PRESCRIBING THE INPUT FOR THE ASET VERSUS RSET ANALYSIS:

IS THIS THE WAY FORWARD FOR PERFORMANCE BASED DESIGN?

C M Fleischmann, PhD, PE
University of Canterbury
Christchurch New Zealand

ABSTRACT
The popularity of Performance Based Design (PBD) has continued to increase over the last two decades
and many consider PBD provides for cost effective and innovative solutions to fire safety challenges.
Fundamental to PBD for life safety, is the principle that the occupants have enough time to exit the
building before being overcome by the fire. In fire engineering terms the Available Safe Egress Time
(ASET) must exceeding the Required Safe Egress Time (RSET) with an appropriate margin of safety.
Currently the necessary input and acceptance criteria are left up to the fire engineer with the approval
from the authority having jurisdiction. Unfortunately the conventional guidance is more qualitative rather
than quantitative which can lead to varying levels of safety in buildings depending on the values chosen
for use in the analysis. This paper describes the necessary input parameters and the appropriate
acceptance criteria for ASET versus RSET analysis and discusses some of the available guidance for
determining these values. The paper ends with a brief description of the framework that the Department
of Building and Housing is proposing for PBD in New Zealand. The framework outlines the design fire
scenarios, design fires, premovement times and acceptance criteria that is currently being field tested.

INTRODUCTION
For nearly two decades Performance Based Design (PBD) has been touted as the future of building design
for fire safety providing for cost effective and innovative solutions to fire safety challenges. Although
PBD continues to grow in popularity and sophistication, Fire Engineering has yet to reach the same level
of understanding compared with the more traditional disciplines where PBD is common place. Fire
engineering is still a rapidly developing discipline with new methodologies and understanding evolving
continuously. For example, it has only been the last five years that CFD modeling has become common
practice for complex fire engineering analysis, where a decade ago only universities and research
institutions had the necessary computing power. Since 1996 the Society of Fire Protection Engineers has
held a biennial international conference on performance based codes and design methods to highlight the
latest developments in performance based fire safety research and design.
At the very heart of PBD for life safety is the fundamental principle that the occupants have enough time
to exit the building before being overcome by the fire. In fire engineering terms, the Available Safe
Egress Time (ASET) must exceed the Required Safe Egress Time (RSET) with an appropriate margin of
safety. There are a number of books, guides, and codes on PBD and many countries allow for
performance based solutions to design problems. One of the most comprehensive codes that include a
performance based option is the National Fire Protection Association Life Safety Code1 (NFPA101).
NFPA101 lays out 8 scenarios that must be used to evaluate a proposed building design. However the
scenarios and supporting performance clauses of the code are very qualitative in nature and do not
provide quantitative advice about the design fire, acceptance criteria, or methodology but simply outline
all of the factors that should be considered by a designer without actually quantifying any of the necessary
input parameters or acceptance criteria. This leaves the designer having to develop their own criteria and
design input with the approval of the Authority Having Jurisdiction (AHJ). This lack of quantified
1

guidance forces the FPE to turn to the literature and pull together the required input and performance
criteria from a number of sources to carry out their analysis.
Under the current approach, without quantified guidance, there is significant variability in the design fire
scenarios, design fires and performance criteria. For example in one building the designer evaluates the
Fractional Equivalent Dose (FED) at 2 m above the floor yet in another design the Fire Engineer
calculates the FED at 1.8 m. In many cases the local AHJ is reluctant to challenge the Fire Engineers
recommendations for the design fire and performance criteria because the AHJ has a lower qualification
than the Fire Engineer. This can lead to inconsistent levels of fire safety in buildings of similar
occupancy.
This paper outlines the input required to carry out the typical ASET versus RSET analysis. A brief
review of the literature is presented for the primary input required for the ASET versus RSET analysis
including fire scenarios, design fires, egress parameters and acceptance criteria. The paper will then
summarize the ongoing work in New Zealand in which the Department of Building and Housing is
developing a framework that will specify the required input for PBD. The framework has been under
development for 2 years and is currently being field tested by fire engineering practitioners. Prior to
releasing the framework for the trial, an internal evaluation was carried out applying the framework to a
number of building designs that complied with the New Zealand compliance documents2 that are
deemed to satisfy the performance based code. The results of the internal trial will also be discussed.

