
mixing. One of the main reasons for a dual maximum sequence, is that it can allow a higher
airflow in heating to prevent stratification without incurring higher reheat energy in
deadband. Therefore, the dual maximum sequences are both modeled with a 50%
maximum airflow in heating and a 20% airflow in deadband. 50% is generally sufficient in
most new buildings in California to meet peak heating loads at low enough supply air
temperatures to prevent stratification.

The basecase model is a typical office building in Sacramento with a packaged VAV and hot
water reheat system. This model was also run in San Francisco, Los Angeles, Chicago and
Atlanta. Numerous parametric analyses were also run to determine the impact of supply air
temperature reset, of single maximum sequences with 40% and 50% minimums, of
oversized zones, of systems that are left running 24/7 and of very lightly loaded buildings.

In the basecase model the Dual Max-VAV saved 5 cents/ft2-yr compared to the single
maximum but the Dual Max-Constant Volume actually used 2 cents/ft2-yr more energy than
the Single Maximum case even though it has a lower flow in deadband (20% versus 30%).
As shown in Figure 162, the Dual Max-VAV savings go down if supply air temperature reset
is employed and go up if the zones are oversized, if the fan runs 24/7 or if the minimum flow
for the Single Maximum sequence is higher than 30%. It is estimated that the average
savings of the Dual Max-VAV sequence for a typical office building would be approximately
10 cents/ft2-yr. The Dual Max-Constant Volume never saves as much as the Dual Max-
VAV and in many cases uses more energy than the Single Maximum. On average, the Dual
Maximum-CV is no more efficient than the Single Maximum control sequence. This is
basically because the higher heating airflow rate for the Dual Maximum-CV causes it to get
“stuck” in heating. Once in heating mode, the cooling load must exceed 50% of peak
cooling load to “unstick” the zone from heating mode.

APPENDIX 9: SIMULATED PERFORMANCE OF THREE ZONE CONTROL SEQUENCES X237X

 FIGURE 162:
ANNUAL UTILITY COST
SAVINGS

Schematics of Modeled Zone Control Sequences

FIGURE 163: Maximum

Airflow Setpoint
SINGLE MAXIMUM ZONE
Reheat Valve Position
CONTROL SEQUENCE

Airflow Setpoint

30%

Heating Loop Cooling Loop

Dead
Band

X238X APPENDIX 9: SIMULATED PERFORMANCE OF THREE ZONE CONTROL SEQUENCES

 FIGURE 164:
Max Cooling
o
90 F Airflow Setpoint DUAL MAXIMUM WITH
VAV HEATING –
Supply Air Temperature Setpoint
(requires discharge temp. sensor) TEMPERATURE FIRST

50%

Airflow Setpoint

20%

Heating Loop Cooling Loop

Dead
Band

Maximum FIGURE 165:

Airflow Setpoint
DUAL MAXIMUM WITH
Reheat Valve Position CV HEATING

50%
Airflow Setpoint

20%

Heating Loop Cooling Loop

Dead
Band

Model Description

A 10,000 ft2 single story five zone office building with 15 ft deep perimeter zones was
modeled in eQuest to evaluate annual energy performance of the three zone control
sequences. The building envelope consists of R-19 metal frame roof and R-13 metal frame
wall with 40% window wall ratio. All windows use double pane glazing. The U value of the
glass is 0.47 and the SHGC value of the glass is 0.31 for non-north facing windows and 0.47
for north facing windows.

The building was modeled to be occupied from 7:00 am to 7:00 pm Monday through Friday
and was closed on Saturday, Sunday and holidays. Building internal loads consist of an

APPENDIX 9: SIMULATED PERFORMANCE OF THREE ZONE CONTROL SEQUENCES X239X

 average 100 ft2 per person occupancy density, 1.3 w/ ft2 lighting power densities and 1.5 w/
ft2 equipment power density.

In order to simulate “real-life” building operation, five occupancy day schedules were
modeled as shown in Figure 166. The simulation models were set up such that on any
weekday, each of the five zones uses one of the schedules shown in Figure 166 and no two
zones use the same schedule on the same day. From Monday to Friday, each zone uses a
different day schedule on a different day. Lighting and equipment schedule are the same as
the occupancy schedule. This is a simplifying assumption which assumes that people turn off
lights in proportion to occupancy, which might not be the case with multi-occupant spaces.
Similarly, equipment loads may not drop off as rapidly as occupancy. On the other hand,
these schedules are somewhat arbitrary so having separate schedules for lighting and
equipment does not necessarily make sense. Furthermore, parametrics were performed to
determine the sensitivity of the results to changes in schedules.
(a) (c)
1.000 1.000
FIGURE 166: 0.900 0.900
0.800 0.800
OCCUPANCY SCHEDULE 0.700 0.700
0.600 0.600
0.500 0.500
0.400 0.400
0.300 0.300
0.200 0.200
0.100 0.100
0.000 0.000
Midnight -1Am
1- 2AM
2-3am
3-4am
4-5am
5-6am
6-7am
7-8am
8-9am
9-10am
10-11am

noon-1pm
1-2pm
2-3pm
3-4pm
4-5pm
5-6pm
6-7pm
7-8pm
8-9pm
9-10pm
10-11pm

Midnight -1Am
1- 2AM
2-3am
3-4am
4-5am
5-6am
6-7am
7-8am
8-9am
9-10am
10-11am

noon-1pm
1-2pm
2-3pm
3-4pm
4-5pm
5-6pm
6-7pm
7-8pm
8-9pm
9-10pm
10-11pm
11-noon

11-midnight

11-noon

11-midnight
(b) (d)
1.000 1.000
0.900 0.900
0.800 0.800
0.700 0.700
0.600 0.600
0.500 0.500
0.400 0.400
0.300 0.300
0.200 0.200
0.100 0.100
0.000 0.000
11-noon
Midnight -1Am
1- 2AM
2-3am
3-4am
4-5am
5-6am
6-7am
7-8am
8-9am
9-10am
10-11am

noon-1pm
1-2pm
2-3pm
3-4pm
4-5pm
5-6pm
6-7pm
7-8pm
8-9pm
9-10pm
10-11pm
11-midnight
1- 2AM
2-3am
3-4am
4-5am
5-6am
6-7am
7-8am
8-9am

noon-1pm
1-2pm
2-3pm
3-4pm
4-5pm
5-6pm
6-7pm
7-8pm
8-9pm
9-10pm
10-11pm
Midnight -1Am

9-10am
10-11am

11-midnight
11-noon

(e)
1.000
0.900
0.800
0.700
0.600
0.500
0.400
0.300
0.200
0.100
0.000
Midnight -1Am
1- 2AM
2-3am
3-4am
4-5am
5-6am
6-7am
7-8am
8-9am
9-10am
10-11am

1-2pm
2-3pm
3-4pm
4-5pm
5-6pm
6-7pm
7-8pm
8-9pm
9-10pm
10-11pm
noon-1pm
11-noon

11-midnight

The building is conditioned by a packaged VAV system with hot water reheats at VAV
boxes. Room temperature setpoint are 75/82 for cooling and 70/64 for heating during
occupied/unoccupied hours. The HVAC system runs from one hour before occupancy to
one hour after occupancy. System supply air temperature is fixed at 55oF in the basecase. A

X240X APPENDIX 9: SIMULATED PERFORMANCE OF THREE ZONE CONTROL SEQUENCES

DOE-2 fan curve that represents a variable speed drive and demand-based static pressure
reset was used for all runs.

The model was run using the weather data representing Sacramento, CA (climate zone 12)
which for California has relatively hot summers and relatively cold winters.

Three VAV control sequences were investigated, the detailed modeling assumptions for each
of the control is are shown in Table 31 and in the screen captures from eQuest below.

APPENDIX 9: SIMULATED PERFORMANCE OF THREE ZONE CONTROL SEQUENCES X241X

 TABLE 31:
Dual max. Dual max.
BASECASE MODELING with CV with VAV
ASSUMPTIONS Case # Single max. heating heating
System Type PVAVS -ditto- -ditto-
Sizing Ratio 1.25 -ditto- -ditto-
Fan Control VSD -ditto- -ditto-
min Fan ratio = 0.1,
Air Flow -ditto- -ditto-
max Fan ratio = 1.1
Fan Eff. SA Fan 53%, RA Fan 53% -ditto- -ditto-
Fan Performance Curve VSD fan with SP reset
Fan static pressure 3.5 in. -ditto- -ditto-
OA ratio Default (calc. from zone OA CFM) -ditto- -ditto-
HVAC differential drybulb, max temperature
Economizer -ditto- -ditto-
limit = 59
System Cooling EIR 0.36 (9.5 EER) -ditto- -ditto-
Min SAT 55.°F -ditto- -ditto-
Max Cooling SAT Reset
59. °F -ditto- -ditto-
Temp
Cooling SAT temp control Constant -ditto- -ditto-
Heating SAT temp control Constant -ditto- -ditto-
No coil at packaged unit, only hot water
Heating Coil -ditto- -ditto-
reheating coil at each zone
RH Coil Valve 3-way valve -ditto- -ditto-
Min Heating Reset Temp 75. °F -ditto- -ditto-
Proportion
Thermostat Proportional ReverseAction
al
Throttling Range 0.1°F -ditto- -ditto-
Cooling Min Flow Ratio 30% 20% 20%
Cooling Max Flow Ratio 100% 100% 100%

Zone Heating Min Flow Ratio 30% 50% 20%

Heating Max Flow Ratio 30% 50% 100%
(each)
Cooling setpoint 75. °F -ditto- -ditto-
Heating setpoint 70. °F -ditto- -ditto-
Cooling setpoint
82. °F -ditto- -ditto-
unoccuppied
Heating setpoint
64. °F -ditto- -ditto-
unoccuppied
Boiler HIR 1.25 -ditto- -ditto-
Boiler Design HWST 180 °F -ditto- -ditto-

Plant Design HW loop dT 40 °F -ditto- -ditto-

HW loop pump control one speed pump -ditto- -ditto-
Exterior wall U value R-13 (code) -ditto- -ditto-
Roof U value R-19 (code) -ditto- -ditto-
Building WWR 40% -ditto- -ditto-
U = 0.47, SHGC = 0.31 (nonnorth),
Envelope Glass Type -ditto- -ditto-
0.47 (north)
100 ft by 100 ft, 15 ft perimeter zone
Area -ditto- -ditto-
depth
2
Occupancy 100 ft /person -ditto- -ditto-
2
Building Lighting 1.3 w/ ft -ditto- -ditto-
2
Internal Equipment 1.5 w/ ft -ditto- -ditto-

Load Occupied 7:00 ~19:00 M-F,

Schedule -ditto- -ditto-
Unoccupied other days

Climate
CZ 12 (Sacramento) -ditto- -ditto-
Zone

X242X APPENDIX 9: SIMULATED PERFORMANCE OF THREE ZONE CONTROL SEQUENCES

 FIGURE 167:
EQUEST PARAMETRIC
RUN INPUTS

Note that Thermal Zone parametric References in Figure 167 only refer to the Core Zone.
This is because the inputs for the perimeter zones are linked to the core zone. Note also that
the REHEAT-DELTA-T in the Single Maximum Basecase is set extremely high but is set
more reasonably in the Dual Maximum cases. This was done to insure that the heating loads
met by the models were to same. Since the Single Maximum has a lower heating airflow
rate, it cannot meet the peak heating loads unless the supply air temperature is allowed to be
unreasonably high. In order to make a fair “apples to apples” energy comparison this higher
REHEAT-DELTA-T is necessary. This also further illustrates the need to for the Dual
Maximum sequence: Single Maximum cannot meet the peak heating loads without
unreasonably high supply air temperatures.

Utility Rates

The PG&E Sch-A10a Electricity Rate was used in all runs. The virtual electricity rate
(including demand charges) was approximately $0.18/kwh for all runs.

The PG&E GNR-1 Gas Rate was used for all runs. The virtual gas rate was approximately
$0.60/therm.

Basecase Results

TABLE 32:
SUMMARY OF ENERGY
CONSUMPTION OF
THREE CONTROL
SEQUENCES

Energy consumption of the three control sequences are listed and compared in Table 32.
Results show that with current modeling assumptions, the Dual Maximum-VAV control
sequences saves about $0.05 $/ ft2/yr, 253.6 Wh/ ft2/yr and 2.9 kBtu/ ft2/yr compared with

APPENDIX 9: SIMULATED PERFORMANCE OF THREE ZONE CONTROL SEQUENCES X243X

 Single Maximum control. The Dual Maximum with CV Heating control used more energy
than the Single Maximum control. This is due to higher space heating, fan and pumping
energy from the higher flow rate in heating mode (50% vs. 30%).

Figure 168 through Figure 171 show the zone VAV damper control, reheating coil output
and fan flow histograms for the basecase scenario. It can be seen that the control sequence is
well simulated. From the histogram it is clear that there are very few hours where the airflow
is above the desired 50% in heating for the Dual Maximum-VAV sequence. Thus it
accurately represents the desired sequence (even though DOE-2 does not have a keyword for
heating maximum).

FIGURE 168: 100%

VAV FLOW RATIO AND
90%
HEATING FOR ONE ZONE Single maximum
80% Dual maximum with CV heating
Dual maximum with VAV heating
70%
VAV Flow Ratio [%]

60%

50%

40%

30%

20%

10%

0%
65 66 67 68 69 70 71 72 73 74 75 76
Zone Temperature [F]

X244X APPENDIX 9: SIMULATED PERFORMANCE OF THREE ZONE CONTROL SEQUENCES

 100%
FIGURE 169:

Single maximum VAV FLOW CONTROL

90%
Dual maximum CV heating FOR ONE ZONE
80% Dual maximum with VAV heating

70%
VAV flow ratio [%]

60%

50%

40%

30%

20%

10%

0%
0 20000 40000 60000 80000 100000 120000 140000
Heating coil output [Btu/hr]

More
FIGURE 170:
1 VAV FLOWRATE
0.9 HISTOGRAM FOR ONE
Dual maximum VAV heating

0.8
Dual maximum CV heating ZONE
Single maximum
0.7
Fan flow ratio bin

0.6

0.5

0.4

0.3

0.2

0.1

0
0 500 1000 1500 2000 2500 3000
# of hours

APPENDIX 9: SIMULATED PERFORMANCE OF THREE ZONE CONTROL SEQUENCES X245X

 More
FIGURE 171:
1
FAN FLOW HISTOGRAM
0.9 Dual maximum VAV heating
Dual maximum CV heating
0.8
Single maximum
0.7

Fan flow ratio bin

0.6

0.5

0.4

0.3

0.2

0.1

0
0 500 1000 1500 2000 2500 3000
# of hours

Parametric Analysis

Parametric runs were carried out to evaluate the energy saving potential of different VAV
control sequences under different scenarios. Descriptions of the parametric runs are as
following.
1. Basecase. Base case runs are described above in Basecase Results;

2. SAT Reset. Supply air temperature reset control was enabled for all three VAV
control sequences. Cooling supply air temperature resets based on the
“warmest” zones and can be reset from 55oF up to 59°F. In the basecase it is
fixed at 55oF. The maximum supply air temperature is limited to 59oF (as
opposed to say 65oF) because DOE-2 seems to almost always peg the supply air
temperature at the maximum allowed which is not necessarily realistic.

3. 40% Single Minimum. The minimum VAV flow ratio for single maximum
control sequence is 40%. While this does not generally comply with code it is
often done in practice in order to meet design heating loads at reasonable
supply air temperatures which limit stratification.

4. 50% Single Minimum. The minimum VAV flow ratio for single maximum
control sequence is 50%. Similarly, this is commonly seen in the field.

5. Oversized Zones. The DOE-2.2 sizing ratio was set to 200% for all three
control sequences compared to 125% in the basecase.

6. 24/7 Operation. The HVAC system runs 24-hours a day and 7 days a week.
The cooling temperature setpoint is kept at 75°F, heating temperature setpoint
is kept at 70°F 24 hours a day and 7 days a week;

X246X APPENDIX 9: SIMULATED PERFORMANCE OF THREE ZONE CONTROL SEQUENCES

 0.000
0.100
0.200
0.300
0.400
0.500
0.600
0.700
0.800
0.900
1.000
0.000
0.100
0.200
0.300
0.400
0.500
0.600
0.700
0.800
0.900
1.000
Midnight -1Am Midnight -1Am
1- 2AM 1- 2AM
7.

