AlllDl 777D3E NATL INST.

OF STAND & TECH V

Reference Publi-
cations AlllOt, MQ75SQ

NBSIR 81-2404

Calculation of Inter-Room Air

Movement for Multi-Room Building
Energy Analysis

U S. DEPARTMENT OF COMMERCE
National Bureau of Standards
National Engineering Laboratory
Center for Building Technology
Washington, DC 20234

November 1 981

Sponsored by
National Bureau of Standards
and
Passive and Hybrid Solar Energy Division
Office of Solar Heat Technologies
_QQ S. Department of Energy
3shington, DC 20585
j^QQ

.U56
z\-im
1931
I
 VATIONAL BURKai
OF ITANDAItU«(
Lirnary

JAN 8 1982

NBSIR 81-2404
e A •
not
Ocioo
^
CALCULATION OF INTER-ROOM AIR . Ui^(o
MOVEMENT FOR MULTI-ROOM BUILDING mo
ENERGY ANALYSIS

George N. Walton

U.S. DEPARTMENT OF COMMERCE

National Bureau of Standards
National Engineering Laboratory
Center for Building Technology
Washington, DC 20234

November 1 981

Sponsored by
National Bureau of Standards
and
Passive and Hybrid Solar Energy Division
Office of Solar Heat Technologies
U.S. Department of Energy
Washington, DC 20585

U.S. DEPARTMENT OF COMMERCE, Malcolm Baldrige, Secretary

NATIONAL BUREAU OF STANDARDS. Ernest Ambler, Director
I

^1
 List of Figures

Figure 1. Vertically adjacent rooms.

Figure 2. Horizontally adjacent rooms.

Figure 3. Envelope surface.

Figure 4. Townhouse floor plan.

Figure 5. Section of townhouse.

Figure 6. Wind effects — computer air flows through major building

surfaces

Figure 7. Inter-room openings.

Figure 8. Effect of interior partition.

Figure 9. Natural ventilation.

Figure A-1. MRLP structure.

i

it
 ABSTRACT

A model is presented for computing the infiltration and air flow between rooms
of a multi-room building in terms of basic principles of fluid mechanics. This
model has been incorporated into a comprehensive loads-predicting computer pro-
gram. Air flows, room temperatures, and heating loads for a typical townhouse
under different conditions of environment and with various construction features
are computed. These calculations show the feasibility of detailed multi-room
air movement analysis. They also indicate that when the inter-room openings of
a low-rise structure are large compared to the envelope openings, the infiltra-
tion and total load can be accurately, and more quickly, computed by assumming
no resistance to air flow between rooms. This property will also allow simpli-
fied calculations for high-rise buildings with many rooms. Methods are proposed
for handling more complex air flow phenomena.

Keywords: building energy analysis; computer simulation; infiltration; natural

ventilation.

iii
 TABLE OF CONTENTS

Page

ABSTRACT ill

1. INTRODUCTION I

1. 1 Background 1
1.2 Objective 1
1.3 Scope 2

2. METHODS 3

2.1 Simulation Capabilities 3

2.2 Room Energy Balance 3
2.3 Natural Air Movements 4
2.3.1 Wind Forces 4
2.3.2 Thermal Forces 6
2.3.3 Flow Equations 7
2.3.4 Solution of the Flow Equations 8

2.4 Sample Problem 9

3. RESULTS AND DISCUSSION 10

3.1 Wind Direction Effects 10

3.2 Inter-Room Openings 10
3.3 Natural Ventilation 11
3.4 Future Developments 12

APPENDIX A: PROGRAM STRUCTURE A-1

APPENDIX B: SIMULATION INPUT B-1
APPENDIX C: TOWNHOUSE INPUT DECK C-1
1. INTRODUCTION

1.1 BACKGROUND

The effort to encourage the design, construction, and retrofit of energy

conserving buildings will be strengthened by the development of comprehensive
methods of evaluating the thermal effectiveness of those buildings. Algorithms
which can accurately determine heat flow quantities, especially for natural
heating and cooling applications, are not presently included in existing energy
analysis computer programs and must be developed and tested over a wide range
of parameters before they can be adapted into standard user-oriented design
tools. Among these design tools are the BLAST [1] and DOE-2 [2] programs.
These programs include simulation of operational and physical characteristics
for many building features which allows the study of the interactions of design
options. Although these programs are extremely versatile and efficient for con-
ventional buildings and for conventional operating procedures, they lack feasi-
bility in analyzing innovative building designs. Moreover, they are difficult
to modify for the testing of new algorithms. One of the major areas neglected
in DOE-2 and BLAST is the energy transfer due to air movement. This includes
both the effects of building construction on infiltration, natural ventilation,
and inter-room air movements in passively heated and cooled buildings.

A research-oriented program is needed to allow the detailed study of simultaneous

air, heat, and moisture transfer in and through a building with complex internal
architecture. A research program must be easily expandable to test the simula-
tion of new features. This requirement implies that the program should have a
simple structure and be well documented. The program should also be able to
handle small or very large problems and be portable so it can be used by many
researchers.

1.2 OBJECTIVES

The primary objective of this effort is to develop a research-oriented building

loads analysis computer program. This program should Include the best avail-
able algorithms for multi-room analysis of heat and air transfer. It must also
include a sufficient variety of building feature models to allow parametric
studies of the interactions of any proposed algorithms with the many complex
features present in real buildings.

[1] Hittle, D.C., The Building Loads Analysis and System Thermodynamics (BLAST)
Program, Version 2.0: Users Manual, U.S. Army Construction Engineering
Research Laboratory Report TR-E-153, June 1979.

[2] DOE-2 Users Guide, Version 2.1 , Lawrence Berkeley Laboratory Report
LBL-8698, May 1980.

1
1.3 SCOPE

The present version of this research program developed in the Environmental

Analysis Group of the National Bureau of Standards is called the Multi-Room
Loads Program (MRLP). It is the first phase in the development of a general
research program. It is based on the BLAST program which uses a very detailed
room heat balance technique. The MRLP adds to that technique an algorithm for
computing infiltration and inter-room air movements for arbitrarily complex
geometry. The MRLP uses a new input language, a revised program structure, and
is written in standard FORTRAN 77 to create a modifiable program. This report
gives an overview of the MRLP and a detailed discussion of the air movement
algorithm (section 2). The preliminary results of sample calculations are dis-
cussed (section 3). The limitations of the proposed algorithm and opportunities
for improved efficiency and other areas for further development are discussed
at the end of section 3. Detailed documentation of the MRLP is in preparation.

2
2 . METHODS

2.1 S IMULAT ION CAPAB IL IT IE S

The Multi-Room Loads Program uses detailed heat balance techniques to compute
room loads, as described below in section 2.2. The program uses a one-hour
time step. Transient features are modeled by a quasi-steady-state analysis
except for thermal conduction. The MRLP may be used to evaluate single "design"
days or use up to one year of measured weather data. A large number of occupant
usage features may be scheduled with hourly, daily, and seasonal variations.
The effects of temperature drifts in response to various control strategies can
be accurately evaluated. Other capabilities can be grouped by the three modes
of heat transfer.

Conduction: Transient conduction effects are modeled by "conduction transfer

functions," which are extensions of response factors. They model the heat stor-
age effects of walls, furnishings, and partitions between rooms. Movable insula-
tion is closely approximated by a massless insulation model with appropriate
controls. This movable insulation may be applied over windows or walls.

Radiation: The transmittance and absorptance of windows is computed from the

optical properties of the window panes. There may be up to four panes which
may be heat absorbing or have reflective films. Shading coefficients may be
used for windows which cannot be described in terms of their components. The
shadowing and distribution of sunlight on surfaces may be calculated in several
levels of detail up to the amount of light transmitted through windows onto
each surface of the room. The effects of "non-black" interior and exterior
surfaces can be modeled. There is a simple model to account for the energy
savings of daylighting.