FIRE SCENARIOS
Fundamental to any fire safety evaluation process are the design fire scenarios. In the context of this
paper, a fire scenario is a qualitative description that characterizes the key events of a potential fire. A
design fire scenario is a description of a specific fire scenario that can be used in a fire safety engineering
analysis. Typically the design fire scenario as used in deterministic analysis may simply dictate particular
performance requirements such as the allowable surface spread of flame in exitways. There are an
infinite number of potential fire scenarios and it is common for the fire engineer to reduce the fire
scenarios to a manageable amount and use deterministic methods to evaluate the consequences of the
scenario in the proposed building against the performance criteria.
There are a number of references which discuss the various aspects of choosing fire scenarios3, 4. The
International Standards Organization technical committee 92 developed ISO/TS16733 Fire safety
engineering Selection of design fire scenarios and design fires5 which outlines a 10 step comprehensive
procedure which includes an event tree to help reduce the number of design scenarios to a manageable
level. The 10 steps are:
1. Location of fire Select fire locations that produce the most challenging fire scenarios.
2. Type of fire Identify the most likely types for fire scenarios and most likely high consequence
fire scenarios based on fire incident statistics.
3. Potential fire hazards identify other critical high consequence scenarios for consideration.
4. Systems impacting on fire - Identify the building and fire safety systems that are likely to have a
significant impact on the fire or development of untenable conditions.
5. Occupant response Identify occupant characteristics and response features that are likely to
have a significant impact on the course of the fire scenarios.
6. Event tree Develop event tree that represents the possible factors that have been identified as
significant.
7. Consider probability Estimate the probability of occurrence of each state using the available
reliability data and engineering judgment when data is not available.
2

8. Consideration of consequence Estimate the consequences of each scenario using engineering

judgment.
9. Risk ranking Rank the scenarios in order of relative risk.
10. Final selection and documentation Select the highest-ranked fire scenario for quantitative
analysis.
The National Fire Protection Association (NFPA) has taken a different approach to developing fire
scenarios for use as a performance based option in their Life Safety Code1 and Building Construction
and Safety Code6. In each of these codes the NFPA has identified 8 scenarios that must be analyzed and
compared to the performance criteria. The 8 required scenarios include:
1. Occupancy specific fire representative of a typical fire for the occupancy.
2. An ultra-fast developing fire in the primary means of egress, with interior doors open at the start
of the fire.
3. A fire that starts in a normally unoccupied room, potentially endangering a large number of
occupants in a large room or other area.
4. A fire that originates in a concealed wall or ceiling space adjacent to a large occupied room.
5. A slowly developing fire, shielded fire protection systems, in close proximity to a high occupancy
area.
6. Most severe fire resulting from the largest possible fuel load characteristics of the normal
operation of the building.
7. Outside exposure fire
8. Fire originating in ordinary combustibles in a room or area with each passive and active fire
protection system independently rendered ineffective.
Additional scenarios may also be specified by the design team or authority having jurisdiction.
According to the NFPA, additional scenarios should be considered and suggest that as a minimum the
following three types of scenarios be considered:
1. High-frequency, low-consequence scenarios
2. Low-frequency, high-consequence scenarios
3. Special problems scenarios
The additional scenarios are intended to take into account the unique characteristics of the building.
Although the NFPA and ISO/TC92 documents give very detailed discussion for developing the fire
scenarios, they do not specify the design fires that are required to carry out the fire safety evaluations for a
building. Thus it is left to the fire engineer to come up with the design fire required.

DESIGN FIRE CHARACTERISTICS

Each of the design fire scenarios are qualitative in nature and require a quantitative design fire for use in a
fire safety assessment. A design fire is intended to represent a credible worse case scenario that will
challenge the fire protection features of the building. Although simple in concept this definition can be
hard to interpret when attempting to quantify the design fire especially in low ceiling spaces where
occupants are expected to be sleeping. Typically the design fire is described in terms of the heat release
rate from the fire. Indeed, the heat release rate history is considered the single most important variable in
describing a fire hazard.7 However, the design fire may also include an estimate of the size of the fire, the
species being produced, and the smoke production rate. Unfortunately it is not possible to derive the
design fire from first principles and can be quite difficult to quantify in practice. The detail required for a
3

design fire is dependent on the issue being investigated and what questions the engineer is trying to
answer. For example it is not much use to have a design fire that includes the decay phase if the engineer
is trying to predict the activation of a sprinkler head. Likewise, the growth phase makes little difference if
the engineer is trying to model the fire resistance of a structural member after four hours of fire exposure.
Thus the nature of the design fire depends on the issues the fire engineer is resolving. Figure 1shows the
idealized fire growth rate highlighting the four phases of conventional fire development and the transition
of flashover.

Flashover
Growth

Fully Developed

Decay

Heat Release Rate

Incipent

Time

Figure 1 Idealized heat release rate history highlighting the 4 phases of conventional fire development
and flashover.