2-3am 2-3am
3-4am 3-4am
4-5am 4-5am
5-6am 5-6am
6-7am 6-7am
7-8am 7-8am
8-9am 8-9am
9-10am 9-10am
10-11am

(c)
10-11am
(a)

11-noon 11-noon
noon-1pm noon-1pm
1-2pm 1-2pm

0.000
0.100
0.200
0.300
0.400
0.500
0.600
0.700
0.800
0.900
1.000
2-3pm 2-3pm
Midnight -1Am 3-4pm 3-4pm
1- 2AM 4-5pm 4-5pm
2-3am 5-6pm 5-6pm
3-4am 6-7pm 6-7pm
4-5am 7-8pm 7-8pm
5-6am 8-9pm 8-9pm
6-7am 9-10pm 9-10pm
7-8am 10-11pm 10-11pm
8-9am 11-midnight 11-midnight
9-10am
10-11am

(e)
0.000
0.100
0.200
0.300
0.400
0.500
0.600
0.700
0.800
0.900
1.000

0.000
0.100
0.200
0.300
0.400
0.500
0.600
0.700
0.800
0.900
1.000
11-noon
noon-1pm Midnight -1Am Midnight -1Am
reduced profile as shown in Figure 172.

1-2pm 1- 2AM 1- 2AM

2-3pm 2-3am 2-3am
3-4pm 3-4am 3-4am
4-5pm 4-5am 4-5am
5-6pm 5-6am 5-6am
6-7pm 6-7am 6-7am
7-8pm 7-8am 7-8am
8-9pm 8-9am 8-9am
9-10pm 9-10am 9-10am
10-11pm 10-11am
(b)

10-11am
(d)

11-midnight 11-noon 11-noon

noon-1pm noon-1pm
1-2pm 1-2pm
2-3pm 2-3pm
3-4pm 3-4pm
4-5pm 4-5pm
5-6pm 5-6pm
6-7pm 6-7pm
7-8pm 7-8pm
8-9pm 8-9pm
9-10pm 9-10pm
10-11pm 10-11pm
11-midnight 11-midnight
Low Load Profile. The daily occupancy/lighting/equipment schedules have a

SCHEDULES
FIGURE 172:
LOW-LOAD DAY

APPENDIX 9: SIMULATED PERFORMANCE OF THREE ZONE CONTROL SEQUENCES X247X

 8. Worst Case Code Compliant. This is the combination of 5, 6 and 7
above. For all three control sequences, the HVAC system is sized to
be 200% of the load, systems runs 24/7, and uses low-load profile.

9. Worst Case. This is the same as 8 above except that 50% minimum
flow rate ratio is used in the single maximum control.

10. Different locations. Four different locations are analyzed besides

Sacramento in the basecase. These are: San Francisco, Los Angeles,
Chicago and Atlanta.

Table 33 shows how the parametric runs are organized. Note that X means run all three
control sequences. The cells highlighted in yellow are illustrated in more detail below.

TABLE 33: Sacramento San Francisco Los Angeles Chicago Atlanta

PARAMETRIC RUNS Basecase X X X X X
SAT reset X
40% single min X
50% single min X
Oversize HVAC X
24/7 X
low load profile X
worst case code
compliant: oversize
HVAC, 24/7, low load X X X X X
worst case: 50% single
min, oversize HVAC,
24/7, low load X

The results of the parametric runs are shown in Table 34.

X248X APPENDIX 9: SIMULATED PERFORMANCE OF THREE ZONE CONTROL SEQUENCES

 Electricity Natural Gas Annual Utility Costs:
Annual Peak Annual Peak ES-E ES-E ES-E ES-E Utility Electrical TABLE 34:
Energy Demand Energy Demand Total Total Electric Electric Cost Energy Natural Gas
Electric Fuel kWh kW Savings Savings Savings SUMMARY OF
kWH kW Therms Therms/Hr $ $ $ $ $/sf/yr Wh/sf/yr kBtu/sf/yr
Base case: Sizing Ratio = 1, No Temperature Reset at Cooling coil, Sacramento
Single Max 68223 55 1710 4.3 12768 1106 9605 2263 - - -
PARAMETRIC RUNS
Dual Max with CV Heating 68330 56 2092 4.8 12776 1323 9590 2286 ($0.02) (10.69) (3.82)
Dual Max with VAV Heating 65688 56 1419 5.2 12448 947 9274 2273 $0.05 253.59 2.91
Sizing Ratio = 1, Temperature Reset at Cooling Coil allow 55 ~ 59 °F SAT
Single Max 65560 56 1215 4.3 12457 834 9281 2276 - - -
Dual Max with CV Heating 65240 56 1332 4.8 12407 902 9224 2283 ($0.00) 32.08 (1.17)
Dual Max with VAV Heating 64251 56 1121 5.2 12291 783 9110 2281 $0.02 130.90 0.94
40% single min
Single Max 71941 55 2351 5.1 13264 1453 10108 2257 - - -
Dual Max with CV Heating 68330 56 2092 4.8 12776 1323 9590 2286 $0.06 361.02 2.60
Dual Max with VAV Heating 65688 56 1419 5.2 12448 947 9274 2273 $0.13 625.30 9.33
50% single min
Single Max 77225 54 3114 5.5 14010 1864 10850 2260 - - -
Dual Max with CV Heating 68330 56 2092 4.8 12776 1323 9590 2286 $0.18 889.49 10.23
Dual Max with VAV Heating 65688 56 1419 5.2 12448 947 9274 2273 $0.25 1153.77 16.96
Oversize sys
Single Max 76109 52 3012 5.5 13811 1808 10710 2201 - - -
Dual Max with CV Heating 77086 53 3825 6.1 13890 2270 10741 2249 ($0.05) (97.68) (8.13)
Dual Max with VAV Heating 70284 53 2226 6.2 13010 1386 9902 2207 $0.12 582.50 7.85
24/7
Single Max 89199 51 6486 2.0 15712 3669 12684 2128 - - -
Dual Max with CV Heating 95991 52 8907 2.6 16623 4987 13573 2150 ($0.22) (679.15) (24.21)
Dual Max with VAV Heating 79400 52 4904 1.9 14330 2828 11284 2146 $0.22 979.94 15.82
Low load
Single Max 45006 42 1981 3.9 8934 1255 6379 1655 - - -
Dual Max with CV Heating 45787 42 2457 4.4 9022 1524 6434 1688 ($0.04) (78.10) (4.76)
Dual Max with VAV Heating 42257 42 1709 4.7 8577 1109 6008 1669 $0.05 274.83 2.73
24/7, LOW-LOAD, OVERSIZE
Single Max 83426 36 10380 2.5 14482 5742 12011 1570 - - -
Dual Max with CV Heating 101622 34 15135 3.5 16904 8306 14481 1523 ($0.50) (1819.59) (47.55)
Dual Max with VAV Heating 69034 34 7723 2.3 12303 4328 9939 1464 $0.36 1439.26 26.58
24/7, LOW-LOAD, OVERSIZE,50% single min
Single Max 119672 44 17648 3.8 20058 9609 17189 1969 - - -
Dual Max with CV Heating 101622 34 15135 3.5 16904 8306 14481 1523 $0.45 1805.00 25.13
Dual Max with VAV Heating 69034 34 7723 2.3 12303 4328 9939 1464 $1.30 5063.85 99.25
San Francisco Base
Single Max 68223 55 1710 4.3 12768 1106 9605 2263 - - -
Dual Max with CV Heating 68330 56 2092 4.8 12776 1323 9590 2286 ($0.02) (10.69) (3.82)
Dual Max with VAV Heating 65688 56 1419 5.2 12448 947 9274 2273 $0.05 253.59 2.91
San Francisco-low load, 24/7, oversize
Single Max 71824 29 9404 2.1 12417 5190 10153 1364 - - -
Dual Max with CV Heating 90422 30 13743 3.0 14974 7513 12711 1363 ($0.49) (1859.87) (43.39)
Dual Max with VAV Heating 58834 27 7121 2.0 10466 3978 8317 1249 $0.32 1298.95 22.82
Los Angelos Base
Single Max 59790 40 1478 3.5 11019 967 8283 1836 - - -
Dual Max with CV Heating 59185 41 1703 3.9 10943 1095 8177 1866 ($0.01) 60.47 (2.25)
Dual Max with VAV Heating 56840 41 1199 4.3 10645 817 7882 1862 $0.05 294.92 2.79
Los Angelos-low load, 24/7, oversize
Single Max 88824 32 8792 2.1 14853 4859 12515 1438 - - -
Dual Max with CV Heating 111111 30 12880 3.0 17902 7046 15584 1419 ($0.52) (2228.72) (40.87)
Dual Max with VAV Heating 73374 29 6470 1.9 12567 3624 10363 1304 $0.35 1544.96 23.22
Chicago Base
Single Max 65378 41 1056 3.1 11843 736 9050 1893 - - -
Dual Max with CV Heating 63881 42 1083 3.6 11655 754 8837 1918 $0.02 149.67 (0.27)
Dual Max with VAV Heating 62312 42 794 3.8 11465 594 8647 1918 $0.05 306.55 2.62
Chicago-low load, 24/7, oversize
Single Max 71206 35 11610 3.3 12817 6436 10489 1428 - - -
Dual Max with CV Heating 87833 32 16076 3.5 15161 8828 12839 1422 ($0.47) (1662.69) (44.66)
Dual Max with VAV Heating 61692 32 10262 7.3 11251 5758 9032 1319 $0.22 951.39 13.47
Atlanta Base
Single Max 71454 51 1690 5.5 13201 1100 10117 2184 - - -
Dual Max with CV Heating 71349 51 1899 5.3 13183 1221 10082 2201 ($0.01) 10.58 (2.08)
Dual Max with VAV Heating 69506 51 1431 6.1 12960 959 9869 2192 $0.04 194.80 2.59
Atlanta-low load, 24/7, oversize
Single Max 98589 38 9284 2.5 16825 5169 14325 1600 - - -
Dual Max with CV Heating 117814 33 13401 3.3 19414 7392 16965 1549 ($0.48) (1922.46) (41.16)
Dual Max with VAV Heating 83880 33 7059 3.5 14576 3992 12214 1462 $0.34 1470.95 22.25

Figure 173 through Figure 176 shows the VAV flow control, reheating coil output and fan
flow histogram for the worst code compliance case in L.A (highlighted in yellow in Table
33).

Since the system is largely oversized, almost all the hours the airflow stays at minimum flow
ratio.

APPENDIX 9: SIMULATED PERFORMANCE OF THREE ZONE CONTROL SEQUENCES X249X

 FIGURE 173: 100%

ZONE AIR FLOW

Single maximum
90%
CONTROL FOR WORST
Dual maximum with CV heating
80% Dual maximum with VAV heating
CODE COMPLIANCE
CASE IN L.A. 70%

VAV Flow Ratio [%]

60%

50%

40%

30%

20%

10%

0%
69.5 70 70.5 71 71.5 72 72.5 73 73.5 74 74.5
Zone Temperature [F]

FIGURE 174:
60%
ZONE AIR FLOW RATE
AND HEATING COIL
50%
OUTPUT FOR WORST
CODE COMPLIANCE Single maxium
CASE IN L.A. 40% Dual maxium with CV heating
VAV flow ratio [%]

Dual maximum with VAV

30%

20%

10%

0%
0 5000 10000 15000 20000 25000 30000 35000 40000
Heating coil output [Btu/hr]

X250X APPENDIX 9: SIMULATED PERFORMANCE OF THREE ZONE CONTROL SEQUENCES

 FIGURE 175:
ZONE AIR FLOW
HISTOGRAM

More
FIGURE 176:
1 FAN FLOW HISTOGRAM

0.9
Dual maximum VAV heating
Dual maximum CV heating
0.8
Single maximum

0.7
Fan flow ratio bin

0.6

0.5

0.4

0.3

0.2

0.1

0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
# of hours

APPENDIX 9: SIMULATED PERFORMANCE OF THREE ZONE CONTROL SEQUENCES X251X

 APPENDIX 10: HOW TO MODEL DCV CONTROL IN EQUEST

This section explains how to model three different ventilation control sequences for
multizone systems with densely occupied zones. The three sequences are:

ƒ Fixed Minimum – The ventilation rate at the zones and at the system is fixed at the
design ventilation rate (i.e. based on peak occupancy)

ƒ DCV Using Sum of Zones – The ventilation rate at the zone is the larger of the area
based ventilation requirement (e.g. 0.15 CFM/ft2) and the occupant-based requirement
(e.g. 15 CFM/person), based on the actual number of occupants in each hour. The
System ventilation is the sum of the zone requirements. This sequence complies with
Title 24 2005.

ƒ DCV Using Critical Zone – This sequence is similar to the DCV Using Sum of Zones
except that the system OA fraction is equal to X/(1+X-Z), where X is the sum of the
current zone minimum flow rates and Z is the minimum flow rate of the critical zone,
i.e. the zone with the highest ventilation ratio. Thus the system ventilation requirement
will be higher than the sum of the zone requirements and less than or equal to the
ventilation rate of the critical zone. This is similar to the ASHRAE Standard 62-2004
multizone approach.

Fixed Minimum Ventilation Control

Figure 177 through Figure 182 are the eQuest screen captures that shows the key inputs at
the system level and zone level in order to model fixed minimum OA control. These inputs
are explained as following:

System Level Inputs

ƒ System Minimum Outside Air Ratio. This input should be left blank. If left blank,
there will not be a minimum OA flow restriction on the system and the system OA flow
will be calculated from the zone level OA flow inputs as explained later. Otherwise, if
the fraction number is larger than the OA flow requirement from zone level inputs, this
number take precedence; if the fraction number is smaller than the OA flow
requirement from zone level inputs, the zone level OA flow requirement is used to be
the actual system OA minimum flow;

ƒ Minimum OA Control method. Set this to “Fraction of Design Flow”. This keyword
determines during hourly simulation, how the system OA flow rate is controlled. This,
together with other keywords mentioned below, ensures the system OA flow rate to be
always the design flow rate, i.e. when system air flow rate decreases, system OA fraction
increases. This assumes that the air handler measures the outdoor airflow, or has
another means to ensure that outdoor airflow is relatively constant.

X252X APPENDIX 10: HOW TO MODEL DCV CONTROL IN EQUEST

ƒ Minimum OA Sizing method. Set this to “Sum of Zone OA” when complying with
Title 24. This keyword determines how the system design OA flow rate is calculated.
There are two choices available: “Sum of Zone OA” or “Set by Critical Zone”. “Sum of
Zone OA” sums the OA requirement from each zone to be the system design OA
requirement. To model ASHRAE Standard 62-2004 OA control, use “Set by Critical
Zone” which scales up the sums of the OA requirement from each zone by the zone
with the largest design OA fraction to be the system design OA requirement.

FIGURE 177:
EQUEST SYSTEM KEY
INPUTS TO MODEL
FIXED MINIMUM OA
CONTROL

Zone Level Inputs

Air Flow Tab

ƒ Zone Minimum Flow Control. Set this keyword to “Fixed/Scheduled” so that the zone
minimum air flow rate will not reset according to zone occupancy.

APPENDIX 10: HOW TO MODEL DCV CONTROL IN EQUEST X253X

 ƒ Zone Minimum Flow Ratio. This should be set to 10-30% for less densely occupied
spaces and left blank for more densely occupied spaces. For example, if you were using
a Dual Maximum Control sequence in a typical office space then this could be set to
about 15% (the lowest controllable flow), but if the zone is a conference room (high
occupant density) then it should be left blank. It is important to understand the
interaction between this keyword and the OA CFM/person and OA CFM/ft2 keywords
on the Zone Outside Air tab. Ideally you would like DOE-2 to set the minimum flow
rate to the maximum of these three: Min Flow Ratio, OA CFM/person and OA
CFM/ft2. Unfortunately DOE-2 does not do this. Min Flow Ratio always takes
precedence over the other two, even if it results in a smaller CFM. For example, if the
zone design total air flow rate is 1200 CFM, design OA flow rate (calculated by using
the larger of 15 cfm/person or 0.15 cfm/ ft2) is 600 CFM, i.e. 50% of design flow, with
Minimum Flow control being “Fixed/Scheduled”, if input for this keyword is 0.3, then
360 CFM (30% of 1200 CFM) instead of 600 CFM is going to be the zone minimum
flow rate in simulation. Furthermore, the Min Flow Ratio depends on the zone load
(e.g. 1 CFM/ft2 versus 2 CFM/ft2) so there is no simple rule of thumb based on
occupant density for when to specify the Zone Flow Ratio and when to leave it blank.
Therefore, the only way to absolutely ensure that DOE-2 is calculating the zone
minimum flow rate correctly is to check design flow rate in the SV-A report and
calculate for yourself which is larger: the desired minimum flow ratio or the required
ventilation. If it is the former then use this keyword to set it accordingly, otherwise
leave this keyword blank.