Convection: The program uses variable interior and exterior convection

coefficients. It can compute infiltration and the natural flow of air between
rooms caused by wind and thermal forces. There is an alternative simple model
for infiltration with options of ventilation cooling, a whole house fan, and
controlled air mixing between two rooms.

The program structure is described in appendix A and the input in appendix B.

2.2 ROOM ENERGY BALANCE

The calculation of the room air temperature is based on a heat balance on the
room air. The convective gains from the room surfaces and internal sources
plus the sensible gains from incoming air plus the energy added by the air
handling system must equal zero.
\

CE + Z HCj*Aj*(Tj - TZ) + Z F^Cp^CTSi - TZ) + L = 0 (1)

where

CE = energy convected from internal sources

Aj = area of surface j

3
 HCj = convection coefficient of surface j
Tj = temperature of surface j
TZ = room air temperature
= mass flow of air through opening i
Cp = specific heat of air
TS^ = temperature of the air through opening i
L = energy from the air handling system

The summations apply to all surfaces/openings which are part of the room. The
values of the surface temperatures T j are obtained by evaluating a heat balance
,

at every suface of the room using an assumed room temperature. A better esti-
mate of the room temperature is then obtained by solving equation 1 for TZ.
Repeating this process leads to successively better approximations of the sur-
face and air temperatures. The process is rapidly convergent. Multiple rooms
can be handled within the iterative scheme given an appropriate data structure
and proper heat balances on the surfaces between rooms. Details of the surface
heat balances are beyond the scope of this report. They are essentially the
same as those Implemented in BLAST, version 3.0. The remainder of this report
will discuss the method for computing the air flows through envelope and
inter-room openings.

2.3 NATURAL AIR MOVEMENTS

Most building surfaces contain openings, either large or small, which permit
the movement of air through them whenever there is a pressure difference across
the surface. Pressure differences arise from naturally occurring wind and
thermal forces. The following discussion of air movements emphasizes the rela-
tionships to basic principles of fluid mechanics.

2.3.1 Wind Forces

Wind pressures on surfaces are proportional to the velocity head of the

undistributed air stream [3]. The wind pressure (PWj^) is thus calculated from
Bernoulli's equation as

PW^' = 0.5 * Da * V**2 ( 2)

The density of air is related to the barometric pressure (PB) and air temperature
(Ta) by

Da = 0.0034838 * PB/(Ta + 273.15) (3)

Wind speed (V) increases with height according to the formula

V = Vq * b * (Z/Zo)**a ( 4 )

[3] ASHRAE Handbook 1977 Fundamentals, p. 21.1.

4
where

Vq = wind speed at standard conditions

Z = height being considered
Zq = standard height of wind measurements (10 m)
a,b = terrain-dependent coefficients given below.

Terrain Parameters for Standard Terrain Classes [4]

Class Description a _b

1 Ocean or other body of water with at least 0.10 1.30

5 km of unrestricted expanse

2 Flat terrain with some isolated obstacles 0.15 1.00

(buildings or trees well separated from
each other)

3 Rural areas with low buildings, trees, etc. 0.20 0.85

4 Urban, industrial, or forest area 0.25 0.67

5 Center of large city (Manhattan) 0.35 0.47

The wind speed is evaluated at the highest point of the surface.

Wind direction effects are dependent on the building’s configuration and its
relationship to surrounding structures. No single correlation can truly model
directional effects on all buildings or even the effects on a single building
in different environments. The following correlation is a compromise derived
by averaging several published correlations. It accounts for the trends
observed in wind pressure measurements [5, 6].

Let Y = angle of incidence of wind -on the surface (0<=Y<=180.)

If(Y<=90.) PWi = (.75-1.05*Y/90.) * PW^’ (5a)

If(Y>=90.) PWj_ = (0.15*Y/90.-.45) * PW^' (5b)

[4] Sherman, M.H. and D.T. Grimsrud, "Infiltration-Pressurization Correlation:

Simplified Physical Modeling," ASHRAE Transactions Vol. 86, Part 2, 1980.
,

[5] Shaw, C.Y. and G.T. Tamura, "The Calculation of Air Infiltration Rates
Caused by Wind and Stack Action for Tall Buildings," ASHRAE Transactions ,

Vol. 83, Part 2, 1977.

[6] Shaw, C.Y., "A Method for Predicting Infiltration Rates for a Tall Building
Surrounded by Lower Structures of Uniform Height," ASHRAE Transactions,
'Vol. 85, Part 1, 1979.

5
2.3.2 Thermal Forces

Temperature differences cause air density differences which produce bouyancy

(or stack) effects. The magnitude of the force on a submerged surface (all
surfaces are submerged in air) is equal to the area times the pressure at the
centroid of the area. It is therefore necessary to know the pressure at the
centroids and the locations of those centroids. The effects of air density on
the pressure on any surface can be considered by a single simple formula. The
pressure at height is equal to the pressure at height Z 2 plus fluid density
times the gravitational constant times (Z 2 - Z^). Figure 1 shows a surface
separating two vertically adjacent rooms. The pressure differences (AP) across
the surface is Pj - P^. It is more useful to know this AP in terms of Pm and
Pj^ which are the reference pressures of rooms n and m.

Pj = Pm + * g * (Z„ - Zj) (6a)

Pi = Pn + Dn * g * (Zn - ^l) (6b)

The same formulae apply to a vertical surface as in figure 2. They also apply
to an envelope surface where Pj is the pressure of the outside air (figure 3).
Ambient air variables are substituted for the room m variables

Pj = Pa + Da * 8 * (Za - Zj) (6c)

Za is the reference height for air pressure measurements. Its location is

arbitrary. The room reference height, Z^ and Zjjj, are also arbitrary. All are
chosen for convenient or physically meaningful representation. The absolute
values for the room pressures (Pn» Pm) pressure (P^) are also
arbitrary since only pressure differences cause air flow.

The pressure difference across surface i is

AP = Pj - Pi

" Pm - Pn + D„*g*(Zm-Zj) '

= Pm - Pn + PSl (7)

where

PSi = Djjj*g*(Z„j-Zj) - Dn*g*(Zn-Zi)

and is the "stack pressure" on surface i.

The height of the centroid of a triangle or rectangle is the average of the

heights of the vertices. Since the stack effect also occurs within the thick-
ness of a horizontal surface and since it is necessary to leave no building
internal height unaccounted, Zi and Zj in figure 1 should both be replaced by
their average.

6
 E
INI hij INI
I
I I
I
I
' I
I
I
' >
I
I

I
1 I
1
I

I
! I I

I I I I

I I I I

I I I I

n
ROOM

Figure 1. Vertically adjacent rooms

7
Figure 2. Horizontally adjacent rooms

8
\\\\<\\\\\^

Figure 3. Envelope surface

9
2.3.3 Flow Equations

The flow through small openings in a surface is based on the orifice equation

F=C*A*D* / 2 * AP/D (8)

where A is the observed opening area and where the flow coefficient, C, is a
function of the Reynolds number and the ratio of the opening size to the entire
surface. When the opening is small, C equals 0.6 for a wide range of Reynolds
numbers [7]. The MRLP assumes this value of C. The user may modify c by using
an effective flow area (FA) which differs from the actual opening area. Experi-
mental results indicate that flow through real surfaces is a combination of
orifice flow (Fo:/AP) and capillary flow (F“:AP) with an exponent of about 0.65
[8] In order to retain the concept of opening area and provide a simple
grouping of all other terms in equation 8, C is multiplied by /T and D divided
by /D to give

F^ = 0.848 * FA^ * /F * AP**Xj^ (9)

where FA^ is the effective flow area of the openings in the surface and D is
the density of Incoming air. The pressure difference for envelope surfaces is

AP = PWi + PSi - Pn (10a)

and the AP for inter-room surfaces is

AP = PSi + Pm - Pn (10b)

The surface flow equation becomes

Fi = 0.848 * FAi * * AP**Xi (AP > 0) (Ha)

or

Fi = -.848 * FAi * * (-AP)**Xi (AP < 0). (Hb)

For each room there must be a mass balance

Fs + S Fi = 0 (12)

where Fg is the net mass flow into the room from the air-handling system.
These mass balances form a system of non-linear algebraic equations which can
be solved for the room pressures.