The Incipient Phase of a fire can last from a few seconds to days depending on the initial fuel involved,
ambient conditions, ignition source, etc. In the case of a flammable liquid spill the incipient phase is
effectively nonexistent. If it is a self- heating to ignition, the incipient phase can last for hours if not days.
In some cases the fire may not grow beyond the incipient phase, consider a cigarette which smolders on a
wool fabric covered chair may never ignite the flammable padding beneath the fabric. There are far too
many variables to allow for reliable modeling of the incipient phase of a fire. Indeed, for the furniture
calorimeter test a gas burner is used simulate a wastepaper basket to eliminate the impact of incipient
phase on the early growth phase.
The Growth Phase is considered to begin when the radiation feedback from the flame governs the
burning rate. Assuming the compartment is vented, the growth rate is primarily governed by the fuel
properties and orientation. During the growth phase the fire spread across the fuels surfaces, increasing
the burning area and corresponding heat release rate. The heat release rate is assumed to be independent
of the fire enclosure and governed more by the flame spread rate. Compartment enhancement due to the
accumulation of hot gases is considered small until the fire nears flashover.
Modeling the actual growth rate is extremely difficult and remains an area of active research. It is
dependent on many factors which are not only a function of the burning object but are also stochastic in
nature such as size and location of the ignition source, orientation of the object, proximity to other object,
proximity to boundaries, proximity to openings, etc. Not withstanding these limitations, the engineer
must rely on judgment when choosing a growth rate. It is true that most fires occurring during the life of
a building will be quite minor and are likely to go unreported; it is the reasonable worst case fire and not
the most likely fire that must be used for design.

There are several approaches to estimating the growth rate for a particular design fire. The most popular
is the t-squared fire growth rate. Originally developed in the 1970's for predicting fire detector activation,
the t-squared fire gained popularity when it was included in the appendix of NFPA728. In NFPA72 there
are three categories for fire growth slow, medium, and fast. These definitions are simply determined by
the time required for the fire to reach 1055 kW (1000 BTU/s). A slow fire is defined as taking 600 or
more seconds to 1055 kW. A medium fire takes more than 150 seconds and less than 300 seconds and
fast fire takes less than 150 seconds to reach 1055 kW. Over time the definition for t-squared fire has
evolved to include an ultra fast fire as well. The common definition for the growth times are shown
below:

q t 2

(1)

where:

q - heat release rate (kW)

- growth constant (kW/s2)
t - time from effective ignition (s)
Classification
Slow
Medium
Fast
Ultra

Growth time
(Time to 1055kW)
(s)
600
300
150
75

(kW/s2)
0.00293
0.0117
0.0469
0.188

The t squared fire growth can be thought of in terms of a burning object with a constant heat release rate
per unit area in which the fire is spreading in a circular pattern at a constant flame speed. Obviously more
representative fuel geometries may or may not produce a t-squared fire growth. However, the implicit
assumption in many cases is that the t-squared approximation is close enough to make reasonable design
decisions9. It should be noted that the t-squared growth rate has been adopted well beyond the original
intent in some cases for fires as large as 30 MW. Such application has been questioned in the literature.10
Flashover occurs when the radiation from the upper layer is so intense that all of the combustible
surfaces in the compartment ignite. Flashover can be thought of as a transition from a small object
oriented fire to full room involvement. This transition typically occurs over a short time span measured
in seconds. Figure 2 is a plot of the heat release rate and upper layer temperature versus time for an ISO
9705 scale compartment with wood cribs and Medium Density Fiberboard wall linings. Flashover occurs
in the cross hatched region of the curve. From an experimental point of view flashover is considered to
occur when the upper layer temperature reaches 500-600C, as seen in Figure 2. The increase in radiation
from the upper layer not only ignites all of the combustibles in the room but also enhances the heat
release rate of all the burning objects. From a design point of view, flashover should be modeled as a
linear transition from a growing fire to a fully developed fire over a very short period of time.

Flashover

4500

900

4000

800

3500

700

3000

600

2500

500

2000

400

1500

300

1000

200

500

100

Temperature (C)

Heat Release Rate (kW)

Fire Extinguished

0
0

60

120

180

240

300

360

420

480

540

Time(s)
Heat Release Rate

Compartment Temperature

Figure 2 - Heat Release Rate History and upper layer temperature from an ISO sized compartment fire
showing the rapid increase in the heat release rate as the upper layer temperature rises above 500 to
600C.