FIGURE 178:
AIR FLOW TAB FOR
TYPICAL ZONES

X254X APPENDIX 10: HOW TO MODEL DCV CONTROL IN EQUEST

 FIGURE 179:
AIR FLOW TAB FOR
DENSELY OCCUPIED
ZONES

Outdoor Air Tab

ƒ OA Air Flow in CFM. Always leave this blank.

ƒ Zone OA Flow per Person and Zone OA Flow per Area. When the OA air flow
keyword is left blank, eQuest automatically uses the one that yields larger flow rate to be
the zone minimum flow rate; when the OA air flow keyword is not blank, the air flow
input overrides inputs of these two keywords.

FIGURE 180:
AIR FLOW TAB FOR
NON-DENSELY
OCCUPIED ZONES

APPENDIX 10: HOW TO MODEL DCV CONTROL IN EQUEST X255X

 Multi-zone DCV Using Sum of Zone Air Method

In this control the zone minimum OA flow rate is the larger of 15 cfm times the actual
number of occupants and 0.15 cfm/ft2. The system minimum OA flow rate is the sum of
zone minimum OA flow rates.

System Level Inputs

ƒ System Minimum Outside Air Ratio. This should be left blank. Ventilation inputs at
the zone level will then be used to determine the system ventilation.

ƒ Minimum OA Control method. Use “DCV return sensor” so that the system OA flow
rate is controlled to be the sum of zone minimum OA flow rates.

ƒ Minimum OA Sizing Method. Corresponding to the “DCV return sensor” choice

under Minimum OA control method above, use “Sum of Zone OA” so that the system
design OA flow rate is calculated to be the sum of zone minimum OA flow rates.

FIGURE 181:
EQUEST SYSTEM LEVEL
KEY INPUT TO MODEL
MULTI-ZONE DCV
USING SUM OF ZONE
METHOD

X256X APPENDIX 10: HOW TO MODEL DCV CONTROL IN EQUEST

Zone Level Inputs

Air Flow Tab

ƒ Zone minimum Flow Control. Use “DCV reset up/down” for densely occupied spaces
(i.e. zones with CO2 control) so that zone minimum flow is reset at each hour to be the
zone minimum OA flow rate that’s calculated from the larger of 15 cfm/per times
number of people in the zone at that hour or 0.15 cfm/ft2. For non-densely occupied
spaces (zones without CO2 control) use “Fixed/Scheduled” so that the Zone Minimum
Flow Ratio is not ignored.

ƒ Zone Minimum Flow Ratio. For zones without DCV, This should be set to the control
sequence minimum (e.g. 10-30%) for most zones to insure DOE-2 is calculating the
minimum flow rate correctly. When “DCV reset up/down” is used this keyword is
ignored. You will need to leave this keyword blank if the zone does not have DCV but
has a relatively high occupant density or a relatively low design load (e.g. less than 1.0
CFM/ft2.) See discussion of Zone Minimum Flow Ratio under Fixed Minimum section
above.

FIGURE 182:
ZONE AIR FLOW TAB
INPUTS TO MODEL
MULTI-ZONE DCV

Outdoor Air Tab

ƒ OA Air Flow in CFM. Always leave this blank.

ƒ Zone OA Flow per Person and Zone OA Flow per Area. Set these based on the
ventilation code at design conditions. Note that when the OA air flow keyword is not
blank it overwrites inputs of these two keywords.

APPENDIX 10: HOW TO MODEL DCV CONTROL IN EQUEST X257X

 FIGURE 183:
ZONE OUTSIDE AIR TAB
INPUTS TO MODEL
MULTI-ZONE DCV

Multi-zone DCV using Critical Zone Method

In this control the zone minimum air flow rate is reset the same way as in DCV Using Sum
of Zones (see section 0), i.e. the larger of 15 cfm times the actual number of occupants and
0.15 cfm/ft2. The system minimum OA flow rate, however, is typically larger than the sum
of zone minimum OA flow rates. If the system OA fraction from sum of zone minimum OA
flow rate is X and the zone with the largest minimum air flow rate fraction (that’s calculated
from zone min OA flow rate) is Z, then the system OA fraction is reset to be X/(1+X-Z).
Refer to the DOE-2 help files for further details.

System Level Inputs

The eQuest zone level key inputs to model this control are the same as for the DCV Using
Sum of Zones. The only difference is the system level inputs on minimum OA control
method and minimum OA sizing method as shown in Figure 184. They are explained as
following.

ƒ Minimum OA control method. Use DCV zone sensors so that the system minimum
OA flow rate is scaled up using critical zone minimum air flow rate.

X258X APPENDIX 10: HOW TO MODEL DCV CONTROL IN EQUEST

ƒ Minimum OA sizing method. Use “Set by Critical Zone” so that the system design
minimum OA flow rate is scaled up using critical zone design minimum air flow rate.
This only affects system design minimum OA CFM and will not impact simulation
results since the minimum OA control uses the “minimum OA control” keyword as
describe above.

FIGURE 184:
EQUEST SYSTEM LEVEL
KEY INPUT TO MODEL
MULTI-ZONE DCV BY
CRITICAL ZONE METHOD

Sample eQuest Simulation Hourly Output

To confirm that eQuest is modeling these three sequences as expected a simple 5 zone model
was simulated. The model includes an interior conference room zone with a design
occupancy of 20 ft2/person. Per the occupancy schedule, this zone is only about 50%
occupied. Figure 185 shows the zone level hourly report variables that were plotted to
confirm the sequences.

APPENDIX 10: HOW TO MODEL DCV CONTROL IN EQUEST X259X

 FIGURE 185:
ZONE HOURLY REPORT

Figure 186 shows conference room airflow and system OA airflow for the fixed minimum
sequence on a typical summer day, July 5th. The conference room ventilation requirement
is as high or higher than the cooling load so the supply flow to this zone is fixed at the
ventilation requirement. The system is in economizer in the morning then on a fixed
minimum OA after 10am.

FIGURE 186:
FIXED MINIMUM 7000

SIMULATION RESULTS (Economizer On) Conf. Room Supply CFM (Fixed Min)
6000
System OA CFM (Fixed Min)

5000
(Economizer Off)

4000
CFM

3000

2000

1000

0
7 8 9 10 11 12 13 14 15 16 17
Fan Operating Hours

Figure 187 is the same model with DCV Sum of Zones control. It can be seen that zone
minimum air flow rate is reset at each hour to be the zone minimum OA flow rate of that

X260X APPENDIX 10: HOW TO MODEL DCV CONTROL IN EQUEST

hour calculated from number of people. System minimum OA flow rates reset each hour to
be the sum of zone OA minimum flow rate. At hour 8 and 9, system is running in
economizer mode.

FIGURE 187:
4500
System OA CFM DCV SUM OF ZONES
4000 System OA required by # of people METHOD SIMULATION
Conf. Room Supply
3500
RESULTS
Conf. Room OA required by # people

3000

2500
CFM

2000

1500

1000

500

0
7 8 9 10 11 12 13 14 15 16 17
Fan operation hours

Figure 188 shows the DCV Critical Zone Method. The zone air flow is same as Figure 187.
System OA flow rate is larger than sum of zone OA flow rate because it is scaled up using the
critical zone OA flow fraction. Since the conference room zone (the zone shown in the
figures) is the critical zone and for many hours has 100% OA flow rate fraction, system OA
fraction is scaled up to be 100% for many hours, too.

6000 FIGURE 188:

System min. OA flowrate
System min. OA flowrate from #people DCV CRITICAL ZONE
Zone air flowrate METHOD RESULTS
5000
Zone min OA flowrate by #people

4000
CFM

3000

2000

1000

0
7 8 9 10 11 12 13 14 15 16 17
Fan operation hours

APPENDIX 10: HOW TO MODEL DCV CONTROL IN EQUEST X261X

 APPENDIX 11: ANALYSIS OF DCV IN DENSELY OCCUPIED
ZONE OF A VAV SYSTEM

Energy savings from two different DCV control sequences for densely occupied zones in a
multizone system were simulated using eQuest. The modeled building consists of 4
perimeter office zones (100 ft2/person) and one interior conference room (20 ft2/person). In
the base case model the conference room minimum ventilation is fixed at 15 CFM/person
and the perimeter office minimum ventilation is fixed at 0.15 CFM/ft2. The two DCV
control sequences are as follows:

ƒ DCV Using Sum of Zones – The ventilation rate in the conference room is the larger of
0.15 CFM/ ft2 and 15 CFM times the actual number of occupants in each hour. The
System ventilation is the sum of the zone requirements. This sequence complies with
Title 24 2005 ventilation requirements.

ƒ DCV Using Critical Zone – This sequence is similar to the DCV Using Sum of Zones
except that the system OA fraction is equal to X/(1+X-Z), where X is the sum of the
current zone minimum flow rates and Z is the minimum flow rate of the critical zone,
i.e. the zone with the highest ventilation ratio. Thus the system ventilation requirement
will be higher than the sum of the zone requirements and less than or equal to the
ventilation rate of the critical zone. This sequence is similar to the ASHRAE Standard
62-2004 multizone approach.

All 16 California Climate Zones were simulated. As shown in Figure 189, the energy savings
are about 3.25~4.75 $/ft2 for the DCV Using Sum of Zones sequence and about 2.4~3.8$/ft2
for the DCV Using Critical Zone sequence. These figures represent the 15 year net present
value of the energy savings 1. These savings are based on a model of a building that is 50%
conference rooms. Similar savings would not be expected for buildings with lower occupant
densities.

1
Savings calculated using Time-Dependent Valuation (TDV) which is the new method for valuing energy in the
performance approach in the 2005 Building Energy Efficiency Standards. Details of TDV can be found at CEC’s
website: http://energy.ca.gov/title24/2005standards/archive/rulemaking/documents/tdv/index.html

X262X APPENDIX 11: ANALYSIS OF DCV IN DENSELY OCCUPIED ZONE OF A VAV SYSTEM
 CZ16 DCV using Critical Zone Method FIGURE 189:
DCV using Sum of Zone method
CZ15 TOTAL TDV SAVINGS
CZ14 USING DCV
CZ13

CZ12

CZ11

CZ10

CZ09

CZ08

CZ07

CZ06

CZ05

CZ04

CZ03

CZ02

CZ01

0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00
Total TDV Savings [$/sf]

Basecase Modeling Assumptions

A 10,000 square foot, five zone, one story building was used for the analysis. The four
perimeter zones are modeled as offices with 100 ft2/person design load and the interior zone
is modeled as a conference room with 20 ft2/person design load. One package VAV unit
with hot water reheat serves all five zones.

Table 31 and Figure 190 to Figure 193 summarize the major assumptions common to all
three control sequences.

Base Case Building Model

TABLE 35:
Wall metal frame wall with R-19 insulation
Roof metal frame roof with R-18 insulation COMMON MODEL
Glazing 40% WWR ratio; North window: SHGC = 0.31, U = 0.47; North ASSUMPTIONS
glazing: SHGC = 0.47, U = 0.47
Occupant Density 100 ft2/person (=12.5 person) for perimeter zones; 20 ft2/person(= 245
person) for core zone (conference rooms)
Perimeter Zones Area 5100 ft2 (51% of total area)
Interior Zone Area 4900 ft2 (49% of total area)
Lighting 1.32 w/ft2
Equipment 1.2 w/ft2
Cooling EER 9.971
Boiler Efficiency 80%
Fan Efficiency 0.44 W/cfm

Figure 190 show the occupancy schedule used for perimeter zones and core zone respectively.

APPENDIX 11: ANALYSIS OF DCV IN DENSELY OCCUPIED ZONE OF A VAV SYSTEM X263X
 FIGURE 190: Building Occupancy Schedules
OCCUPANCY SCHEDULES 1
Perimeter Zone Occupancy
Schedule
0.9 Core Zone Occupancy
Schedule
0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

FIGURE 191:
ZONE AIR FLOW INPUTS

Note that Min Flow Ratio for the Core Zone (interior conference room) is left blank while it
is set to 30% for the perimeter office zones. This is because the Min Flow Ratio in DOE-2
takes priority over the OA Flow/Person and the OA Flow/Area. For the perimeter zones the
zone minimum flow rate is driven by the 30% minimum flow ratio but for the interior
conference room zone it is driven by the 15 CFM/person ventilation requirement. The
design air flow for the interior zone is approximately 0.75 CFM/ft2 (it varies slightly by
climate zone) so the design ventilation rate is approximately 100% of the design flow rate.

FIGURE 192:
ZONE OUTSIDE AIR
INPUTS

X264X APPENDIX 11: ANALYSIS OF DCV IN DENSELY OCCUPIED ZONE OF A VAV SYSTEM
 FIGURE 193:
SYSTEM OUTSIDE AIR
INPUTS

DCV Modeling Assumptions

Figure 194 shows how the zone and system inputs were modified to model the two DCV
control schemes. For more detail on the meaning of these keywords and inputs refer to the
DOE-2.2 Dictionary and to the Appendix on “Modeling DCV in EQUEST”.

FIGURE 194:
DCV MODELING INPUTS

To evaluate the energy impact of demand control ventilation, the control of the core zone
VAV box and the outside air damper of the core zone system are changed. Under demand
control ventilation, each hour, the model calculates the outside air requirement based on
15cfm/person and number of people in the zone from the occupancy schedule. The outside
air damper was controlled to maintain just enough outside air flowrate for that hour. The
VAV damper is controlled by the zone cooling/heating load indicated by the thermostat as
well as the CO2 sensor, whichever gives the larger flowrate requirement.

APPENDIX 11: ANALYSIS OF DCV IN DENSELY OCCUPIED ZONE OF A VAV SYSTEM X265X
 Same changes are made to perimeter zone VAV boxes. However, since the 100 ft2/person
occupancy density at 15 cfm/person is equivalent to 0.15 cfm/ft2, the DCV control doesn’t
change air flowrate for perimeter zones.

Figure 195 shows system supply airflow rate and building total electrical energy
consumptions on a typical summer day, July 5th, with and without DCV control. It can be
seen that, without DCV control, core zone supply airflow rate is maintained at a constant
value: the design OA flowrate, even when the zone is not fully occupied. As results, more fan
energy is consumed due to large air flowrate. More cooling energy is consumed to cool more
summer outside air. And more reheat energy is consumed because of a higher zone minimal
flowrate that’s defined by the high design OA flowrate. With DCV control, zone minimal
airflow rate can be reset each hour to provide just enough airflow for the number of people
in the zone at that hour. As a result, fan energy, cooling and heating energy are reduced.