[ 7 ]
ASHRAE Handbook 1977 Fundamentals , p 4.10.

[8] ASHRAE Handbook 1977 Fundamentals, p .21.4.

10
2.3.4 Solution of the Flow Equations

There are several methods for solving such a set of equations. Most involve
computing the following partial derivatives:

3Fi
= -Xi * F^/AP (13a)
9Pn

3Fi
= Xj * Fj/AP (inter-room surface) (13b)
9 Pm

or

9 Pi
= 0 (envelope surface) (13c)
9 Pm

Note that it is not necessary to evaluate AP**(X£-1) to compute these

derivatives. The simplicity of the derivative calculation indicates that
methods which approximate the derivative are not necessary. Note also that as
AP approaches zero these derivatives cannot be evaluated. This condition
applies for large openings which cannot support a large pressure difference
(except with very high flow rates). This a potential source of trouble which
must be considered.

Two methods for solving the air flow equations were studied. First was the
standard Newton's method for systems of equations [9]. This method was found
to give fast solutions in several test cases but to be non-convergent in other
very similar cases. Convergence could be obtained in those cases by the choice
of a suitable relaxation coefficient. The tests showed no pattern of non-
convergence which would allow prediction of a usable relaxation coefficient.

The second method tested was a simple iterative scheme from Conte and De Boor
[10]. It called for successively approximating each zone pressure according to

Pn^^'*'^^ = ^ Fi/2
^
—
9Pn
(14)

where k is the iteration number. This method was often slower than Newton's
method, but it converged for all of the simple test cases. It was therefore
chosen for use in the MRLP.

[9] Stoecker, W. F., Design of Thermal Systems , McGraw-Hill, 1971, pp. 92-93.

[10] Conte, S.D. and De Boor, C., Elementary Numerical Analysis, an Algorithmic
Approach, McGraw-Hill, 1972, p. 88.

11
2.4 SAMPLE PROBLEM

A study of the combined multi-room heat balance with air flow simulation was
done by using a "typical design" townhouse [11]. The floor plan is shown in
figure 4. and figure 5 is a section of the townhouse. There is a total of
(1315 ft^) 122 m^ living area on the two levels and (10850 ft^) 307 m^ of
conditioned space.

The MRLP input deck describing this building is given in appendix C. The
building is divided into 5 thermal zones: the first floor and stairwell (room
1), each of the three bedrooms (rooms 2, 3, and 4), and the attic (room 5).
Conductive heat transfer was modeled with 98 surfaces, and air flows were
modeled in 37 openings. The model probably uses much more detail than is neces-
sary to obtain a reasonable estimate of the energy requirements, but the detail
is useful in testing the capabilities of the program.

The opening areas were assumed according to ASHRAE [12] which suggests openings
of about .02 percent to .05 percent of the total surface area for frame walls
and .01 percent to .04 percent for ceilings. The largest values were chosen
for the test building. Inter-room partitions and floors were also given .05
percent openings. Additional inter-room opening areas of 0.046 m^ (0.5 ft^) for
the (almost closed) doors were used in the base case. The attic had a soffit
area of 0.186 m^ (2.0 ft^) on the north side only and a porosity of .10 percent
for the roof.

[11] Hastings, S.R., "Three Proposed Typical House Designs for Energy
Conservation Research," National Bureau of Standards Interagency Report
(NBSIR) 77-1309, 1977.

[12] ASHRAE Handbook 1977 Fundamentals, p. 21.7.

12
 Q£
O

ft^)

ft^
(10,840

(1315

307

-
122

volume

area

Total Total

.001

Figure 4. Townhouse floor plan

13
3. RESULTS AND DISCUSSION

3.1 WIND DIRECTION EFFECTS

Figure 6 shows the results of three test computer runs for a winter weather
condition with only the wind direction changed. When the wind blew perpendicu-
lar to the exposed sides of the house (cases 6a and 6c), an infiltration rate
of just over one air change per hour (ach) was produced. The location of the
soffit is the primary reason for the difference in results between the cases
of north and south winds. When the wind blew parallel to the townhouse (case
6b), infiltration was only .22 ach. This was due mostly to stack effect.
These values compare favorably to ASHRAE design air change rates for residences
[13]. ASHRAE recommends 1 ach for rooms with windows or exterior doors on one
side and 1.5 ach for rooms with windows or exterior doors on two side. For
rooms with weatherstripped windows or with storm sash, two thirds of these
values should be used. The estimated loads (case a = 7.41 kW, case b = 4.68 kW,
case c = 7.87 kW) show that infiltration and wind direction effects on infil-
tration are important components of total heating load of this typical structure
Computation times varied from about 7 to 15 seconds per simulation day.

3.2 INTER-ROOM OPENINGS

Figure 7 shows four cases comparing the effects of the size of inter-room
openings on the air change rates, loads, and calculation times. The small
openings simulation (case 7a) assumes 0.009 m^ (0.1 ft^) openings between rooms
instead of the standard 0.046 m^ (0.5 ft^) (case 7b) openings. The third case
(7c) simulated infinite openings between rooms by modeling the four rooms of
the townhouse as a single room. There is no significant change in the calcu-
lated loads in these three cases (7.34 kW, 7.41 kW, and 7.42 kW) which go from
the smallest reasonable inter-room openings to no inter-room flow resistance.
Although the total load in the small openings case is only one percent smaller
than the base case, there are much greater differences in the loads of the
Individual rooms (room 1 = +1 percent, room 2 = +6 percent, room 3 = -8 percent,
room 4 = -8 percent). This is noticeable in real buildings where closed-off
rooms on the windward side become colder than the leeward rooms.

The fourth case (7d) used the air change rate computed for the single room
model to compute loads based on a known air change rate instead of on the pres-
sure and air flow balance model. This gave the same load, as it should, and
required only 3.9 seconds/day simulation time. The time of other cases in
excess of this amount is the computation cost of the detailed air flow algorithm
These calculation times are good only for relative comparsions because they are
greatly affected by the numbers of surfaces, rooms, and openings and the conver-
gence limits of the simulation. Comparing case 7c to 7d, the increase in com-
putation time of 18 percent appears justified because of the importance of
infiltration in the total heating load of the building. However, in case 6b
infiltration is making a much smaller contribution to the load and calculation

[13] ASHRAE Handbook 1977 Fundamentals, p. 21.5, table 1.

15
 6.7
NORTH WIND m/s
(15 mph)
-14.4°C
(6°F)

I.02 ach
6.7
7.41 kWh
II. 1 sec/day
m/s

WEST WIND

0.22 ach

4.68 kWh
14.7 sec/day
6.7 m/s

SOUTH WIND

1.17 ach
7.87 kWh
7.0 sec/day

Figure 6. Wind effects - computer air flows through

major building surfaces

16
21 JAN SMALL INTER-ROOM OPENINGS

6.7 m/s
-14.4°C

1.00 ach

7.34 kWh
8.2 sec/day

STANDARD OPENINGS

I. 02 ach
7.41 kWh
II. 1 sec/day

SINGLE ZONE

1.03 ach

7.42 kWh
4.6 sec/day

CASE 7d
AIR CHANGE METHOD: using same ach
gives same load
3.9 sec/day

Figure 7 . Inter-room openings

17
time is almost 15 seconds. In this situation the computational cost is not
justified.