4000 50
FIGURE 195:
COMPARISON OF 45
3500
AIRFLOW AND
40
ELECTRICAL ENERGY
3000
CONSUMPTION ON A 35
TYPICAL SUMMER DAY
2500
30

KWH
CFM

2000 25

20
1500

15
1000
DCV Controled Zone CFM 10
No DCV Control Zone CFM
500
DCV Controled Elec. Energy 5

No DCV Control Elec. Energy

0 0
7 8 9 10 11 12 13 14 15 16 17
Fan Operating Hours

Simulation Results

The Fixed Minimum control, multi-zone DCV with sum of zone method, and multi-zone
DCV with critical zone method are run under all 16 California climate zones. The Time
Dependent Valuation formula was used to estimate gas and electrical costs. Time-Dependent
Valuation (TDV) is the new method for valuing energy in the performance approach in the
2005 Building Energy Efficiency Standards. TDV is based on the cost for utilities to provide
the energy at different times. TDV gas series is a series of 8760 values each represents the
cost of natural gas in the $/therm for that particular hour. The same TDV gas series is
applied to all 16 climate zones; TDV electricity series are 16 series of 8760 values, each

X266X APPENDIX 11: ANALYSIS OF DCV IN DENSELY OCCUPIED ZONE OF A VAV SYSTEM
represent the cost of electricity in $/kWH for that particular hour in that particular climate
zone. TDV rates are based on 15 years of operation and thus represent the 15 year net
present value of the energy savings. Details of TDV series can be found at CEC’s website:

http://energy.ca.gov/title24/2005standards/archive/rulemaking/documents/tdv/index.html

For reference the TDV virtual rate for the gas and electricity is shown in Figure 196.
FIGURE 196:
DCV using Sum of Zone method DCV using Critical Zone method
Climate
Total TDV Electricity TDV Elec Virtual Total TDV
ENERGY AND COST
Electricity TDV Elec Virtual Gas TDV Gas Virtual Gas TDV Gas Virtual
Zone
Savings Savings Elec Rate Savings Savings Gas Rate Savings Savings Savings Elec Rate Savings Savings Gas Rate Savings
(kWh/ft2) ($/ft2) ($/kWh) (Therm/ft2) ($/ft2) ($/therm) ($/ft2) (kWh/ft2) ($/ft2) ($/kWh) (Therm/ft2) ($/ft2) ($/therm) ($/ft2) SAVINGS PER SQUARE
CZ01 0.67 1.30 1.69 0.19 2.54 13.70 3.84 0.73 1.40 1.69 0.19 2.53 13.70 3.92
CZ02 0.74 1.47 1.77 0.18 2.39 13.87 3.86 0.35 0.75 1.77 0.18 2.34 13.88 3.08 FOOT
CZ03 0.88 1.71 1.77 0.19 2.47 13.88 4.18 0.68 1.35 1.77 0.19 2.46 13.87 3.81
CZ04 1.00 2.02 1.78 0.18 2.38 13.91 4.40 0.44 0.96 1.79 0.18 2.35 13.91 3.31
CZ05 0.99 1.85 1.72 0.18 2.45 13.80 4.31 0.79 1.51 1.72 0.18 2.44 13.79 3.95
CZ06 1.06 2.24 1.99 0.17 2.26 13.83 4.50 0.73 1.57 1.99 0.17 2.26 13.83 3.83
CZ07 1.13 1.97 1.63 0.16 2.15 13.82 4.12 0.88 1.40 1.65 0.16 2.15 13.82 3.55
CZ08 1.14 2.36 1.99 0.16 2.09 13.86 4.45 0.69 1.44 1.99 0.16 2.08 13.86 3.52
CZ09 1.15 2.51 2.05 0.16 2.09 13.75 4.60 0.59 1.30 2.06 0.16 2.08 13.76 3.38
CZ10 1.02 2.10 2.00 0.14 1.94 13.82 4.03 0.50 1.09 2.00 0.14 1.92 13.82 3.02
CZ11 0.76 1.57 1.82 0.16 2.16 14.03 3.73 0.24 0.65 1.82 0.16 2.10 14.03 2.75
CZ12 0.89 1.80 1.83 0.18 2.35 14.01 4.15 0.41 0.90 1.83 0.17 2.31 14.02 3.21
CZ13 0.92 1.81 1.79 0.15 2.03 14.05 3.84 0.10 0.43 1.78 0.15 1.99 14.05 2.43
CZ14 0.70 1.55 2.03 0.13 1.83 14.00 3.39 0.26 0.69 2.02 0.13 1.76 14.01 2.45
CZ15 1.31 2.80 2.05 0.10 1.44 13.80 4.24 0.45 1.10 2.04 0.10 1.43 13.81 2.54
CZ16 0.49 1.03 1.82 0.16 2.17 13.89 3.20 0.33 0.68 1.83 0.14 1.89 13.91 2.57

APPENDIX 11: ANALYSIS OF DCV IN DENSELY OCCUPIED ZONE OF A VAV SYSTEM X267X
 APPENDIX 12: FAN SYSTEMS

Overview

This section of the report covers the analysis of fan systems including the selection and
operation of the fan, motor, belts and variable speed drives. Specific issues addressed in this
section include:

ƒ Development and testing of fundamental fan system models

ƒ Comparison of fan type and sizing

ƒ Staging and isolation of multiple fans in parallel

ƒ Supply pressure reset

The comparison of fan types, fan sizing, fan staging and supply pressure reset are dealt with
in brief in this report. They are elaborated on in the HPAC article “A Fresh Look At Fans”
(Hydeman and Stein, 2003).

Throughout this section of the report we will use the term fan system to include the fan,
motor, physical drive (gears or belts) and variable speed drive (if appropriate). These
components are depicted in Figure 197.

FIGURE 197:
FAN SYSTEM
COMPONENTS
PHOTO COURTESY OF:
TAYLOR ENGINEERING

Fan System Models

A fan system model was developed to evaluate the impact of fan selection and control on
large building energy usage. For use in this project we needed a model that predicted energy
usage as a function of airflow (cfm) and fan static pressure (inches of water column). We
also have measured data for variable speed drive input (% speed) that we can convert to fan

X268X APPENDIX 12: FAN SYSTEMS

speed (rpm) using a correlation between the EMCS signal and tachometer readings at the fan
and motor.

We sought a model that had the following characteristics:

ƒ Accurate at predicting fan system energy over a range of full- and part-load operating
conditions

ƒ Easy to calibrate from manufacturer’s or field monitored data

ƒ Ability to identify operation in the manufacturer’s “do not select” or “surge” region

ƒ Relatively simple to integrate into existing simulation tools

ƒ Ability to separately model the performance of the fan system components including the
motor, the mechanical drive components, the unloading mechanism (e.g. VSD) and the
fan.

ƒ The model must be relatively simple to calibrate from data readily available from
manufacturers.

An existing gray-box regression model presented in the ASHRAE Secondary Systems

Toolkit 2 (Brandemuehl et. al, 1993) produces fan efficiency as a function of dimensionless
airflow and pressure. Although this model can be readily calibrated to manufacturers data,
this model does not directly work in existing simulation programs. This is due to the fact
that the it correlates efficiency to a dimensionless flow term which includes both airflow and
fan speed. Simulation tools like DOE2 use airflow and fan pressure as inputs to the fan
system model. The fan speed can only indirectly be obtained through iteration or other
mathematical solution. A second problem is that this model relies on the fan laws for
extrapolation between fan wheel sizes, it does not account for the improvement of fan
efficiency with wheel size that is apparent in manufacturer’s data.

The existing model in DOE2 was deemed unsuitable as it does not account for the variation
in the efficiency of each of the fan system components and assumes that the fan always rides
on a fixed system curve.

Energy usage of a fan system is driven by the efficiency of several components: the fan, the
fan belt, the motor, and possibly the variable speed drive. Each of the components has a
unique characteristic that changes its efficiency as a function of fan load. Our model is
composed of separate submodels for each component.

2
We believe that this model was originally documented in the HVACSIM+ program (Clark, 1985).

APPENDIX 12: FAN SYSTEMS X269X

 Characteristic System Curve Fan Model

We developed a gray box model based on the fan laws (referred to as the Characteristic
System Curve Model). This model is based on application of the perfect fan laws for the
variation of fan performance as a function of fan speed. The core assumption is that the
efficiency of a fan is constant as the fan rides up and down on a particular system curve.
Extensive testing with fan selection software shows this assumption appears to be true for all
manufacturer's fan data in both the surge and non-surge regions. For our model we defined
a “characteristic system curve” as a second order equation equating fan static pressure to
airflow (cfm) with zero constant and first order coefficients. A system curve is characterized
by a single coefficient, which we are calling SCC (system curve coefficient). The equation
for any system curve is:

SP
SCC =
CFM 2 (Equation 1)

Using this assumption it is only necessary to find fan performance at a single point on a
characteristic system curve to define its performance along that curve at all speeds. As
depicted in Figure 198, there are 3 system curves of particular importance: the curves at the
minimum and maximum ends of the tuning data set and the curve that represents the
highest efficiency for the fan. As described below and depicted in Figure 201 fans behave
very differently at each side of this peak efficiency. For plenum fans the “do not select” or
surge line is the same as this line of peak efficiency. For all other fans it appears to be to the
left of this peak efficiency line.

FIGURE 198:
Tuning Data for 660 CPL-A Model
TUNING DATA FOR 660
8
CPL-A CHARACTERISTIC 7.75
7.5 Points from Software Program
7.25
SYSTEM CURVE MODEL Effic = 33.0%
7
6.75 Effic = 12.4%
6.5 Peak Effic = 65.9% (Surge Line)
6.25
6
5.75
5.5
5.25
5
4.75
4.5
4.25
DP

4
3.75
3.5
3.25
3
2.75
2.5
2.25
2
1.75
1.5
1.25
1
0.75
0.5
0.25
0
0 20000 40000 60000 80000 100000 120000
CFM

X270X APPENDIX 12: FAN SYSTEMS

Figure 198 shows several points of data for a particular Cook fan with system curves drawn
through three of the points: two extreme points and a point on the system curve of highest
efficiency. These points were all taken from the Cook selection software. The efficiency is
calculated from the BHP reported by the software and using the equation:

CFM ∗ DP
FanEffic =
6350 ∗ BHP (Equation 2)

The model can be used to predict the fan power for any point whose system curve is between
the two extreme system curves. Figure 199 is the same data as Figure 198 but overlaid on
top of the fan curve from the Cook catalog. Notice that our surge line is almost exactly on
top of the manufacturer’s “Do Not Select” line.

When a fan enters the surge (a.k.a stall or pulsation) region not only does the efficiency drop
but the fan begins to vibrate which creates audible noise and vibrations that can damage the
fan, bearings, drive and attached ductwork. The further the fan moves into the surge region
the greater the vibration. Catastrophic failure can occur if the fan moves well into the surge
region at high power (high static). Some manufacturers appear to be more conservative than
others in terms of what amount of vibration is acceptable. Moving into the surge region at
low power (low static) is not likely to cause catastrophic failure or unacceptable vibration but
it will reduce fan life. From our experience, fans with variable speed drives commonly
operate for extended periods of time in the surge region, but it is usually at low power.

APPENDIX 12: FAN SYSTEMS X271X

 FIGURE 199: Points from Software Program
Tuning Data for 660 CPL-A Model
Effic = 33.0%
TUNING DATA FOR Effic = 12.4%
CHARACTERISTIC Peak Effic = 65.9% (Surge Line)
10
SYSTEM CURVE MODEL
ON TOP OF
9

MANUFACTURER'S FAN
CURVE 8

6
DP

0
0

17500

35000

52500

70000

87500

105000

122500

140000

157500

175000
CFM

Figure 200 shows fan efficiency plotted against system curve coefficient (SCC) for this data.
If we divide the data into surge and non-surge regions then we can fit a polynomial function
to each side of the data. These equations can accurately predict the efficiency in each region.

X272X APPENDIX 12: FAN SYSTEMS

 FIGURE 200:
FAN EFFICIENCY VS.
Efficiency vs System Curve Coefficient
SYSTEM CURVE
70.0% COEFFICIENT

Tuning data - Surge Region

60.0%
Tuning data - Non Surge Region
Poly. (Tuning data - Non Surge Region)
50.0% Poly. (Tuning data - Surge Region)
Fan Efficiency

40.0%

30.0%
y = 1E+31x 4 - 6E+23x 3 + 1E+16x 2 - 9E+07x + 0.7515
R2 = 0.9983
20.0%
y = -2E+54x 6 + 8E+45x 5 - 2E+37x 4 + 2E+28x 3 - 9E+18x 2 + 3E+09x + 0.0182
R2 = 0.9999
10.0%

0.0%
0.00E+0 2.00E-09 4.00E-09 6.00E-09 8.00E-09 1.00E-08 1.20E-08 1.40E-08 1.60E-08 1.80E-08 2.00E-08
0

System Curve Coefficient

The efficiency curve is easier to visualize and to fit a regression equation if plotted as a
function of the negative of the log of the system curve coefficient (see Figure 201). The log
causes the efficiency curves to become nearly straight lines and the negative plots flips the
surge and normal regions so that it matches manufacturer’s curves (i.e. surge to the left,
normal operation to the right). The base of the log does not seem to make much difference.
We use base 10 but other bases such as base “e” (natural log) also seem to work well. We
arbitrarily selected the name “Gamma” for the negative of the log of the system curve
coefficient.

Gamma = − log(SCC ) (Equation 3)

Critical Gamma is the gamma that corresponds to the system curve of highest fan efficiency.
One way to confirm the Critical Gamma is by trial and error using the manufacturer’s
software by comparing efficiency as you select points in the vicinity of the Critical Gamma.
For a particular fan, any gamma value less than the Critical Gamma is in surge and any
gamma greater than the critical gamma is in the non-surge region.

Fan efficiency can be very accurately predicted as a function of gamma. The most accurate
prediction comes from breaking the function into two parts: an equation for gammas in the
surge region and another for gammas in the non-surge region. A polynomial fits the data

APPENDIX 12: FAN SYSTEMS X273X

 nicely. A first order (i.e. linear) is reasonably accurate but a third order appears to provide
the best balance between fit and rational function behavior between calibrating data points.
While breaking the function into two parts is the most accurate, a single equation can
actually predict both the surge and non-surge regions fairly well. Figure 201 shows the R-
square term for both natural log and log10 regressions of various orders for a particular fan in
both the surge and non-surge regions.

FIGURE 201: Fan Effic as a Function of Gamma for 660 CPL-A Tuning Data

FAN EFFICIENCY AS A Surge

Non Surge
Poly. (Surge)
FUNCTION OF GAMMA 70.0% Poly. (Non Surge)

LN Curve Fits - Surge Region Log10 Curve Fits - Surge Region

First Order R2 = 0.9950 First Order R2 = 0.9984
60.0% Second Order: R2 = 0.9954 Second Order: R2 =0.9991
Third Order: R2 = 0.9997 Third Order: R2 = 0.9992
Fourth Order: R2 = 0.9998 Fourth Order: R2 = 0.9994
Fifth Order: R2 = 0.9999 Fifth Order: R2 = 0.9995
Sixth Order: R2 = 1.0000 Sixth Order: R2 = 0.9999
50.0%
Fan Efficiency

40.0%

y = -2.3737x6 + 118.4x5 - 2459.4x4 + 27230x3 - 169484x2 + 562293x - 776844

R2 = 0.9999
30.0% y = -0.0644x4 + 2.6282x3 - 40.101x2 + 270.59x - 680.26
R2 = 0.9998

20.0% LN Curve Fits-NonSurge Region: Log10 Curve Fits-NonSurge:

First Order R2 = 0.9984 First Order R2 = 0.995
Second Order: R2 = 0.9991 Second Order: R2 = 0.9954
Third Order: R2 = 0.9992 Third Order: R2 = 0.9997
10.0% Fourth Order: R2 = 0.9994 Fourth Order: R2 = 0.9998
Fifth Order: R2 = 0.9995 Fifth Order: R2 = 0.9999
Sixth Order: R2 = 0.9999 Sixth Order: R2 = 1.0

0.0%
0 2 4 6 8 10 12
Gamma = -Log10(Coefficient)

Figure 202 shows the accuracy of the Characteristic System Curve Fan model. This
particular model is based on 6th order polynomials of gamma with separate equations in the
surge and non-surge region. Table 36 depicts the fit results of 3rd order polynomials across a
range of manufacturers and fan types (plenum and housed, airfoil, forward curved and
backwardly inclined).

X274X APPENDIX 12: FAN SYSTEMS

 FIGURE 202:
Accuracy of the System Curve Fan Model for 660 CPL-A Tuning Data
ACCURACY OF
50
CHARACTERISTIC
45
SYSTEM CURVE FAN
40 MODEL

35

30
Predicted BHP

25

20

15

10

0
0 5 10 15 20 25 30 35 40 45 50
BHP (from manufacturer)

Left Region Right Region TABLE 36. FIT RESULTS

Count Min Max Average Min Max Average FOR 43 FANS
43 0.0% 5.7% 0.5% 0.1% 3.7% 1.7%

Figure 203 below depicts the predicted fan efficiency from the Gamma model for a plenum
fan. The predicted efficiency is plotted on the Z-axis as a function of the airflow (cfm, X-
axis) and fan static pressure (“H2O, Y-axis). The efficiency is computed between the
minimum and maximum characteristic system curves.