Note in the multi-room cases that as the areas of the inter-room openings
increase, the calculation time goes up dramatically. In fact, it was not
possible to simulate a case with 0.186 m^ (2.0 ft^) openings between rooms.
This situation called for the development of a simple test case to study the
effects of inter-room opening area on the calculation technique. The results
are shown in figure 8. The first case represents a single room consisting of
two walls, each with opening areas of 0.01 m^ (0.11 ft^) one facing into a
6.7 m/s (15 mph) wind and one facing away from it. The flow through the room
is computed to be 2.98 m^/h (105. ft^/h) and requires only 2 Iterations. The
next case shows the original room divided into two rooms by a partition contain-
ing a 0.01 m2 opening. This reduces the air flow rate to 1.95 m^/h (68.9 ft^/h)
and raises calculation time to only 3 iterations. As the area in the partition
opening is doubled in succeeding cases, the air flow rate approaches the single
room flow rate while the calculation time increases rapidly. For example, the
partition with 0.08 m^ opening area has a flow within one percent of the single
room value but requires over twelve times as long to calculate. The next larger
case, 0.16 m^, could not be calculated within the iteration limit. However, it
is obvious that assuming the single room air flow value would be correct to
within one percent.

This situation may be considered a problem or an opportunity. Doorways,

ductwork, and the lack of vapor barriers cause interior partitions to have much
larger flow areas than exterior walls in typical low-rise construction. If
those openings are to be modeled exactly, another algorithm must be developed
to solve the air flow equations. On the other hand, the small effect that
large inter-room openings have on the infiltration and the loads indicates that
it may be possible to separate the problem. Infiltration could be computed by
assuming that relatively large inter-room openings are effectively infinite.
Then the inter-room flows would be computed based on the net Infiltration into
each room. High-rise buildings are usually constructed to have relatively small
openings between floors to reduce the stack effect, but all the rooms on a
single floor are usually well coupled. Therefore, some separation should still
be possible.

3.3 NATURAL VENTILATION

Other tests of the two-room model indicated that it is the relative size of the
inter-room and envelope openings, not the absolute size, which is affecting
computation time. Solutions were reached rapidly for all cases where the
envelope openings were larger than the partition openings. This property per-
mitted modeling the soffit of 0.186 m^ (2.0 ft^) with no problems. It suggested
that it should be possible to simulate the condition where the envelope openings
are about the same size as the inter-room openings. This is the case when win-
dows have been opened to get natural ventilation. Figure 9 shows how ventila-
tion can affect the building conditions. A warm, sunny fall afternoon was
chosen as the test weather condition. The top figure (case 9a) shows the air
flow rates and room temperatures for the standard configuration of envelope and
partition openings. Note that the south bedroom is at 35.8 C (96.4 F). In

18
WALL PTN WALL aOW ITERATIONS

2.98 mVhr
a
0.01 FLOW AREAS 0.01 m2

I 1.95
I
O01 0.01 aoi

I 1 2.56
I I I
001 ao2 aoi

I
2.85 8
I
O01 ao4 aoi

1
2.95 25
I
O01 0.08 aoi

I unsolved >80
I I
0.01 0.16 0.01

Figure 8. Effect of interior partition

19
21 SEP
3 P.M. CLOSED WINDOWS — 3.3 m/s
26.7X
(80°F)

0.41 ach
16.0 sec/day

OPEN WINDOWS SAME CONDITIONS

5.26 ach

7.5 sec/day

Figure 9. Natural ventilation

20
case 9b windows and Inter-room doors have all been opened to 0.186 (2.0 ft^).
This produces greater than 5 ach and reduces the south bedroom temperature to
29 C (84.2 F).
0
3.4 FUTURE DEVELOPMENTS

The MRLP air movement algorithm is theoretically capable of modeling high and
low openings in a single wall. This was not tested in the sample problem. It
has potential use in the case where one room is significantly hotter than the
other and a stack effect is developed between the high and low openings. This
condition occurs in a Trombe wall where the space between the cover and the
wall may be considered and extremely narrow room. The narrowness of the room
adds a further resistance to air flow (besides that of the wall openings) which
should be added to the air flow model.

The air flow model does not address two important classes of openings. One is
the doorway where bi-directional flows can occur. It may be possible to handle
this by using a high and a low opening. The other case is the opening between
floors as it is affected by thermal stratification. A very rapid air exchange
will take place between a warm room below a cold room, whereas almost no air
will move between a cold room below a warm one. This diode effect will add a
further complication to the already non-linear air flow equations. Two openings
may set up a large convection pattern with much greater air movement than a
single opening of equivalent area.

The possibility of separating the envelope from the inter-room air flow
calculation suggests a way of handling these additional features. ASHRAE [14]
shows that the flow through an orifice is no longer proportional to the square
root of AP at very low Reynolds number. In fact, it becomes proportional to
AP. The Reynolds number of air flow between rooms connected by an open door
appears to be typically in a transition region. If it can be modeled to suffi-
cient accuracy by the linear assumption, solution of the multi-room flow problem
would be greatly simplified. Only simultaneous linear equations would have to
be solved. It may also be possible to superimpose bi-directional effects.

The general problem of thermal stratification in the room air should also be
addressed. Two simple models are possible. The first would assume a linear
temperature profile from the room's coldest point at the floor to its warmest
point at the celling. The slope of this profile would be determined experimen-
tally. The second method would divide the room air into two parts: a warm
upper layer and a cooler bottom layer. Energy would be convected between these
layers and from walls and internal sources to each layer according to experi-
mentally determined coefficients. Heat balances could then be performed for
each layer.

The current air flow model and the possible extensions all require experimental
validation to show if they are sufficiently accurate. If it is also shown that
the computation cost is low enough, the method can be recommended for use in

[ 14 ] ASHRAE Handbook 1977 Fundamentals, p 4.10.

21
energy analysis design tools. It would then be necessary to develop a data
base of openings areas and flow exponents for different building constructions
for design ;v^5e.

22
 APPENDIX A: PROGRAM STRUCTURE

The program is written almost entirely in FORTRAN 77. This is a standard

language which has been or is being implemented on most large and medium size
computers. The handling of character data and direct access files help to make
this language machine-independent. The MRLP should be portable except for some
conversions from single to double precision and replacing the use of BLAST
weather tapes. The PARAMETER feature of FORTRAN 77 is used extensively in
modifiable array dimensions so the MRLP can handle different size problems.

The MRLP is divided into two subprograms: the input processor (IP) and the
loads calculator (LG). This structure is illustrated in figure Al. The IP
checks the user input, prepares input files for the LC, and reports the array
sizes that the LC will require. The LC performs the building loads simulation
and reports the results. This separation of subprograms allows a reduction in
computer core requirements. The IP uses minimal array sizes because it processes
the input one element (surface, schedule, material, etc.) at a time. The LC
uses large arrays which consider these elements simultaneously. The LP and LC
may be run as separate jobs when it is necessary to recompile the LC for correct
array sizes, or they may be run as a single job if the LC is known to have
sufficiently large arrays.

The input processor is divided into four major sections. These correspond to
the input language sections which they process. This sectioning allows further
reductions in core requirements by the use of overlays (or segments). A small
input processor is important to allow the fast job turn-around necessary to
quickly debug the user input deck. Each section contains several subroutines
which process the input language "keys" plus others for special elements of
certain keys. The two largest of these special routines are the groups of sub-
routines which evaluate the conduction transfer functions and the window optics.

The loads calculator is divided into three sections. The first section reads
the file created by the IP and fills the LC arrays. It checks for sufficient
array space and performs some checks which require that all data for the problem
description be present in the arrays. The second section uses the geometric
and optical descriptions of the building surfaces to compute solar gain factors
for various periods of the year. The third section does the actual heat balances
necessary to compute the multi-room loads on a hour-by-hour basis.

A-1
 LOADS CALCULATOR

Figure A-1. MRLP structure

 APPENDIX B: SIMULATION INPUT

It is intended that the MRLP input deck be an on-line file which is created and
modified by a text editing system. It can also be a deck of punched cards.
There is a simple, consistent input structure. The input is based on a keyword
system (keyword = value) to produce a free-fomat, self-descriptive input lan-
guage. In some ways it is similar to NAMELIST input which is a feature of some
versions of FORTRAN. The keyword system is organized in a logical heirarchy
which groups related concepts and data. Defaulting is used extensively. This
applies to individual values and to groups of values representing entire elements
of building operation or description. Some groups of often used data can be
named and then used many times by referring to the name.

The input deck for the test case (appendix C) demonstrates the features of the
input language. On the first page of the townhouse input deck nine lines below
"LIBRARY" is the line:

MATL(NAME=GYP/DESC=' GYPSUM WALLBOARD /K=. 42/D=100. /CP=. 2)

'

This line of input describes the conductivity (K=.42), density (D=100.), and
specific heat (CP=.2) of a gypsum wallboard, and gives this material the name
GYP. Other properties of this material which are not mentioned, such as its
roughness, solar absorptance, and thermal emittance, are given default values
by the input processor. Twelve lines further down is the input:

CONS (NAME=PTN-S/DESC=' PARTITION - STUD AREA’/

MATL=. 0417, GYP,. 292, WD-S,. 0417, GYP)

This describes the partition walls at the studs and gives this construct the
name PTN-S. It consists of two .0417 ft (.5 in) thick layers of the already
defined GYP (wallboard) with a 0.292 ft (3.5 in) thick layer of WD-S (softwood)
between them. This construct is then used in the input at the bottom of the
second page

SRF( C0NS=PTN-S/ IN/OP/BS/AZM=90/SIZE=3 .5,8.5)

This line is describing the stud area of the partition along the east side of
the first zone of the townhouse. IN means this is an interior rather than enve-
lope surface. This will determine the type of surface heat balance performed.
OP means it is opaque instead of transparent and determines which algorithm is
used to model the effects of solar energy and other short-wavelength thermal
radiations on the surface. BS means it is a base surface. That is, its loca-
tion and orientation are not based on another surface. This surface faces east
(AZM=90) and is 3.5 ft wide by 8.5 ft high (SIZE=3.5,8.5) . This width is the
combined width of all studs in the wall.

B-1

I
'JK"
APPENDIX C; TOWNHOUSE INPUT DECK

PROJECT C
RC<XREF/DEH/UIN=EN6LISH/U0UT=METRIC/
DESC=M ZONE HASTINGS TOWN HOUSE PLUS ATTIC')
RPT<RNOS = 8» 21)
LOC(DESC='COLUMBIAj M0» '/LATD=38*9/L0N6=?2,3/TZ=6)
GRND(GRT=56»55..56j!=i9j61f65f68j70f69f A6»62f59)
DDY(DESC=' COLUMBIA WINTER /HIGH=26/L0W=6/WB=20/DATE=21JAN/PRES=405/
'

WS=1320/DIR^0/CLR=1 ./SAT)
1

LIBRARY C
LIB(NMATL=32/NC0NS=32)
MATL(NAME=C-HW/DESC='CONCRETE - SAND i GRAVEL /K=l 104/D=140 /CP= .2 ' . .

MATL(NAME=R5/DESC='R5 POLYSTYRENE /R=S' .

HATL NAME=B1/DESC= 'AIRSPACE RESISTANCE '/R= 91

( .

MATL(NAME=R11/DESC='R11 INSULATION'/R=ll .

MATL(NAME=R19/DESC='R19 INSULATION' /R=19 ,

MATL NAME=WD-S/DESC= SOFTWOOD /K= 0675/D=32 /CP= 33

( ' ' . . .

MATL(NAME=WD-H/DESC=' HARDWOOD '/K=.092/D=45./CP=. 3)

MATL NAME=GYP/DESC^ GYPSUM WALLBOARD /K= 42/D= 100 /CP= 2
( ' ' . . .

MATL(NAME=SHTH/DESC=' INTERMEDIATE DENSITY BUILDING BOARD SHEATHING'/

K=. 034/11=22. /CP=. 31)
MATL (NAME=AL/DESC=' ALUMINUM SIDING' /K=128 /D=171 ./CP= .214
.

MATL(NAME=CPT/DESC='CARPET X FIBROUS PAD'/R=2.08)

MATL(NAME=E4/DESC='CEILING AIRSPACE'/R=1 .0)
MATL<NAME=SPDS/DESC='SINGLE PANE» DOUBLE STRENGTH GLASS'/
GLASS/R=.0236/TRNS=.87)
MATL(NAME=DIRT/K=.1/D=65./A=.70)
MATL (NAME=ASH/DESC=' ASPHALT SHINGLES '/K=18.16/D=70/CP=. 3)
CONS(NAME=PTN-A/DESC='PARTITION - AIR SPACE AREA'/
MATL=.0417»GYP» 0»B1? .0417»GYP)
CONS(NAME=PTN-S/DESC='PARTITION - STUD AREA'/
MATL=.0417»6YPj .292 »WD-Sj .0417jGYP)
CONS(NAME=WL-I/DESC='EXTERIOR WALL - INSULATION AREA'/
MATL=.0052»AL> .0417»SHTHj 0»R11j .0417»6YP)
CONS(NAME=WL-S/DESC='EXTERIOR WALL - STUD AREA'/
MATL=.0052»AL» .0417»SHTH» .292jWD--S» .041?jGYP)
CONS (NAME=WDW/DESC=' SINGLE PANE WINDOW OR GLASS DOOR /MATL=0 SPDS ' f

C0NS(NAME=CLG2-I/DESC='CEILIN6 UNDER ATTIC - INSULATION AREA'/

MATL=0jR19f .0417 jGYP)
C0NS(NAME=CLG2-T/DESC='CEILING UNDER ATTIC - TRUSS AREA'/
MATL=.458fWD-Sf .0417»GYP)
C0NS(NAME=FLR3-I/DESC='ATTIC FLOOR - INSULATION AREA'/
MATL=.0417»GYP» 0»R19)
C0NS(NAME=FLR3-T/DESC='ATTIC FLOOR - TRUSS AREA'/
MATL-^.0417fGYPf .458^WD-S)
C0NS(NAME=CLG1-A/DESC='FIRST FLOOR CEILING - AIR SPACE AREA'/
MATL-^0»CPT» .0625»WD-S> 0»E4j .0417fGYP)
C0NS(NAME=CLG1-I/DESC='FIRST FLOOR CEILING - INSULATED AREA'/
MATL=.0625>WD-S» 0»R11» 0»B1 .0417fGYP)
j

C0NS(NAME=CLG1-T/DESC='FIRST FLOOR CEILING - TRUSS AREA'/

MATL=0»CPT» .0625»WD-S» .792fWD-S» .0417^GYP)