APPENDIX 12: FAN SYSTEMS X275X

 FIGURE 203:
COOK 660 CPL-A
PLENUM FAN
EFFICIENCY MAP USING
THE GAMMA MODEL

Extending the Characteristic System Curve Model to Multiple Diameters

ASHRAE Standard 51/AMCA Standard 210 (ASHRAE, 1999) specifies the procedures and
test setups that fan manufacturers use to test fans. Manufacturers are not required to test all
fan sizes. According to the standard, test information on a single fan may be used to
extrapolate the performance of larger fans that are geometrically similar using the perfect fan
laws. The following formulas are used to extrapolate performance:

3
⎛D ⎞
CFM 1 = CFM 2 × ⎜⎜ 1 ⎟⎟ (Equation 4)
⎝ D2 ⎠

2
⎛D ⎞
TP1 = TP2 × ⎜⎜ 1 ⎟⎟ (Equation 5)
⎝ D2 ⎠

2
⎛D ⎞
SP1 = SP2 × ⎜⎜ 1 ⎟⎟ (Equation 6)
⎝ D2 ⎠

5
⎛D ⎞
BHP1 = BHP2 × ⎜⎜ 1 ⎟⎟ (Equation 7)
⎝ D2 ⎠

X276X APPENDIX 12: FAN SYSTEMS

Figure 203 shows gamma curves for several fans including 5 sizes of Cook CPL-A plenum
airfoil fans. The 54 to 73 inch diameter CPL-A fans have virtually identical curves, just
shifted along the x-axis, but the 49 in. version has a different peak efficiency and curve shape.
This suggests that Cook tested the 54 in. fan and extrapolated the performance to the 60 to
73 in. sizes. Figure 2 0 4 also shows housed fans (CADWDI) and a flat blade plenum fan
(CPL-F). Each fan type has a unique curve shape but a single curve shape might be used for
multiple fans. To satisfy ourselves that Cook did in fact extrapolate the 54 in. data to the
larger sizes, we used the fan laws to extrapolate the manufacturer’s data for the 54 in. to a 60
in. diameter. We then compared our extrapolation to the manufacturer’s data for the 60 in.
fan. As Figure 2 0 4 shows, the extrapolated data was the same as the manufacturer’s data.

FIGURE 204:
Fan Efficiency vs Gamma for Several Fans
80% FAN EFFICIENCY VS.
75% GAMMA FOR SEVERAL
70% COOK FANS (BLUE
65% DASHED LINE IS THE

60% FLAT BLADE PLENUM

55% FAN)
Fan Efficiency

50%

45%

40%

35% 490 CPL-A (plenum - airfoil)

540 CPL-A
30% 600 CPL-A
660 CPL-A
25% 730 CPL-A
490 CADWDI (housed airfoil)
20% 600 CADWDI
660 CPL-F (plenum - flat blade)
15%

10%
7 7.5 8 8.5 9 9.5 10 10.5
Gamma

APPENDIX 12: FAN SYSTEMS X277X

 FIGURE 205:
Fan Efficiency vs Gamma for Two Cook Plenum Airfoil Fans
EXTRAPOLATION OF FAN 70%

DATA USING FAN LAWS 65%

60%

55%

50%

45%
Fan Efficiency

40%

35%

30% 540 CPL-A (Manufacturer's Data)

600 CPL-A (Manufacturer's Data)
25%
600 CPL-A extrapolated from 540 CPL-A

20%

15%

10%
7 7.5 8 8.5 9 9.5 10 10.5
Gamma = -Log10(SCC)

We believe that fan curves could be shifted and scaled as follows:

1. For each fan type develop a gamma model from a single fan using
manufacturer’s data.

2. Use the fan laws (Equations 4 to 7) to recreate fan curves for fans of
other diameters in that product line.

3. Provide an efficiency offset for peak efficiency.

Generalized Fan Model

Rather than relying on manufacturer’s data for specific fans, generalized models of each fan
type can be used to compare fan types and sizes.

A generalized fan model can be developed based on one or more manufacturers’ data. A
single generalized model can be used for all fans of a specific type, i.e. one model for all
housed airfoil fans another for all plenum airfoil fans, etc.

The generalized model could be used by DOE-2 or other simulation tools.

The generalized model is based on the assumption that all fans of a specific type have gamma
curves of the same shape but the exact location of the gamma curve (i.e. location of the peak
efficiency point) will vary based on diameter. When we refer to location of the curve we are
referring to the gamma at which the efficiency peaks and the value of the peak efficiency, i.e.
the x,y coordinates of the peak point in efficiency vs gamma space.

X278X APPENDIX 12: FAN SYSTEMS

Peak Efficiency Offset (Translating Up/Down)

Figure 206 shows data for the entire line of Cook housed airfoil fans. While it is clear that
data for many of the sizes were extrapolated from smaller sizes, it is also clear that the shape
of the gamma curves are fairly constant, they are simply translated up and down and left and
right.

120 CADWDI FIGURE 206:

Cook Housed Airfoil Fans 135 CADWDI
150 CADWDI ALL COOK HOUSED
165 CADWDI
180 CADWDI
90% 195 CADWDI AIRFOIL FANS
210 CADWDI
225 CADWDI
80% 245 CADWDI
270 CADWDI
300 CADWDI
70% 330 CADWDI
365 CADWDI
402 CADWDI
Static Fan Efficiency

60% 445 CADWDI

490 CADWDI
540 CADWDI
600 CADWDI
50% 660 CADWDI
730 CADWDI

40%

30%

20%

10%

0%
13 14 15 16 17 18 19 20 21 22 23 24 25 26
Gamma = -ln(SCC)

Figure 207 shows data from the entire line of Greenheck housed airfoil fans and the Cook
housed airfoil fans. Figure 208 is a subset of Figure 207. Again, the shapes of the curves are
similar across sizes and across manufacturers.

APPENDIX 12: FAN SYSTEMS X279X

 FIGURE 207: 120 CADWDI
Housed Airfoil Fans 135 CADWDI
150 CADWDI
HOUSED AIRFOIL FANS (CADWDI = Cook, AFDW = Greenheck) 165 CADWDI
180 CADWDI
195 CADWDI
90% 210 CADWDI
FROM TWO 225 CADWDI
245 CADWDI
270 CADWDI
MANUFACTURERS 80% 300 CADWDI
330 CADWDI
365 CADWDI
402 CADWDI
445 CADWDI
70% 490 CADWDI
540 CADWDI
600 CADWDI
660 CADWDI
60%
Static Fan Efficiency

50%

40%

30%

20%

10%

0%
13 14 15 16 17 18 19 20 21 22 23 24 25 26
Gamma

FIGURE 208: 180 CADWDI

Airfoil Fans 270 CADWDI
SUBSET OF HOUSED (CADWDI = Cook, AFDW = Greenheck)
490 CADWDI
90% 730 CADWDI
AIRFOIL FANS
18 AFDW
80% 27 AFDW
49 AFDW
73 AFDW
70%

60%
Static Fan Efficiency

50%

40%

30%

20%

10%

0%
13 14 15 16 17 18 19 20 21 22 23 24 25 26
Gamma

Figure 209 shows only the highest point on each gamma curve from Figure 207. It is a little
easier to see from this figure which fans the manufacturers tested and which they
extrapolated. For example, it is clear that both Cook and Greenheck tested their 30 in. fans.
Cook then extrapolated it all the way up to 73 in. (The variability in the peak efficiency of
the Cook 30 in. to 73 in. fans is due to rounding error and the fact that we only sampled at
discrete increments). Greenheck only extrapolated the 30 in. up to 36 in., then they tested

X280X APPENDIX 12: FAN SYSTEMS

the 40 in. and extrapolated that all the way to 73 in. The Cook 30 in. is more efficient than
the Greenheck 30 in. but not more efficient than the Greenheck 40 in. Had Cook tested a
40 in. (or larger) fan they might have found that it had higher efficiency than equally sized
Greenheck fans.

Figure 209 also shows that a curve can be fit to the data (using one or more manufacturers
data) in order to develop an equation for peak efficiency as a function of diameter. A more
accurate function may be possible using a log-log scale (See Figure 210).

FIGURE 209:
Housed Airfoil Fans: Peak Efficiency vs Diameter
HOUSED AIRFOIL FANS:
80% PEAK EFFICIENCY VS.
79%
78% DIAMETER
77%
76%
75%
74%
73%
Peak Efficiency

72%
71%
70%
69%
Cook Airfoils
68%
67% Greenheck Airfoils
66% Both
65% Poly. (Both)
64%
63% y = 1E-06x3 - 0.0002x2 + 0.0135x + 0.5078
62% R2 = 0.9112
61%
60%
0 10 20 30 40 50 60 70 80
Diameter (inches)

FIGURE 210:
All Airfoil: Natural Log of Peak Effic vs Natural Log of Diameter
PEAK EFFICIENCY VS.
0.00 DIAMETER ON LOG-LOG
2 2.5 3 3.5 4 4.5
-0.05 SCALE

-0.10
y = -0.072x2 + 0.6006x - 1.507
-0.15 R2 = 0.903
ln(Peak Effic)

-0.20

-0.25

-0.30

-0.35

-0.40

-0.45

-0.50
ln (Diameter)

APPENDIX 12: FAN SYSTEMS X281X

 Peak Gamma Offset (Translating Left/Right)

As noted above, fan laws can be used to translate gamma curves to the left and right based on
diameter. Another method of translating is to fit a regression to a full set of data from one or
more manufacturers. While it is a little difficult to tell from Figure 207 and Figure 208, fans
of the same size from different manufacturers peak at slightly different places. Figure 211
shows the relationship between diameter and the gamma location of the peak efficiency
point. Notice, for example, that the 40 in. Cook fan peaks at a different gamma than the 40
in. Greenheck fan. Figure 211 also shows that a reasonably accurate function can be
developed for gamma at peak as a function of diameter.

FIGURE 211:
Housed Airfoil Fans
GAMMA AT PEAK VS.
DIAMETER FOR HOUSED 22
AIRFOIL FANS
21

20
Gamma at Peak Efficiency

19 y = -0.0017x2 + 0.2485x + 12.257

2
R = 0.9862
18

17
Cook Airfoils
Greenheck Airfoils
16
Both
Poly. (Both)
15

14
0 10 20 30 40 50 60 70 80
Diameter

Motor Model

The next component for a fan system is the motor. We borrowed a model from the
Department of Energy’s Motor Challenge market transformation program
(http://www.oit.doe.gov/bestpractices/motors/). This model was presented to us by Gil
McCoy of Washington State University. In this model the efficiency of any motor consists
of a rated efficiency at nominal motor horsepower (MHP) and a part load function for
efficiency as a function of percent load that is defined as follows:

BHP
% Load = (Equation 8)
MHPno min al

X282X APPENDIX 12: FAN SYSTEMS

Note that the percent load does not correlate to the percent speed (one might expect it to be
a cube law relationship) because air profiles do not follow a single system curve. Thus the
percent speed for a fan with static pressure reset will produce a lower percent load than the
same percent speed for a fan with fixed static pressure setpoint (i.e. one operating at a higher
pressure for the same airflow).

Motor efficiency data can be found in the Department of Energy’s MotorMaster+ program
(http://mm3.energy.wsu.edu/mmplus/default.stm). This program has a database of
hundreds of motors from a range of manufacturers. Each motor is rated at full load, 75%
load, 50% load and 25% load. The same data is also available from Oak Ridge National
Laboratory’s Pumping System Assessment Tool (http://public.ornl.gov/psat/). The
rd
MotorMaster+ data can be fit using two equations: a 3 order polynomial from 25% to
100%MHP and the following function from 0 to 25%MHP:

BHP
MotorEfficiency 0− 25% = (Equation 9)
BHP + FixedLosses

Where fixed losses are calculated from motor efficiency at 25%:

0.25 ∗ MHP
FixedLosses = − 0.25 * MHP (Equation 10)
Effic25%

FIGURE 212:
SAMPLE MOTOR DATA
FROM PSAT SOFTWARE

APPENDIX 12: FAN SYSTEMS X283X

 FIGURE 213:
25 HP Energy Efficient Motor
EXAMPLE MOTOR
EFFICIENCY AS A 93.5%
FUNCTION OF LOAD
93.0%

Motor Efficiency
92.5%

92.0%
3 2
y = -0.0043x - 0.0736x + 0.1191x + 0.8862
91.5% 2
R =1
91.0%
0% 20% 40% 60% 80% 100% 120%
Percent MHP

Variable Speed Drive Model

The variable speed drive model is a 2nd order equation of natural log of percent load. The
calculation of percent load is done in the following steps:
1. The fan speed at the current cfm and fan static pressure is calculated
from a Secondary Toolkit fan model (Brandemuehl et. al, 1993)
solving for the dimensionless flow coefficient from the fan efficiency.

2. If this speed is above the minimum speed, the percent load is

calculated directly.

3. If this speed is below the minimum speed, the fan, motor and belt are
recalculated at the minimum speed with the static pressure adjusted
for the fan riding its curve. The percent load is than calculated

4. The VSD energy is calculated from the percent load from either step
2 or 3 above.

RPM Model

Using the Secondary Toolkit fan model, RPM is calculated from phi, the dimensionless flow
coefficient in two steps. First PHI is calculated from fan efficiency using two 3rd order
equations for above and below the peak efficiency point. Second, RPM is calculated from
PHI as follows:

CFM
RPM = , where Diameter is in feet (Equation 11)
Phi ∗ Diameter 3

X284X APPENDIX 12: FAN SYSTEMS

The tuning data from the manufacturer is used to develop equations for phi as a function of
fan efficiency. These equations can then be used, along with the output of the Characteristic
System Curve Fan Model, to predict Phi and RPM for any operating condition in the tuning
range.

In order to develop equations for phi as a function of fan efficiency, the fan efficiency tuning
data must be divided into two regions: left and right of peak efficiency. For the data we
analyzed a 3rd order polynomial fit both regions well. Figure 214 shows the equations
developed for phi as a function of fan efficiency in the surge and non-surge regions for the
660 CPL-A tuning data. This figure also shows that the relationship between phi and fan
efficiency is identical for the 600 CPL-A.

FIGURE 214:
Phi as a Function of Fan Efficiency
(tuning data for 660 CPL-A and 600 CPL-A) PHI AS A FUNCTION OF
1.2
Phi - 660 CPL-A (Non-Surge) FAN EFFICIENCY
Phi - 660 CPL-A (Surge Region)
600 CPL-A
1 Poly. (Phi - 660 CPL-A (Surge Region))
Poly. (Phi - 660 CPL-A (Non-Surge))

0.8

y = -433.9x6 + 989.51x5 - 903.06x4 + 419.05x3 - 104.28x2 + 12.817x + 0.3805

R2 = 0.9987
Phi

0.6

0.4

0.2

y = 4.4164x3 - 3.6764x2 + 1.5562x - 0.1077

R2 = 0.9991

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Fan Efficiency

Variable Speed Drive Model

Gilbert McCoy, at Washington State University, provided VSD performance data to Taylor
Engineering that he received from Saftronics, a VSD manufacturer (see Appendix A). The
combined efficiency of the MotorMaster/Saftronics data is reasonably consistent with similar
data provided by ABB (another manufacturer) and data in an ASHRAE paper by researchers
at the University of Alabama (Gao et. al, 2001 see Figure 215).

APPENDIX 12: FAN SYSTEMS X285X

 FIGURE 215:
1.2
COMBINED MOTOR AND

Percent of full load motor & VFD efficiency

DRIVE EFFICIENCY DATA
FROM FOUR SOURCES 1

0.8

Manuf #1
0.6
Manuf #2
0.4
From Saftronics/MotorMaster (assumes 100% speed
= 100% MHP)
0.2 from ASHRAE preprint of Transactions paper by
Gao

0
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percent MHP

Belt Model

According to AMCA Publication 203-90 (AMCA, 1990), drive loss is a function of motor
output (i.e. depends only on the BHP and not on the MHP). This is depicted in Figure
216.