C-1
 C0NS(NAME=FLR2-A/DESC='SEC0ND FLOOR FLOOR - AIR SPACE AREA'/
MATL^»0417fGYPf 0jE4f »0625»WD-B» O.CPT)
C0NS(NAME=FLR2-T/DESC='SEC0ND FLOOR FLOOR - TRUSS AREA'/
HATL^»0417y6YPf 792»UD-Sf »0625fUD-S» OjCPT)
C0NS(NAME=FLR2-I/DESC='SEC0ND FLOOR FLOOR - INSULATION AREA'/
HATL=»03125fWD-SF 0»R11> »0625 jWD-S.. OjCPT)
CONS(NAME=PTN-I/DESC='PARTITION - INSULATION AREA'/
MATL=.0417»GYP, OfRll» »0417fGYP)
CONS(NAME=ROOF-A/DESC='ROOF - BETWEEN TRUSSES'/
MATL^,01.04f ASH» 04l7fWD-S)
CONS<NAME=ROOF-T/DESC='ROOF - AT TRUSSES'/
MATL=^0104..ASH^ 04176?WD-S» f292»WD-S)
CONS (NAME=BLAB/DESC=' SLAB FLOOR'/
MATL:--,5fIiIRTf 0fR5» »333i.C-HW. OjCPT)
CONS (NAHE=HOOR/DESC=' FRONT ENTRY /MATL=» 146 jWD-H)
'

DS(NAME=LBWD/DESC=' LIGHTS - BEDROOMS - WINTER - WEEKDAY'/

FFC=6*0»2**5»9*0>6*»5j0)
DS(NAME=LKWD/DESC='LIGHTS - KITCHEN - WINTER - WEEKDAY'/
FFC=6*0»2*»8>9*0>2**7f4**4»0)
DS<NAME=LLWD/DESC='LIGHTS - LIVING ROOM - WINTER - WEEKDAY'/
FFC=6Y0f2*.2>9)t:0f2*»3.3)|c»7.. »3 j0)
DS( NAME=CONST/DESC= 'CONSTANT '/FFC=24*D
DS(NAME=EKWD/DESC='EOIJIPMENT - KITCHEN - WINTER - WEEKDAY'/
FFC=6*0j2*»5»9*0jlf2!»:»2j4*0)
WS(NAME=LBW/DESC='LIGHTS - BEDROOMS - WINTER'/WD=LBWD)
WS(NAME=LKW/DESC='LIGHTS - KITCHEN - WINTER'/WD=LKWD)
WS(NAME=LLW/DESC='LIGHTS - LIVING RM - WINTER'/WD=LLWD)
WS(NAME=CONST/DESC=' CONSTANT '/ALL=CONST)
WS(NAME=EKW/DESC='EQUTPMENT - KITCHEN - WINTER'/WD=EKWD)
CPFL NAME=HT/DESC= 'NORMAL HEATING /TMP=70 70/CAP=1 0
( ' j j

CPFL NAME=SB/DESC= SETBACK HEATING /TMP=65 65/CAP= 1 0

( ' ' . f

CPFL(NAME=CL/DESC='COOLING'/TMP=65 j65j78»78/CAP=1»0»O>-1)
CDS (NAME=WD/DESC=' WINTER - WEEKDAY '/CPFL = 6!|!SB.2)KHTj9*SB.6*HTj SB)
CWS(NAME=WINTER/DESC=' WINTER '/WD=WD)
D
BUILDING C
SIM<SLDS=2/HTB=1/FL0W=.0001 »0001>80 »9/CNVG=»Ol5 »001» *1 j12)
j j

$ NORTH SIDE SHADING

DSS(AZM=90/0RG=»29»34j0/SIZE=2j17,17)
DSS(AZM=270/0RG=19.71?36»0/SIZE=2..17*17)
$ SOUTH SIDE SHADING
DSS(AZM=270/0RG=20f2.0/SI7E=17»2.2)
3

ZONE L

$ FIRST FLOOR + STAIRWELL

GE0M(NAME=FLR1/Z=»67/CLHT^8.5)
$ WEST PARTITION + STAIRWELL - 311.
SRF(C0NS=PTN-A/IN/0P/BS/AZM=270/SIZE=32.62f8.5)
SRF(C0NS=PTN-S/IN/0P/BS/AZM=270/SIZE=4.00»8.5)
$ EAST PARTITION - 272.
SRF(C0NS=PTN-A/IN/0P/BS/AZM=90/SIZE=28.5»8.5)
SRF(C0NS--PTN-S/IN/0P/BS/AZM=90/SIZE-^3.5*8.5)

C-2
$ INTERNAL PARTITIONS - 167 .4 »24 »9 80
» *

SRF(C0NS=PTN-A/IN/OP/BS/AZM=180/SIZE^19»44f7»67/Z0NE=liBIiRM)
SRF<C0NS=PTN-S/IN/0P/BS/AZM=180/SIZE=2»39,7»67/Z0NE=MBIiRM)
AFS(AREA=»5/ZS=12.5)
BRF<C0NS=PTN-A/IN/0P/BS/AZM=90/SIZE=2.9f7.67/Z0NE=BDRM2)
SRF(C0NS=PTN-S/TN/0P/BS/AZM-=90/SIZE=t35»7»67/Z0NE=BDRM2)
AFS(AREA=»5/ZB=12.5)
SRF(C0NS=PTN-A/IN/0P/BS/AZH=0/SIZE=9.36j7»67/Z0NE=BDRM3)
SRF(C0NS=PTN-S/IN/0P/BS/AZM=0/SIZE=1»15»7.67/Z0NE=BDR«3)
AFS(AREA=.5/ZS=12»5)
i INTERNAL MASS
MASS(C0NS=PTN-A/AREA=795)
MASS(C0NS=PTN-S/AREA=98)
% NORTH EXTERIOR WALL - 165.1
SRF(C0NS=WL-S/EX/0P/BS/AZM=0/SIZE^2.0?8.5/0RG^2.29j34j .67)
AFS(AREA=.050)
AFS(AREA=.047/ZS=.l)
SRF(C0NS=WL-I/EX/0P/BS/AZM=0/SIZE=17.42j8.5/0R6=19.71r34» .67)
$ GLASS DOOR
SRF(C0NS=WDU/EX/TR/SS/SIZE=5.73j6.72/0RG=2.24>.47j-.17)
$ WINDOW
SRF(C0NS=WDW/EX/TR/SS/SIZE=4.98j3.83/0RG=12.2j3.5»-.17)
$ SOUTH EXTERIOR WALL - 222.9
SRF(C0NS=PTN-I/EX/0P/BS/SIZE=5.3»8.5)
0SC(TZP=.3/TAP=.7)
SRF<C0NS=WL-I/EX/0P/BS/AZM=180/SIZE=6.67j8/0RG=3.75j3.67j 67)
SRF(C0NS=D00R/EX/0P/SS/SIZE^3.1f6.97/0RG^^0» .IfO)
SRF(CONS=WL-S/EX/OP/BS/AZM=180/SIZE=7»8/ORG=10.4 ?jO> 67)
SRF(C0NS=WDW/EX/TR/SS/SIZE=5.73j6.72/0RG=.63f .33J-.17)
SRF(C0NS=WL-I/EX/0P/BS/AZM=270/SIZE=2.77 j8/0RG=10.42j 2.77» .67)
SRF(C0NS=WL-S/EX/0P/BS/AZM=270/SIZE=.9f8/0RG=l0.42f3.67j .67)
SRF(CONS=WL-I/EX/OP/BS/AZM=90/SIZE=2 j8.5/ORG=17.4jOj 67)
SRF(C0NS=WL-I/EX/0P/BS/AZM=180/TILT=90/SIZE-2.6j8.5/0R6=17.4f2. .67)
AFS<AREA=.060)
AFS(AREA=.058/ZS=.l)
$ FLOOR
SRF(CONS=SLAB/EX/0P/BS/AZM=18O/TILT=180/SIZE=20f31.6/ORG=0.34j .67)
0SC(TGP=1)
$ CEILINGS - 129. 1j 216. 4j 129.? 104.3
SRF(C0NS^CLG2-I/1N/0P/BS/TILT=0/SIZE=11.97?10/Z0NE=ATTTC)
SRF(C0NS=CLG2-T/IN/0P/BS/TILT=0/SIZE=.94?10/Z0NE=ATTIC)
AFS(AREA=.049/ZS=16.5)
SRF(C0NS=CLG1-A/IN/0P/BS/TILT=0/SIZE=19.27?10/Z0NE=MBDRM)
SRF(C0NS=CLG1-T/IN/0P/BS/TILT=0/SIZE==2.37?10/Z0NE=MBDRM)
AFS(AREA=.167/ZS=8.5)
SRF<C0NS=CLG1-A/IN/0P/BS/TILT=0/SIZE=11.49?10/Z0NE=BDRM2)
SRF(C0NS^CLG1-T/IN/0P/BS/TILT=0/SIZE-=1.41?10/Z0NE=BDRM2)
AFS(AREA=.065/ZS=8.5)
SRF(C0NS=CLG1-A/IN/0P/BS/TILT=0/SIZE=9.29»10/Z0NE=PDRM3)
SRF(C0NS=CLG1-T/IN/0P/BS/TILT^0/SIZE-1.14?10/Z0NE=BDRM3)
AFS(AREA=.080/ZS=8.5)
SRF(C0NS=CLGl-I/EX/0P/BS/TILT=0/AZM=0/SIZE=7»2/nRG=10.4»2?8)

C-3
% SCHEDULES
LIT(WS=LKW/CAP=»7)
LIT(WS=LLW/CAP=1*2/RAD=»8/UIS=,1)
EQP(US=EKW/CAP=5»9)
EQP(WS=C0NST/CAP=1»5)
CTRL(CWS=WINTER/)
3

ZONE C
$ MASTER BEDROOM
GE0M(NAME=MBDRM/Z=9.17/CLHT=8)
$ PARTITIONS - 76»5j 137. 167. 41.4
SRF(C0NS=PTN-A/IN/0P/BS/AZM=180/SIZE=8.88?7.67)
SRF(C0NS=PTN-S/IN/0P/BS/AZM=270/SIZE=1.09»7.67)
SRF(C0NS=PTN-A/IN/0P/BS/AZM=90/SIZE=15.96f7.67)
SRF(C0NS=PTN-S/IN/0P/BS/AZM=90/SIZE=1.96»7.67)
SRF(C0NS=PTN-A/IN/0P/BS/AZM=^0/SI7E = 19.44»7.67/Z0NE=FLR1)
SRF<C0NS=PTN-S/IN/0P/BS/AZM=0/SIZE=2.39»7.67/Z0NE=FLR1)
AFS(AREA=.5/ZS=4.
SRF(C0NS=PTN-A/IN/0P/BS/AZM=0/SIZE=4.8j7.67/Z0NE=BDRM2)
SRF(C0NS=PTN-S/IN/0P/BS/AZM=0/SIZE=.6f7.67/Z0ME=BDRM2)
AFS(AREA=.019/ZS=4.
SRF CONS=PTN-A/ IN/OP/BS/SI ZE=9 8
( .

i SOUTH WALL - 189.5

SRF(C0NS=WL-I/EX/0P/BS/AZM=270/SIZE=1.78»8/0RG=0.2r0)
SRF(C0NS=WL-S/EX/0P/BS/AZM=270/SIZE=.22 j8/0RG=0j1.78j0)
SHD(LFIN/SIZE=8f 10/0R6=0f0j0)
SRF(CONS=WL-I/EX/OP/BS/AZM=180/SIZE=9.6 j8/ORG=OjO»0)
SRF(C0NS=WL-S/EX/0P/BS/AZM=180/SI7E=1.2»8/0RG=9.6j0,0)
SRF(C0NS=WL-I/EX/0P/BS/AZM=90/SIZE=1.78j8/0RG=10.75j0j0)
SRF(C0NS=WL-S/EX/0P/BS/AZM=90/SIZE=.22j8/0RG=10.75.1.78j 0)
SRF(C0NS=WL-I/EX/0P/BS/AZM=180/SIZE=7.92»8/0RG=10.75j2f0)
SRF(C0NS=WDU/EX/TR/SS/SIZE=5.73?6.72/0RG=0 .25j-.17)
j

SRF(C0NS=WL-S/EX/0P/BS/AZM=180/SIZE=.97»8/0R6=18.67j2j0)
AFS(ARFA=.055)
AFS(AREA=.054/ZS=8.
$ FLOOR
SRF(CONS=FLR2-I/EX/OP/BS/AZM=180/TILT=180/SIZE=3.27>10.42/ORG=0»3.67f0)
SRF(C0NS--FLR2-T/EX/0P/6S/AZM=1.B0/TILT = 180/SIZE=.40f 10.42/0RG=0f .40.-0)
SRF(C0NS=FLR2-A/IN/0P/BS/TILT=180/SIZE=19.27»10/Z0NE=FLR1)
SRF(C0NS=FLR2-T/IN/0P/BS/TILT=180/SIZE--2.37f 10/Z0NE=FLR1)
AFS(AREA=.167)
$ CEILING - 256.7
SRF(C0NS=CLG2-I/IN/0P/BS/TILT=0/SIZE=23.8j 10/Z0NE=ATTIC)
SRF(CONS=CLG2-T/IN/OP/BS/TILT=O/SIZE=1.87»10/ZONE=ATTIC)
AFS(AREA=.096/ZS=8.)
% SCHEDULES
LIT(WS=LBW/CAP-=.6/RAD=.8/VIS=.l)
CTRL(CWS=WINTER)
3

ZONE C
$ BEDROOM *2
GE0M(NAME=BDRM2/Z=9.17/CLHT=8>

C-4
$ PARTITIONS - -> 82.8
SRF(C0NS=PTN-A/IN/0P/BS/AZM=90/SIZE=14.16»7.67)
SRF<C0NS=PTN-S/IN/0P/BS/AZM=90/SIZE=1 .74»7.67)
SRF(C0NS=PTN-A/IN/0P/BS/AZ«=180/SIZE--^4,8»7.67/Z0NE=MBDRM)
SRF(C0NS=PTN-S/IN/0P/BS/AZM=180/SIZE=.6»7.67/Z0NE=fiBDRM)
AFS(AREA=.019/ZS=4.
SRF(C0NS=PTN-A/IN/0P/BS/AZM=270/SIZE=2.9»7.67/Z0NE=FLR1)
SRF(C0NS=PTN-S/IN/0P/BS/AZM=270/SIZE=.35j7.67/Z0NE=FLR1)
AFS(AREA=.5/Z8=4.)
SRF(C0NS=PTN-A/IN/0P/BS/AZM=270/SIZE=9.6»7.67/Z0NE=BDRM3)
SRF(C0NS=PTN-S/IN/0P/BS/AZh=270/SIZE^1.2f7.A7/Z0NE=BDRM3)
AFS(AREA=.038/ZS=4.
$ NORTH WALL - 77.7
SRF(C0NS=WL-I/EX/0P/BS/AZM=0/SIZE^-8.65f8/0R6=19.72»34j0)
SRF<C0NS=WDW/EX/TR/SS/SIZE=5.0»4.17/0R6=2.35>3.5>-.17)
SRF(C0NS=WL-S/EX/0P/BS/AZM=0/SIZE=1.O6>8/ORG=11.O7»34 jO)
AFS(AREA=.022)
AFS(AREA=.020/ZS=8.)
$ FLOOR
SRF(CONS=FLR2-A/IN/OP/BS/TILT=180/SIZE=11.49»10/ZONE=FLR1)
SRF(C0NS=FLR2-T/IN/0P/BS/TILT=180/SIZE=1.41j10/Z0NE=FLR1)
AFS(AREA=.065)
i CEILING - 192.
SRF(C0NS=CLG2-I/IN/0P/BS/TILT=0/SIZE=11.9,10/Z0NE=ATTIC)
SRF(C0NS=CL62-T/IN/0P/BS/TILT=0/SIZE=.94»10/Z0NE=ATTIC)
AFS(AREA=.048/ZS=8.
% SCHEDULES
LIT(WS=LBW/CAP=.4/RAD=.8/VIS=.l)
CTRL<CWS=WINTER)
3
ZONE C