FIGURE 216:
AMCA BELT LOSSES
DATA

X286X APPENDIX 12: FAN SYSTEMS

 Ln(belt loss) vs Ln(BHP) FIGURE 217:
AMCA BELT LOSSES
3 DATA
Medium Efficiency
2.5
Ln(Belt Losses)

1.5

0.5

0
-2 0 2 4 6 8
Ln(BHP)

Here is an approximation of the AMCA data:

< 0.3 BHP 0.3 to 100 BHP >100 BHP TABLE 37:
High Effic (low loss belts) 92% 97.2%
APPROXIMATE BELT
Med Effic 89% 96%
Low Effic (high losses) 84% 94% LOSS E

In the absence of any information on the type or quality of the belts, we have been assuming
medium efficiency belts for our fan scenario analyses. Tom Webster at the UC Berkeley
Center for the Built Environment has done some field research on belt efficiency at the NBI
PIER sites and is also finding that medium efficiency belts is a reasonable assumption.

FIGURE 218:
Belt Efficiency vs BHP
BELT EFFICIENCY
100% FUNCTIONS

95%
Belt Efficiency

90%
High Efficiency (low losses)
Med Efficiency
85% Low Efficiency (high losses)

80%

75%
0 5 10 15 20 25 30 35 40 45 50
BHP

APPENDIX 12: FAN SYSTEMS X287X

 Fan Type and Sizing

Comparison of fan types and sizes is relatively easy. For each fan we produce a Characteristic
Curve fan model as described above. Each of these fans is run across the measured cfm and
static pressure. To neutralize the inherent error in predicting energy use the base case staging
is modeled as well (as opposed to using the measure fan energy).

To compare classes of fans (like housed vs plenum) we can add a fixed amount of pressure to
the pressure for each individual fan at a given hour.

Fan Staging and Isolation

Fan staging is handled by comparing the operation of all available fans at a given record and
selecting the combination that is the most efficient.

Fixed losses for isolation devices such as inlet backdraft dampers or outlet isolation dampers
can be added to allow a fan specific pressure at each record. This is the same feature used for
housed fans under type and sizing.

Supply Pressure Reset

Supply pressure reset is achieved by mapping each cfm to a system curve representing the
amount and degree of reset. The system pressure is then read from the curve and used for
the calculation of fan energy.

Appendix 12 References

Hydeman, M.; and Stein, J. May 2003. A Fresh Look At Fans. HPAC Engineering.
Penton Publishing. Cleveland, OH.

Gao, X.; McInerny, S.A.; and Kavanaugh, S.P. June 2001. Efficiencies of an 11.2 kW
Variable Speed Motor and Drive. ASHRAE Transactions. American Society of Heating
Refrigeration and Air-Conditioning Engineers, Atlanta GA.

ASHRAE. 1999. ANSI/ASHRAE Standard 51-1999 (ANSI/AMCA Standard 210-99),

Laboratory Methods of Testing Fans for Aerodynamic Performance Rating. Atlanta GA:
American Society of Heating Refrigeration and Air-Conditioning Engineers.

Brandemuehl, M.J.; Gabel, S.; and Andresen, I. 1993. HVAC 2 toolkit: Algorithms and
Subroutines for Secondary HVAC System Energy Calculations. Atlanta GA: American
Society of Heating Refrigeration and Air-Conditioning Engineers.

X288X APPENDIX 12: FAN SYSTEMS

AMCA Publication 203-90, "Field Performance Measurement of Fan Systems." 0203X90A-
S. The Air Movement and Control Association International, Inc. Arlington Heights,
Illinois.

Clark, D.R. 1985. HVACSIM+ building systems and equipment simulation program:
Reference Manual. NBSIR 84-2996, U.S. Department of Commerce, Washington D.C.

APPENDIX 12: FAN SYSTEMS X289X

 Saftronic VSD Data

X290X APPENDIX 12: FAN SYSTEMS

APPENDIX 12: FAN SYSTEMS X291X
 APPENDIX 13: FOUR INCH AND ODD SIZE VAV BOXES

Most VAV box manufacturers offer VAV boxes with inlet sizes of 4, 5, 6, 7, 9, 10, 12, 14,
16, and 24x16 inches. Many designers only use even sized boxes starting from 6 inches.
Here are some reasons why you might want to avoid using 4 in., 5 in., 7 in., or 9 in. VAV
Boxes.

It is very difficult to tell if a box that is already installed is a 4 in. box or a 6 in. box with a
pancake reducer because the 4 in. box is a 6 in. box with an adapter, i.e. it has the same
dimensions as the 6 in. box.

FIGURE 219:
ADAPTER
PHOTO COURTESY OF:
TAYLOR ENGINEERING

FIGURE 220:
FOUR INCH BOX
(SUPPOSEDLY)
INSTALLED IN THE FIELD
PHOTO COURTESY OF:
TAYLOR ENGINEERING

X292X APPENDIX 13: FOUR INCH AND ODD SIZE VAV BOXES
 FIGURE 221:
DUAL DUCT BOX
PHOTO COURTESY OF:
TAYLOR ENGINEERING

Figure 221 shows a dual duct box (the engineer specified a 9 in. cold inlet and a 7 in. hot
inlet. The contractor supplied a box with two 10 in. inlets and pancake reducers on the
inlets.)

The flow probe in a 4 in. box is installed in the 6 in. neck but the assumption is that all the
air goes through the middle 4 in. of the neck because there is an adapter that funnels the air
to the middle. It is a little hard to believe that a 6 in. flow cross with an adapter works as
well as a flow cross without the adapter, i.e. the 4 in. box may not really have a lower
controllable minimum than a 6 in. box.

Even if the controllable minimum on a 4 in. box is lower than a 6 in. box, the difference in
minimum flow is only about 10 CFM (20 CFM min for a 4 in., 30 CFM min for a 6 in.).
So even if you had (10) 4 in. boxes on a job you could only save 100 CFM. The time you
would spend explaining the issue and field verifying the 4 in. boxes is not worth it. Save
your fee for something that can really make a difference.

Any savings from the lower controllable minimum on a 4 in. box could be offset by the
higher pressure drop of the 4 in. inlet duct.

Four inch or odd size spiral duct is not commonly produced and is not less expensive than
the next larger even size spiral duct. Thus there is a pressure drop (and energy) penalty but
no cost advantage.

Since a 4 in. or odd size spiral duct is not commonly produced it is likely the contractor will
use the next larger even size with a reducer at the box. Reducers add pressure drop and also
are a source of leakage.

One option to avoid 4 in. or odd size boxes is to add a diffuser to serve another nearby space
such as a corridor so that the total design flow for the box is more acceptable. For example,
suppose box 1-1 was going to serve Office 101 with a design flow of 100 CFM and box 1-2
was going to serve Offices 102 (300 CFM) and corridor 123 (200 CFM). Rather than use a
4 in. for 1-1 and a 7 in. for 1-2, you could move the corridor to 1-1 and use 6 in. boxes for
both zones.

APPENDIX 13: FOUR INCH AND ODD SIZE VAV BOXES X293X
 APPENDIX 14 - ZONE DEMAND
BASED RESET CASE STUDIES

Supply Air Temperature Reset

Table 31 and Table 32 present a summary of supply air temperature reset case study results
from several systems. The temperature data in Table 1 correspond to periods during which
the outdoor air temperature is within the reset range (generally outside air temperature below
60-70oF), whereas the data in Table 2 correspond to all times during which the systems are
operating. Setpoint data is presented where available, otherwise actual supply air
temperatures are shown. The two systems from the San Ramon Valley Conference Center
are examples where the resets operate effectively and the average supply air setpoint
temperatures are about 57 to 58oF. Results from AHU-14 are presented in more detail
below. The two units at the San Ramon Tech Center and several units on the UC Merced
campus are examples where the supply air resets are not operating effectively because one or
more “rogue zones” are consistently unsatisfied and drive the systems to their minimum
setpoints (for example, zone VAV2-9 from the SRTC AHU-2, shown in Figure 222). In
each case, the average supply air temperature is near the minimum at about 53 to 55oF. In
the case of the UC Merced systems, the minimum and maximum supply air temperatures
span a wide range because of poorly tuned SAT control loops. All of these temperature reset
systems feature PID control except the cold deck of the dual duct system at the UC Merced
Classroom building, which utilizes trim and respond logic.

Cooling Load (Percent

TABLE 38: SAT During Reset ( oF) Design of Design VFD Speed) Seq-
Project / Airflow Number Data uence
SUPPLY AIR System Type Max Min Avg (cfm) of Zones Period Max Min Avg Type Notes
San Ramon
TEMPERATURE RESET Valley
3 wk
Conference Setpoint 60.0 53.1 57.1 -- 13 100 4 89 PID
CASE STUDY SUMMARY Center /
summer
AHU-18
– OAT IN RESET RANGE San Ramon
Valley
(E.G. BELOW 70OF) Conference Setpoint 63.3 53.2 57.9 -- 18
1 wk
-- -- -- PID
summer
Center /
AHU-14
One
San Ramon or
1.5 mo
Tech Center Setpoint 60.1 53.0 54.2 -- -- 60 10 35 PID more
summer
Main AC-1 rogue
zones
One
San Ramon or
1.5 mo
Tech Center Setpoint 54.1 53.0 53.2 -- -- 52 10 38 PID more
summer
Main AC-2 rogue
zones

X294X APPENDIX 14 - ZONE DEMAND BASED RESET CASE STUDIES

 Overall SAT Cooling Load (Percent
o
( F) Design of Design VFD Speed) TABLE 39:
Airflow Number Data Sequence
Project / System Type Max Min Avg (cfm) of Zones Period Max Min Avg Type Notes SUPPLY AIR
UC Merced Library
and Information One or more TEMPERATURE RESET
Temperature 69.1 47.8 53.8 40,300 64 October 73 58 64 PID
Technology Center / rogue zones
AHU-1 CASE STUDY SUMMARY
UC Merced Library
and Information One or more – ALL OAT
Temperature 72.1 43.9 54.0 33,000 41 October 93 73 81 PID
Technology Center / rogue zones
AHU-2
UC Merced Library
and Information One or more
Temperature 60.6 47.8 53.9 28,000 ~ 40 October 100 65 92 PID
Technology Center / rogue zones
AHU-3
One or more
UC Merced Library
rogue zones;
and Information
Temperature 69 46.8 54.0 28,000 ~ 40 October 100 67 94 PID reset logic
Technology Center /
programmed
AHU-4
wrong
UC Merced One or more
Temperature 79.3 43.5 55.6 35,000 185 July 100 5 34 T&R
Classroom and rogue zones
Office Building / One or more
Temperature 82.4 43.7 55.4 35,000 185 August 100 5 44 T&R
AHU-1 Cold Deck rogue zones
Setpoint does
not respond
Morgan Hill Rec
Setpoint 65 53 62 24,000 26 Sept-Dec T&R until 3 zones
Center AC-1
request colder
air

VAV2-20 VAV2-21 VAV2-11 VAV2-22 VAV2-23 VAV2-24 VAV2-25 FIGURE 222:

VAV2-27 VAV2-30 VAV2-31 VAV2-6 VAV2-7 VAV2-28 VAV2-9 EXAMPLE OF ROGUE
ZONE AT SAN RAMON
100
TECH CENTER
Cooling Loop Output (%)

80

60

40

20

0
6/27/06 3:00 6/27/06 6:00 6/27/06 9:00 6/27/06 12:00 6/27/06 15:00 6/27/06 18:00 6/27/06 21:00

APPENDIX 14 - ZONE DEMAND BASED RESET CASE STUDIES X295X

 Supply Air Temperature Reset Case Study – San Ramon Valley Conference Center
AHU-14

At the San Ramon Valley Conference Center, the SAT setpoint is reset from T-min (53oF)
when the outdoor air temperature is 60oF and above, proportionally up to T-max when the
outdoor air temperature is 55oF and below. T-max ranges from 55oF to 63oF based on a PID
control loop that maintains the zone furthest from setpoint at setpoint plus 1oF.

Figure 223 shows the SAT setpoint as a function of the outdoor air temperature for AHU-14
over a period of 4 days during the summer. Actual setpoint values are represented by points
and the upper and lower range of expected setpoints (based on the value of T-max) are
represented by solid lines. Actual data points would be expected to fall anywhere between
the upper and lower boundaries. The setpoint reset with the outdoor air temperature along
the upper limit, suggesting that all zones were satisfied with the warmer supply air during
this period. Figure 224 shows a time-series plot of the same data along with the cooling
requests. During this period, the outdoor air temperatures were only cool enough to allow
the supply air temperatures to reset at night when the cooling loads are low. All of the
cooling requests (zones needing more cooling) during this period occur during the day when
the outdoor air temperatures are above 60oF and supply setpoint is already at its minimum.

FIGURE 223:
64
SUPPLY AIR
Supply Air Temperature Setpoint (oF)

TEMPERATURE RESET –
62
SAN RAMON VALLEY
CONFERENCE CENTER 60

58

56

54

52
52 54 56 58 60 62 64

Outdoor Air Temperature (oF)

X296X APPENDIX 14 - ZONE DEMAND BASED RESET CASE STUDIES

 FIGURE 224:
95 SAT Setpoint 9
Outdoor Air Temperature SUPPLY AIR
90 8
Cooling Requests TEMPERATURE RESET –

Number of Cooling Requests

85 7 SAN RAMON VALLEY
CONFERENCE CENTER
80 6
Temperature (oF)

75 5

70 4

65 3

60 2

55 1

50 0
7/28 7/28 7/29 7/29 7/30 7/30 7/31 7/31 8/1 0:00 8/1
0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 12:00

The San Ramon Tech Center is another building on the same campus as the SRVCC that
utilizes supply air temperature reset, although with a slightly higher outdoor air temperature
limit. Figure 225 shows reset data from AC-1 at the SRTC during a 6 week period over the
summer. In this case, during all but a few nights, the supply air resets with T-max at its
lower bound, indicating that generally there was at least one zone that could not meet its
cooling setpoint. Because the supply air never resets up to its upper limit, the full energy
savings potential of the reset sequence is not realized. Zones that are consistently
insufficiently cooled (potentially due to inadequate design air flows, for example) and force
the temperature reset to its minimum are referred to as “rogue zones”. The reset problem
could potentially be corrected by supplying more flow to these zones or by ignoring these
rogue zones from the reset logic. Note that ignoring rogue zones from the reset logic would
improve the efficiency of the system at the expense of the thermal comfort in the excluded
rogue zones.

APPENDIX 14 - ZONE DEMAND BASED RESET CASE STUDIES X297X

 FIGURE 225:
64
SUPPLY AIR

Supply Air Temperature Setpoint (oF)

TEMPERATURE RESET–
62
SAN RAMON TECH
CENTER
60

58

56

54

52
52 54 56 58 60 62 64 66 68
Outdoor Air Temperature (oF)

Static Pressure Reset

Static pressure reset results from several systems, all utilizing trim and respond logic, are
summarized in Table 40. Generally, the reset examples provided here are working effectively
to reduce fan energy and appear to operate robustly. The San Ramon Tech Center systems
averaged about 1.2 and 1.4 inches of static pressure over an extended period during the
summer. The Sacramento Courthouse systems, which averaged between 0.5 and 0.8 inches
of pressure over short periods during the winter and spring, feature control logic that allow
certain zones to be “locked out” or excluded from the reset sequences. Where there are
rogue zones that may otherwise drive the reset logic to the maximum pressure, this feature
allows the overall system to reset at the risk of potentially not meeting airflow setpoints at the
“locked out” zones.

The systems that were not operating as intended include the air handlers at the UC Merced
Library and Information Technology Center where a simple mistake in the control
programming only allowed the static pressure to reset within a range of 1.3 to 1.5 inches of
pressure.

More detailed examinations of static pressure reset case studies are presented below.