% BEDROOM *3
GE0M(NAME=BDRM3/Z=9.17/CLHT=^8)
i PARTITIONS - 80. 6» 82. 8» 92.8
SRF(C0NS=PTN-A/IN/0P/BS/AZM=180/SIZE^9.36»7.67/Z0NE=FLR1)
SRF<C0NS=PTN-S/IN/0P/BS/AZM=180/SIZE=1.15»7.67/Z0NE=FLR1)
AFS(AREA=.5/ZS^4)
SRF(C0NS=PTN-A/IN/0P/BS/AZM=90/SIZE=9.6»7.67/Z0NE=BDRM2)
SRF(C0NS=PTN-S/IN/0P/BS/AZM=90/SIZE=1.2»7.67/Z0NE=BDRM2)
AFS(AREA=.038/ZS=4)
SRF(C0NS=PTN-A/IN/0P/BS/AZM=270/SIZE=10.8j7.67)
SRF(C0NS=PTN-S/IN/0P/BS/AZM=270/SIZE=1.3>7.67)
% NORTH WALL - 77.7
SRF(C0NS=WL-I/EX/0P/BS/AZM=0/SIZE--8.63f8/0RG=10j34>0)
SRF(C0NS=WDW/EX/TR/SS/SIZE=5.0»4.17/0RG=2.35»3.5»-.17)
SRF(C0NS=WL-S/EX/0P/BS/AZM=0/SIZE-=1.06»8/0RG=1.35 j34j0)
AFS<AREA=.022)
AFS(AREA=.020/ZS=8.)
% FLOOR - 104.3
SRF(C0NS^FLR2-A/IN/0P/BS/TILT=180/SIZE=9.29j10/Z0NE=FLR1)
SRF(C0NS=FLR2-T/IN/0P/BS/TILT=180/SIZE=1.14>10/Z0NE=FLR1)
AFS(AREA=.080)

C-5
% CEILING - 104»3
SRF(C0NS=CLG2-I/IN/QP/BS/TILT^0/SIZE=9^67Jl0/Z0NE=ATTIC)
SRF(C0NS=CLG2-T/IN/0P/BS/TILT=0/SIZE=.76j10/Z0NE=ATTIC)
AFS(AREA=»039/ZS=8»
$ SCHEDULES
LIT(WS=LBW/CAP=,4/RAD=»8/UIS=»1)
CTRL(CWS=WINTER)
:
ZONE C

$ ATTIC
GE0M(NAME=ATTIC/Z=17.17/CLHT=2»2)
$ ATTIC FLOOR - 129»1j 256»4> 129»» 104*3
SRF(C0NS-^FLR3-I/IN/OP/BS/TILT = 18O/SIZE = ll*97f 10/Z0NE=FLR1)
SRF(C0NS=FLR3-T/IN/0P/BS/TILT=180/SIZE=,94»10/Z0NE=FLR3
AFS(AREA=*049)
SRF(CONS=FLR3-I/IN/OP/BS/TILT=180/SIZE=23*8>10/Z0NE=MBDRM)
SRF(C0NS^FLR3-T/IN/0P/BS/TILT=180/SIZE=1*87j10/Z0NE=MBDRM)
AFS(AREA=*096)
SRF(C0NS=FLR3-I/IN/0P/BS/TILT=180/SIZE=ll*9f 10/Z0NE=BDRM2)
SRF(C0NS:^FLR3-T/IN/OP/BS/TILT=180/SIZE=*94 j10/ZONE=BDRM2)
AFS(AREA=*048)
SRF(C0NS=FLR3-I/IN/0P/BS/TILT=180/SIZE=9*67»10/Z0NE=BDRM3)
SRF(C0NS=FLR3-T/IN/0P/BS/TILT=180/SIZE=*76j10/Z0NE=BDRM3)
AFS(AREA=*039)
SRF(CONS=RO0F-A/EX/OP/BS/AZM=O/TILT = 180/ORG=19*73 »34>0/SIZE=19*42»2)
AFS(AREA=2,/EXP=*5)
$ ROOF
SRF(C0NS=R00F-A/EX/0P/BS/AZM=180/TILT-22»6/SIZE=20f 18*8/0RG=0f0F0)
AFS(AREA=*35)
SRF(CONS=ROOF-A/EX/OP/BS/AZH=0/TILT=22*6/SIZE=20jl8,8/ORG=20»34fO)
AFS(AREA=*35)
$ NO SCHEDULES
:

C-6
 21 JAN SMALL INTER-ROOM OPENINGS

6.7 m/s
-14.4°C

I
1.00 ach
I

I
7.34 kWh
8.2 sec/day
ii Fh

STANDARD OPENINGS

I. 02 ach
7.41 kWh
II. 1 sec/day

SINGLE ZONE

1.03 ach

7.42 kWh
4.6 sec/day

CASE 7d
AIR CHANGE METHOD; using same ach
gives same load
3.9 sec/day
NBS-114A (REV. 2-80
U.S. DEPT. OF COMM. 1, PUBLICATION OR 2. Performing Organ. Report No. 3. Publication Date
REPORT NO.
BIBLIOGRAPHIC DATA NBSIR 81-2404 December 1981
SHEET (See instructions)
4. TITLE AND SUBTITLE

CALCULATION OF INTER-ROOM AIR MOVEMENT FOR MULTI-ROOM BUILDING ENERGY ANALYSIS

5. AUTHOR(S)
George N. Walton 742.01
6. PERFORMING ORGANIZATION (if joint or other than NBS, see instructions) 7. Contract/Grant No.

NATIONAL BUREAU OF STANDARDS

DEPARTMENT OF COMMERCE 8 . Type of Report & Period Covered
WASHINGTON, D.C. 20234

9. SPONSORING ORGANIZATION NAME AND COMPLETE ADDRESS (Street. City, State, ZiP)
DEPARTMENT OF ENERGY
NBS and Passive and Hybrid Solar Energy Division
Office of Solar Heat Technologies
Washington, DC 20585
10. SUPPLEMENTARY NOTES

I I
Document describes a computer program; SF-185, FIPS Software Summary, is attached.
11. ABSTRACT (A 200-word or less factual summary of most significant information. If document includes a significant
bi bliography or literature survey, mention it here)

A model is presented for computing the infiltration and air flow between rooms of a
multi-room building in terms of basic principles of fluid mechanics. This model
has been incorporated into a comprehensive loads-predicting computer program.
Air flows, room temperatures, and heating loads for a typical townhouse under
different conditions of environment and with various construction features are
computed. These calculations show the feasibility of detailed multi-room air
movement analysis. They also indicate that when the inter-room openings of a low-
rise structure are large compared to the envelope openings, the infiltration and
total load can be accurately, and more quickly, computed by assuming no resistance
to air flow between rooms. This property will also allow simplified calculations
for high-rise buildings with many rooms. Methods are proposed for handling more
complex air flow phenomena.

12. KEY WORDS (Six to twelve entries; alphabetical order; capitalize only proper names; and separate key words by semicolon s)
building energy analysis; computer simulation; infiltration; natural ventilation

13. AVAILABILITY 14. NO.OF

PRINTED PAGES
i
Unlimited
37
I I
For Official Distribution. Do Not Release to NTIS
I I
Order From Superintendent of Documents, U.S. Government Printing Office, Washington, D.C.
20402. 15. Price

$6.50
Order From National Technical Information Service (NTIS), Springfield, VA. 22161

USCOMM-DC 6043-P80
1