X298X APPENDIX 14 - ZONE DEMAND BASED RESET CASE STUDIES

 TABLE 40:
Cooling Load STATIC PRESSURE
(Percent of
Static Pressure (in Design VFD RESET CASE STUDY
WG) Design Speed)
Airflow Number Data Sequence
Project / System Data Type Max Min Avg (cfm) of Zones Period Max Min Avg Type Notes
SUMMARY
Actual
2.92 0 1.24
San Ramon Tech Center Pressure 1.5 mo
-- -- 60 10 35 T&R
/ Main AC1 Pressure summer
2 0.5 1.25
Setpoint
Actual
2.09 0 1.41
San Ramon Tech Center Pressure 1.5 mo
-- -- 52 10 38 T&R
/ Main AC2 summer
Pressure
2 0.5 1.41
Setpoint
Sacramento Courthouse / Pressure 20,000 ~ 16 per 2 days One zone locked out
1.1 0.46 0.73 -- -- -- T&R
(Average of 32 AHUs) Setpoint each AHU winter (6 % of total)
Sacramento Courthouse / Pressure ~ 16 per 2 days One zone locked out
1 0.3 0.81 20,000 55 35 40 T&R
AH-7A1 Setpoint AHU March (6 % of total)
Sacramento Courthouse / Pressure ~ 16 per 2 days One zone locked out
1.2 0.4 0.47 20,000 50 45 45 T&R
AH-7A2 Setpoint AHU March (6 % of total)
UC Merced Library and Max/min setpoints
Actual
Information Technology 1.55 1.1 1.32 40,300 64 October 73 58 64 T&R programmed as 1.5
Pressure
Center / AHU-1 and 1.3 inWG
UC Merced Library and Max/min setpoints
Actual
Information Technology 1.55 1.15 1.31 33,000 41 October 93 73 81 T&R programmed as 1.5
Pressure
Center / AHU-2 and 1.3 inWG
UC Merced Library and Max/min setpoints
Actual
Information Technology 1.65 0.25 1.23 28,000 ~ 40 October 100 65 92 T&R programmed as 1.5
Pressure
Center / AHU-3 and 1.3 inWG
UC Merced Library and Max/min setpoints
Actual
Information Technology 1.55 0.25 1.19 28,000 ~ 40 October 100 67 94 T&R programmed as 1.5
Pressure
Center / AHU-4 and 1.3 inWG
2.15 0.15 0.89 July 100 5 34 Two rogue zones.
UC Merced Classroom
Actual Max reset setpoint
and Office Building / 35,000 185 T&R
Pressure pressure changed
AHU-1 Cold Deck 1.95 0.2 1.32 August 100 5 44
several times.
Setpoint does not
respond until 3
Morgan Hill Rec Center Actual
1.5 0.5 0.6 24,000 26 Sept-Dec T&R zones request more
AC-1 Pressure
pressure (see case
study below)
Santa Clara Service Actual
1.0 0.5 0.75 -- 18 July PID See case study below
Center Building #3 Pressure

APPENDIX 14 - ZONE DEMAND BASED RESET CASE STUDIES X299X

 FIGURE 226:
Max, Min, Avg Pressure Setpoint (in. w.g.)
SUMMARY OF STATIC 0 0.5 1 1.5 2 2.5
PRESSURE RESET CASE
STUDIES San Ramon Tech
Center / Main AC1

San Ramon Tech

Center / Main AC2

Sacramento
Courthouse / (Average
of 32 AHUs)

Sacramento
Courthouse / AH-7A1

Sacramento
Courthouse / AH-7A2

UC Merced Library /
AHU-1

UC Merced Library /
AHU-2

UC Merced Library /
AHU-3

UC Merced Library /
AHU-4

UC Merced Classroom
Bldg / AHU-1 Cold Deck

Morgan Hill Rec Center

AC-1

Santa Clara Service

Center Building #3

Static Pressure Setpoint Reset Case Study – San Ramon Tech Center

The static pressure setpoint at the San Ramon Tech Center is reset using a trim and respond
approach within the range of 0.5 in. to 2.0 in. and an initial setpoint of 1.0 in. At each zone,
a “static pressure request” is made when the damper position is greater than 90%. Every 2
minutes, the setpoint is trimmed by -0.05 in. and responds by 0.05 in. times the number of
pressure requests, but no more than 0.2 in. Figure 227 shows trend results for the pressure
reset for a one day period. Most of the day, the pressure was reset down to its minimum,
however, increased setpoints occurred during periods with increased pressure requests due to
most-open damper positions above 90%. Toward the late afternoon, the pressure is
consistently forced up to its maximum.

X300X APPENDIX 14 - ZONE DEMAND BASED RESET CASE STUDIES

 Static Pressure Requests Static Pressure Setpoint FIGURE 227:
8 120
Most-Open Damper Position Outdoor Air Temperature STATIC PRESSURE
RESET CASE STUDY –
7
100 SAN RAMON TECH

Most-Open Damper Position (%) and

Number of Pressure Requests and

CENTER (AC-1)
Static Pressure Setpoint (inWG)

Outdoor Air Temperature (oF)

80
5

4 60

3
40

20
1

0 0
7/27 6:00 7/27 8:00 7/27 10:00 7/27 12:00 7/27 14:00 7/27 16:00 7/27 18:00

Figure 228 shows the static pressure setpoint at the SRTC as a function of the fan drive
speed over a two month period. Although the pressure setpoint is not controlled by the
drive speed, a clear correlation exists. The static pressure setpoint is maintained at its
minimum with a drive speed of about 40 percent and proportionally up to its maximum
with a speed of 70 percent or greater.

FIGURE 228:
2.5
STATIC PRESSURE
Static Pressure Setpoint (inWG)

RESET CASE STUDY –

2
SAN RAMON TECH
CENTER (AC-1 JUNE-
1.5 JULY 2006)

0.5

0
0% 20% 40% 60% 80% 100%
Fan VFD Speed

Figure 229 shows a plot of the actual static pressure as a function of the pressure setpoint.
The control loop performs well at maintaining the setpoint, as indicated by the alignment of

APPENDIX 14 - ZONE DEMAND BASED RESET CASE STUDIES X301X

 the data points along the line of slope = 1. The data are filtered by fan status but some
outlying points may occur during the transitional periods at system startup or shut down.

FIGURE 229:
3
STATIC PRESSURE
RESET CASE STUDY –
SAN RAMON TECH
2.5
CENTER (AC-1 JUNE-
JULY 2006)

Static Pressure (inWG) 2

1.5

0.5

0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Static Pressure Setpoint (inWG)

Figure 230 shows a histogram of static pressure reset data from AC-1 at the SRTC during
the months of June and July 2006. The histogram graphically depicts the frequency with
which the static pressure setpoints fall within certain ranges. The most common range of
pressures is 1.2 to 1.4 inches. The frequencies decline at transitional pressures above and
below this range, but show spikes again at the upper and lower limits of the reset logic.

X302X APPENDIX 14 - ZONE DEMAND BASED RESET CASE STUDIES

 FIGURE 230:
35%
STATIC PRESSURE
30% RESET CASE STUDY –
SAN RAMON TECH
25%
CENTER (AC-1 JUNE-
Frequency

20% JULY 2006)

15%

10%

5%

0%
0.0 - 0.2 - 0.4 - 0.6 - 0.8 - 1.0 - 1.2 - 1.4 - 1.6 - 1.8 -
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Static Pressure Setpoint (inWG)

Static Pressure Reset Case Study – Santa Clara Service Center Building #3

In 2004 The County of Santa Clara replaced an existing dual duct air distribution system
with a new single duct VAV system with hot water reheat for a 6000 ft2 wing of the Service
Center Building #3 on Berger Drive in San Jose, CA. The building is basically a small office
building. New equipment included a new rooftop air handler and 18 new DDC VAV
boxes. The control sequences included static pressure reset based on damper position and
supply air temperature reset based on zone demand.

After the contractor completed start-up, test and balance and commissioning, a 3rd party
commissioning agent was brought in to evaluate the system. The commissioning agent
found that at least two zone controllers were incorrectly calibrated such that the controller
was unable to measure design airflow at any duct static pressure. Rather than recalibrate
these controllers, the balancer had simply decided that 1.5 in. must be sufficient and set the
maximum setpoint to 1.5 in. As a result the static pressure setpoint was “pegged” at 1.5 in.
After recalibrating the zones in question and locking out other zones with obvious problems,
the maximum static pressure was reduced to 1.0 in. and the system was able to reset the
setpoint below the maximum on a consistent basis. Here is an excerpt from the
commissioning agent’s report:

Balancing Issues
1. Using a flow hood, we compared the energy management system
(EMS) air flow readings for VAV-1 near the maximum and
minimum design flows and found the EMS readings to be reasonably
accurate.

APPENDIX 14 - ZONE DEMAND BASED RESET CASE STUDIES X303X

 2. We discovered that the calibration factors for at least two zones were
incorrect. VAV-14, for example, was unable to achieve design flow
even at 1.5 in. static pressure. However, after adjusting the calibration
factors we were able to achieve design flow at 1.0 in. static pressure
with the damper 40% open. Similar issues were found at box 7A.

3. We adjusted the control sequences to ignore feedback from the

following zones for purposes of static pressure and supply air
temperature reset:

o VAV-14 – this serves the old mechanical room, which is now a storage
room.

o VAV-7A – this serves the men’s toilet and locker room. This box is the
closest one to the air handler and should never be starved but for some
reason it was unable to achieve design flow, possibly due to undersized
diffusers or obstructions in the ductwork. Regardless, locking this zone
out of the reset sequences is not likely to have adverse effects because the
design flow is 850, which is quite high for a space of this nature,
especially given the amount of transfer air required for the exhaust fan.
(We measured about 700 CFM at 1.0 in. SP setpoint) Furthermore, if
it gets a little too hot in the toilet/locker room it is not likely to be a
significant problem because of the transient nature of the space.)

o VAV-7B – women’s room -similar issues to 7A.

o VAV-11 and VAV-13 – These boxes serve the dispatch or “graveyard”

room. This room is also served by two Samsung heat pumps as well as a
dedicated outside air fan. The controls specification calls for the heat
pumps to be tied into the control systems so that they can be shut down
during the day when the air handler is running and turned on at night
during the “graveyard” shift. According to the controls contractor, the
heat pumps do not have contacts for shutting them down and the
manufacturer does not recommend cutting power to the units as a
means of control. Therefore, they are not interlocked. On the day of
the testing we noticed that one of the heat pumps was set to 74°F and
was in full heating while VAV-13 was set to 73°F and was in full
cooling and was driving the reset sequences for the rest of the building.
I suspect the heat pumps are more than capable of meeting the peak
heating and cooling loads so locking the VAVs out of the resets should
not result in under-heating or under-cooling but it does not entirely
resolve the issue of “fighting” between the VAVs and the heat pumps.
4. After locking out the above zones from the reset sequences we were
able to change the maximum static pressure setpoint from 1.5 in. to

X304X APPENDIX 14 - ZONE DEMAND BASED RESET CASE STUDIES

 1.0 in., i.e. with all zones at design flow and the fan maintaining 1.0
in. the dampers for all the other zones were less than 100% open.

Figure 231 shows that after the rogue zones were addressed the system was able to satisfy all
other zones with an average setpoint of about 0.75 in.

FIGURE 231:
Berger Building #3
STATIC PRESSURE
1.2
Sup.Static.Avg
RESET AT BERGER
Sup.Static.Setpt
DRIVE BUILDING #3
1

0.8
inches w.c.

0.6

0.4

0.2

0
7/11/2004 7/12/2004 7/12/2004 7/13/2004 7/13/2004 7/14/2004 7/14/2004 7/15/2004
12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00
date time

Static Pressure and Supply Air Temperature Reset Case Study

AC-1 is a Trane packaged VAV unit serving 26 zones in a recreation center in Morgan Hill.
Static Pressure setpoint is reset from 0.3 in. to 1.5 in. using trim and respond controls with a
setpoint of 2 zones, i.e. setpoint does not respond until 3 zones request more pressure. The
Trane unit’s internal control modulate fan speed to maintain duct static pressure at setpoint.
As seen in the figure below the unit’s internal controls have a minimum duct static setpoint
of 0.5 in. Supply air temperature setpoint is reset per the following sequence: Setpoint is
reset from T-min (53°F) when the outdoor air temperature is 70°F and above,
proportionally up to T-max when the outdoor air temperature is 55°F and below. T-max
shall be reset using trim and respond logic within the range 55°F to 65°F. When fan is off,
freeze T-max at the maximum value (65ºF). While fan is proven on, every 2 minutes,
increase the setpoint by 0.3ºF if there are no zone cooling requests. If there are more than
two (adjustable) cooling requests, decrease the setpoint by 0.3ºF. A cooling request is
generated when the cooling loop of any zone served by the system is 100%. The Trane
unit’s internal controls modulate the economizer and compressors to maintain supply air
temperature at setpoint.

APPENDIX 14 - ZONE DEMAND BASED RESET CASE STUDIES X305X

 Figure 232 shows two days of data for AC-1. The unit appears to operate from 1:00 am to
11:00 pm. For some reason the static pressure setpoint resets to 1.0 in. each morning at
4:00am but then quickly drops to the minimum setpoint of 0.3 in. and the actual static
pressure is stable at 0.5 in. throughout. The supply air temperature is also pegged at it’s limit
of 65oF most of the time. During the middle of the day, as the AC-1 outside air temperature
rises, the amount of temperature reset allowed by the sequence is limited and the supply air
temperature setpoint goes down accordingly. Two outside air temperature series are plotted
hear because the AC-1 outside air temperature sensor is not believed to be falsely influenced
by direct sunlight. The nearby IEC unit’s outside air temperature sensor is believed to be
more accurate. The Trane unit’s internal controls do not do a very good job of maintaining
supply air temperature at setpoint, particularly when the supply air temperature setpoint is
low.

FIGURE 232: AC-1 - Current Duct Pressure SP AC-1 - Supply Duct Pressure
1.2 AC-1 - Supply Air Temp. AC-1 - Supply Temp. SP 70
IEC - Outside Air AC-1 - Outside Air Temp.
PRESSURE AND
TEMPERATURE RESET AT
1 65
MORGAN HILL FOR TWO
DAYS IN DECEMBER
0.8 60

0.6 55

0.4 50

0.2 Unstable Unit Off 45

SAT control when
pressure is 0

0 40
12/19/06 12/19/06 12/19/06 12/19/06 12/19/06 12/20/06 12/20/06 12/20/06 12/20/06 12/20/06 12/21/06
12:00 AM 4:48 AM 9:36 AM 2:24 PM 7:12 PM 12:00 AM 4:48 AM 9:36 AM 2:24 PM 7:12 PM 12:00 AM

Static Pressure Reset Case Study – Underfloor System

Figure 233 reflects two weeks of monitored data from a 30,000 CFM VAV air handler
showing the static pressure setpoint being reset from 0.9 in. down to 0.15 in. It also shows
that the actual static pressure is closely tracking the setpoint, although it increases its hunting
as the fan speed is reduced.

X306X APPENDIX 14 - ZONE DEMAND BASED RESET CASE STUDIES

 FIGURE 233:
MONITORED DATA
ILLUSTRATING STATIC
PRESSURE RESET

Demand Controlled Ventilation

Results from CO2 sensors from several zones featuring demand controlled ventilation are
summarized in Table 41. Most of the zones have had occurrences where the maximum CO2
concentration has exceeded the DCV setpoint. In the case of the VAV zones at the UC
Merced Library and Information Technology Center, spikes in CO2 concentration are
typically brief and coincide with rising space temperatures. The corresponding increases in
zone airflow help to control both the CO2 and the space cooling. As such, the average CO2
concentrations in these zones are generally low and correspond to ambient concentrations.
In contrast, the average CO2 concentrations at the UC Merced Classroom building are well
above ambient and the DCV setpoint is frequently exceeded for prolonged periods. The
trend review for this building determined that the DCV control sequences were incorrectly
implemented; in particular, the maximum zone flow setpoint for the dual duct system was
not programmed as intended. The DCV control logic properly increased the flow setpoints
to their maximums in these zones when the DCV setpoint was exceeded, but the maximum
setpoints were insufficient to adequately reduce the CO2 concentrations. As a result, the
frequent peaks in CO2 concentration occurred over several hours at a time and the average
concentration was well above ambient.

The average CO2 concentration at the San Ramon Valley Conference Center was also well
above ambient but may have possibly been due to calibration error. The DCV system
appeared to respond correctly to high CO2 concentrations both at the zone and air handler; a
detailed examination of this system is presented below.

APPENDIX 14 - ZONE DEMAND BASED RESET CASE STUDIES X307X

 TABLE 41: CO2 Concentration
Zone or
DEMAND CONTROLLED Project AHU Ave Max Min Setpoint Period System Type Notes
UC Merced Library and
VENTILATION CASE August to
Information Technology VAV-206 384 622 342 1000 VAV reheat
September
STUDY SUMMARY Center
UC Merced Library and
Information Technology VAV-231 410 1040 369 1000 October VAV reheat
Center
UC Merced Library and
June to
Information Technology VAV-232 383 1982 268 1000 VAV reheat
September
Center
UC Merced Library and
March to
Information Technology VAV-238 372 573 330 1000 VAV reheat
September
Center
Maximum zone
UC Merced Classroom DDV- October to
528 1275 344 1000 Dual Duct airflow incorrectly
and Office Building 101 November
programmed
Maximum zone
UC Merced Classroom DDV- September to
715 1567 526 1000 Dual Duct airflow incorrectly
and Office Building 103 October
programmed
Maximum zone
UC Merced Classroom DDV- September to
608 1854 327 1000 Dual Duct airflow incorrectly
and Office Building 122 November
programmed
March to
Sacramento Courthouse AHU-7 453 599 367 1000 VAV reheat
August
San Ramon Valley VAV22- 2 days
646 1167 527 1100 VAV reheat
Conference Center 37 summer
San Ramon Valley VAV22- 2 days
620 1432 552 1100 VAV reheat
Conference Center 33 summer
OA damper at
San Ramon Valley 2 days AHU responds to
AHU-1 680 1432 561 N/A VAV reheat
Conference Center summer CO2 control loop
signals above 50%

Demand Controlled Ventilation Case Study – San Ramon Valley Conference Center

At the San Ramon Valley Conference Center, twelve VAV zones served by a common air
handling unit are equipped with CO2 sensors for demand controlled ventilation. Minimum
zone airflow rates vary from the normal zone minimum to the zone maximum based on the
outputs of PID loops with CO2 concentration setpoints of 1100 ppm. At the air handler,
the minimum outdoor air damper position is mapped from the minimum position for
ventilation to 100 percent open for loop outputs of 50 to 100 percent from any zone.

Figure 234 shows trends for two of the zones and the air handler over a two day period to
demonstrate the successful operation of the DCV control sequences. The zone airflows
generally modulate during occupied hours to meet the cooling setpoints but also respond to
high CO2 concentrations. The CO2 concentration in zone VAV22-37 slightly exceeded the
setpoint in the late morning each day and caused brief bursts in airflow that effectively
reduced the CO2 concentration below the setpoint on the first day. On the second day,
modulations in the airflow may have been responding to a combination of the cooling loop
and CO2 loop outputs. In zone VAV22-23, the airflow clearly responded to a large and
sustained CO2 concentration spike (up to a maximum of 1432 ppm) on the second day.

At the air handler, the OA damper minimum position was typically set at about 33 percent,
whereas the actual damper position varied from the minimum up to 100 percent open for
economizer operation. On the first day, the air handler did not respond to the small spikes
in CO2, indicating that the CO2 loop output did not reach 50 percent. On the second day,

X308X APPENDIX 14 - ZONE DEMAND BASED RESET CASE STUDIES

however, the sustained spike resulted in a loop output of 100 percent forcing the minimum
OA damper position to its maximum (the damper was already fully open in economizer
mode).

1500 600
FIGURE 234:
VAV22-37 CO2 CO2 Setpoint VAV22-37 Airflow Setpoint DEMAND CONTROLLED
500

Zone Airflow Setpoint

1250 VENTILATION CASE
CO2 Concentration

400
STUDY
1000 300

200
750
100

500 0
7/30/06 12:00 7/31/06 0:00 7/31/06 12:00 8/1/06 0:00 8/1/06 12:00

1500 800

VAV22-23 CO2 CO2 Setpoint VAV22-23 Airflow Setpoint 700

Zone Airflow Setpoint

1250 600
CO2 Concentration

500

1000 400

300

750 200

100

500 0
7/30/06 12:00 7/31/06 0:00 7/31/06 12:00 8/1/06 0:00 8/1/06 12:00

1500 AHU-1 High CO2 AHU-1 OA Damper AHU-1 OA Damper Min 150

125
CO2 Concentration (ppm)

Damper Position (%)

1250
100

1000 75

50
750
25

500 0
7/30/06 12:00 7/31/06 0:00 7/31/06 12:00 8/1/06 0:00 8/1/06 12:00

APPENDIX 14 - ZONE DEMAND BASED RESET CASE STUDIES X309X

 APPENDIX 15 – REFERENCES

General

Commercial Building Survey Report. Pacific Gas and Electric Company, San
Francisco CA. 1999. A useful resource for existing building stock
characteristics in California.

The Control System Design Guide and Functional Testing Guide for Air Handling
Systems. Available for no-cost download at http://buildings.lbl.gov/hpcbs/FTG.
The control design guide portion is targeted at designers but will also be a
useful support tool for commissioning providers. It includes information on
the control design process, standard point list templates for various air
handling system configurations, valve sizing and scheduling tools, damper
sizing and scheduling tools, information on sensing technologies and
application recommendations, and sample standard details that can be opened
in AutoCAD® and used as starting points by designers.

The functional testing guide portion is targeted at commissioning providers

but will also be useful support tool for designers. It includes information on
testing basics as well as information on testing the air handling system at a
component level and an integrated system level. Each chapter includes tables
that outline the energy and resource benefits associated with testing that
particular component, the purpose behind testing in the area that is the subject
of the chapter, the instrumentation requirements, the time required, the
acceptance criteria, and a listing of potential problems and cautions. Many
chapters also contain a table that outlines design issues related to successfully
commissioning the component that is the subject of the chapter. In many
instances, this information is linked to additional information providing the
theory behind the issues. The PG&E Commissioning Test Protocol library is
fully embedded into the guide, allowing users to open and modify publicly
available tests for their own use based on information in the guide and the
requirements of their project. A calculation appendix illustrates the use of
fundamental equations to evaluate energy savings or solve field problems
including examples from projects where the techniques have been employed.

The guide also includes reference appendix listing numerous references that
would be useful to those involved with the design, installation, commissioning,
and operation of air handling systems and their related control and utility
systems.

X310X APPENDIX 15 – REFERENCES

Energy Design Resources. http://www.energydesignresources.com/. This site has a
number of design briefs covering a range of topics from simulation to chilled
water plant design.

Kammerud, Ron, PhD, Ken Gillespie, and Mark Hydeman. Economic

Uncertainties in Chilled Water System Design. June 1999. ASHRAE, Atlanta
GA. SE-99-16-3. This paper explores the accuracy of simulation components
like equipment model calibration and the accuracy of the load profile on the
resulting cost-benefit analysis.

Controls

Hartman, Tom. “Improving VAV Zone Control.” ASHRAE Journal. June 2003.
Schemes for integration of zone controls including occupancy, lighting and
temperature. Presents a challenge to existing practices for both comfort and
energy performance.

Hartman, Tom. “Ultra Efficient Cooling with Demand-Based Control.” HPAC.

December 2001. Schemes for integration of zone controls including
occupancy, lighting and temperature. Presents a challenge to existing practices
for both comfort and energy performance.

Taylor, Steve. ASHRAE Fundamentals of HVAC Control Systems Self-Directed

Learning Course. 2001. Atlanta GA. An excellent primer in HVAC control
design.

The Iowa Energy Center website at www.DDC-Online.org provides a lot of useful

information regarding DDC theory in general and a generic apples-to-apples
comparison of the offerings of most of the major control vendors.

Supply Air Temperature

Bauman, F., T. Borgers, P. LaBerge, and A. Gadgil. "Cold Air Distribution in

Office Buildings: Technology Assessment for California." ASHRAE
Transactions, Vol. 99, Pt. 2, pp. 109-124, June 1992. Previously published by
Center for Environmental Design Research, University of California, Berkeley,
61 pp.

Xiangyang Chen and Kazuyuki Kamimura. Vote Method of Deciding Supply Air
Temperature Setpoint for VAV System, ASHRAE Transactions.

APPENDIX 15 – REFERENCES X311X

 Yu-Pei Ke and Stanley Mumma. “Optimized Supply Air Temperature in VAV
Systems,” Energy, Vol. 22, No. 6, 1997.

Night Flushing

Braun, James E., Ph.D., Montgomery, Kent W., Chaturvedi, Nitin. “Evaluating
the Performance of Building Thermal Mass Control Strategies.” ASHRAE
Journal of HVAC&R Research, Vol. 7, No. 4. October 2001.

Braun, James E. Load Control using Building Thermal Mass.

Braun, James E. Reducing Energy Costs and Peak Electrical Demand Through
Optimal Control of Building Thermal Storage.

Keeney, Kevin R., Braun, James E., Ph.D. Application of Building Precooling to
Reduce Peak Cooling Requirements.

Load Calculations

Brown, Karl. "Setting Enhanced Performance Targets for a New University

Campus: Benchmarks vs. Energy Standards as a Reference?" In Proceedings of
the 2002 ACEEE Summer Study of Energy Efficiency in Buildings. 4:29-40.
Washington, D.C.: American Council for an Energy-Efficient Economy.
Presents an innovative procedure used to prevent oversizing of central plant
and building services at the new UC Merced campus.

VAV Box Sizing

Bauman, Fred, Charlie Huizenga, Tengfang Xu, and Takashi Akimoto. 1995.
Thermal Comfort With A Variable Air Volume (VAV) System. Center for
Environmental Design Research, University of California, Berkeley, California.
Presents research on ADPI for diffusers over a range of flows.

Fisk, W.J., D. Faulkner, D. Sullivan, and F.S. Bauman. "Air Change Effectiveness
And Pollutant Removal Efficiency During Adverse Conditions." Indoor Air;
7:55-63. 1997.Denmark: Munksgaard.

X312X APPENDIX 15 – REFERENCES

Persily A.K. and Dols W.S. "Field measurements of ventilation and ventilation
effectiveness in an office/library building", Indoor Air, Vol 3, 1991.

Persily A.K. "Assessing ventilation effectiveness in mechanically ventilated office

buildings," International Symposium on Room Air Convection and
Ventilation Effectiveness, Tokyo, 1992

Offerman F.J, Int-Hout D. Ventilation effectiveness and ADPI measurements of a

forced air heating system," ASHRAE Transactions 94(1), 1988. pp. 694-704.

Fans and Fan Systems

AMCA Publication 200 Air Systems.

AMCA Publication 201 Fans and Systems.

AMCA Publication 202-88 Troubleshooting.

AMCA. 1990. AMCA Publication 203-90, Field Performance Measurement of Fan

Systems. 0203X90A-S. The Air Movement and Control Association
International, Inc. Arlington Heights, Illinois. This guide has many useful
tidbits including field measurement protocols, data on belt performance and
others.

ASHRAE. ANSI/ASHRAE Standard 51-1999 (ANSI/AMCA Standard 210-99),

Laboratory Methods of Testing Fans for Aerodynamic Performance Rating. 1999.
Atlanta GA: American Society of Heating Refrigeration and Air-Conditioning
Engineers. Details the methods of testing and rating fan performance. This
includes the process used by manufacturers to extend limited data sets to a
family of fans.

ASHRAE. ASHRAE Handbook of Fundamentals, Chapter 32, Duct Design. Design

guides for HVAC duct design and pressure loss calculations.

ASHRAE. ASHRAE Handbook of Air-Handling Equipment, Chapter 18, Fans.

Details on fan selection and performance.

ASHRAE, Duct Fitting Database CD, 2002.

Brandemuehl, Michael, Shauna Gable, Inger Anderesen. HVAC 2 Toolkit, A

Toolkit for Secondary HVAC System Energy Calculations. ASHRAE, Atlanta
GA. 1993. A compendium of component models for air-side systems.

APPENDIX 15 – REFERENCES X313X

 The Energy Design Resources briefs titled “Design Details”, “Document Review”,
and “Field Review” discuss the resource and first cost savings associated with
providing good detailing on HVAC contract documents. The first brief
focuses on the details themselves. The second focuses on making sure the
details are properly reflected on the contract documents. The third focuses on
making sure the installation reflects the requirements detailed on the contract
documents. All are available for free down load at
www.energydesignresources.com. There are also numerous other design briefs
on the EDR site, some of which are highly applicable to air handling system
design including topics like Integrated Energy Design, Economizers, Drive
Power, Building Simulation, and Underfloor Air Distribution.

Hydeman, Mark, Jeff Stein. “A Fresh Look at Fans”. HPAC. May 2003. Presents
a detailed evaluation of fan selection and control for a commercial office
building.

Hydeman, Mark, Jeff Stein. Development and Testing of a Component Based Fan
System Model. ASHRAE, Atlanta GA. January 2004. Presents a new
component based fan system model that can be used for simulations of airside
system design. This includes details for modeling of motors, belts and VSDs.

SMACNA HVAC Systems Duct Design. 1990. Design guides for HVAC duct
design and pressure loss calculations.

Stein, Jeff, Mark Hydeman. Development and Testing of the Characteristic Curve
Fan Model. ASHRAE, Atlanta GA. January 2004. Presents a new fan model
that can be used for simulations of airside system design.

Wang, Fulin, Harunori Yoshida, Masato Miyata. Total Energy Consumption

Model of Fan Subsystem Suitable for Automated Continuous Building
Commissioning. ASHRAE, Atlanta GA. January 2004. Presents a new
component based fan system model that can be used for simulations of airside
system design.

Filters

Burroughs, H.E. Barney. “The Art and Science of Air Filtration in Health Care”.
HPAC. October 1998.

Burroughs, H.E. Barney. “Filtration: An Investment in IAQ”. HPAC. August

1997.

X314X APPENDIX 15 – REFERENCES

Chimack, Michael J. and Dave Sellers. “Using Extended Surface Air Filters in
Heating Ventilation and Air Conditioning Systems: Reducing Utility and
Maintenance Costs while Benefiting the Environment.” Available from PECI
at http://www.peci.org/papers/filters.pdf

NAFA Guide to Air Filtration. 1996. (available from National Air Filtration
Association website or ASHRAE website). This manual provides a complete
source for information about air filtration; from the basic principles of
filtration, and different types of filtration devices, to information about testing,
specialized applications, and the role of filtration in Indoor Air Quality.

Outside Air Dampers

ASHRAE Guideline 16-2003. Selecting Outdoor, Return, and Relief Dampers for
Air-Side Economizer Systems. An excellent and detailed reference for
specification of dampers for air-side economizer systems.

The mixing and economizer section chapter in the Functional Testing Guide (see
reference above under “General”) along with is supplemental information
chapter contains a lot of information on dampers, economizers, and their
controls. The control design guide contains information on damper sizing as
well as a linked spreadsheet that provides the user with the framework for a
damper schedule, illustrates some typical sizing calculations, and includes the
characteristic curves or opposed and parallel blade dampers.

CO2 and DCV

Emmerich, Steven J. and Andrew K. Persily. “State-of-the-Art Review of CO2

Demand Controlled Ventilation Technology and Application.” NISTIR 6729,
March 2001. A thorough review of DCV technology and research.

Part 1 - Measure Analysis and Life-Cycle Cost – DRAFT 2005 California Building
Energy Efficiency Standards. California Energy Commission, Sacramento CA.
P400-02-011, April 11, 2002. Details on life-cycle cost analysis of demand
ventilation controls for single zone systems.

Schell, M.B. and D. Int-Hout. “Demand Control Ventilation Using CO2.”

ASHRAE Journal. February 2001. An excellent primer on DCV control
system design.

APPENDIX 15 – REFERENCES X315X

 Schell, M.B. “Real Time Ventilation Control.” HPAC, April 2002.

Project Reports

The following reports, available at www.energy.ca.gov/research/index.html, or at

www.newbuildings.org/pier were also produced during the research leading to
development of this design guideline:

Integrated Energy Systems: Productivity & Building Science – PIER Program Final
Report. This report contains the objectives, approach, results and outcomes for
the six projects of this PIER program. A full summary of the Integrated Design
of Large HVAC Systems project is included. Publication # P500-03-082

Large HVAC Building Survey Information (Attachment A-20 to Publication

#P500-03-082), October 2003. This document contains the following three
reports published by this PIER project: A Database of New CA Commercial
Buildings Over 100,000 ft2, the Summary of Site Screening Interview, and the
Onsite Inspection Report for 21 Sites.

Large HVAC Field and Baseline Data (Attachment A-21 to Publication # P500-03-
082), October 2003. This document contains the following three reports
published by this PIER project: Field Data Collection (comprised of Site
Survey Data Form, Site Survey Letter and Site Survey Schedule), Sensitivity
Analysis and Solutions Report.

Large HVAC Energy Impact Report (Attachment A-22 to Publication # P500-03-

082), October 2003. This report describes the estimated energy savings due to
measures recommended in the guideline on both a per-building and statewide
basis.

X316X APPENDIX 15 – REFERENCES