In the Post Flashover/Fully Developed phase of the fire all of the combustible objects in the
compartment are burning and the heat release rate is either limited by the fuel surface area or the available
air supply. Typically it is the available air supply that governs the post flashover phase except in the
cases of very large openings or low combustible surface areas. The mass of air that flows into an opening
can be estimated using the well know A square root H correlation first identified by Kawagoe when
reducing post flashover fire data in 195811.

air 0.5A O H O
m

(2)

where:

air - mass flow rate of air into the compartment (kg/s)

m
AO
HO

- area of the opening (m2)

- height of the opening (m)

The heat release rate within the compartment can then be estimated using the assumption that most fuels
release a constant amount of energy per unit mass of air consumed, that is 3.0 MJ/kgair. Thus turning
Equation 2 into a heat release rate equation:

q inside 1.5A O H O

(3)

where:

q inside - heat release rate inside the compartment (MW)

It should be reiterated that this is the energy that is released inside the compartment. In many cases the
burning objects actually release more fuel than can be consumed within the compartment, i.e. the fire is
ventilation limited which can fuel very long flames out of the opening12.

The Decay Phase occurs when the fire has consumed much of the available fuel and the heat release rate
starts to diminish. During the Decay phase the fire will typically transition from ventilation controlled to
surface area controlled. This is primarily of interest when determining the required fire resistance of
structural elements. This phase of the design fire curve is the least studied and least understood. In most
cases fire fighting intervention prevents or at least interferes with the fires decay.

PERFORMANCE CRITERIA
The performance criteria can be as challenging as the design fire scenarios and design fires themselves.
The appropriate performance criteria are dependent on the particular fire scenario and the portion of the
design being evaluated. For building to building fire spread the performance criteria could be an
allowable radiative heat flux or surface temperature on the adjacent building or boundary, for structural
performance it could be a prescribed time in a specific standard fire test, for surface finish it could be a
performance in a standard flame spread test. Quantifying the performance is much more challenging
when predicting the impact of the fire on the occupants. The fires impact on life safety is commonly
broken down into four categories; thermal effects, narcotic gas effects, irritant gas effect and visibility.
The most comprehensive review on the hazard to occupants from the fire gases is given by Purser in The
SFPE Handook of Fire Protection Engineering13. In this section, Purser gives a compendium of the
available literature on the hazards that smoke poses to humans and provides the engineering tools
necessary to allow the designer to estimate the hazard that the smoke may have on egressing occupants.
The assessment is usually in the form of the Fractional Effective Dose (FED) which is defined as the ratio
of the exposure dose to the exposure dose necessary to produce incapacitation. The FED can be defined
for asphyxiant toxicants, irritant gases, or radiative and convective heat exposure. For information on
calculating the hazard for occupants posed by the smoke and heat the reader is directed to references 13,
14 &15.
Ultimately the performance criteria must be selected for life safety. Although an FED of 1 is considered
to be the point at which a person might be expected to be incapacitated, it is considered prudent, for two
primary reasons, to use a value less than one for conservatism. Firstly, the uncertainty in calculations is
high because of the limited amount of data available for comparison. The data used to develop the
relationships are based on both human and animal research. To further refine the results, additional
experiments would be necessary but exposing humans to dangerous toxic species is considered unethical
and is not expected to ever be available. The second reason is that the data used was for young healthy
adult humans and animals which represent the least vulnerable population. Certain subpopulations such
as elderly and the young are expected to be more vulnerable to the effects of fire and must be considered
in design. Thus documents such as Published Document 7947-6:200416 recommend the use of the
FED<0.3 as the acceptance criteria and visibility of 10 m. In cases where the occupants are considered to
be a vulnerable subpopulation the FED may be set even lower.

REQUIRED SAFE EGRESS TIME (RSET)

The Required Safe Egress Time (RSET) can be defined as the time from ignition of a fire until the time
when the occupants reach a place of safety16. The RSET is a function of four time based values:
tdet - detection time: time from ignition to detection by an automatic system or time for
occupants to detect the fires cues.
talarm alarm time: time from detection to a general alarm.
tpre pre-movement time: time from alarm to time when occupants start to egress the building.
This time includes two components: time for the occupant to recognize the alarm and time for the
occupant to respond to the alarm and start evacuation.
7

ttravel travel time: time it takes for the occupants to travel from their location in the building to
a safe place. This commonly comprises two parts, the walking time and the flow time. The
walking time is based on the speed that the occupants are expected to walk when egressing. The
flow time is the time it takes for the occupants to flow through the exit which includes flows
through a doorway or down stairs. This can also include the time an occupant is in a queue
waiting to evacuate a space.

Premovement Times
For a detailed description on how to determine the values listed above the reader is directed to references
15-17. In an RSET analysis the detection time is calculated using a deterministic model to estimate the
time a detection device will activate. Originally this was carried out using the program DETACT which
estimated the detection time based on the ceiling jet temperature and velocity and the Response Time
Index (RTI) of the detection device. However as our understanding of detection theory has improved so
have the models for predicting the detection time. The reader should consult reference 18 for more detail
on detection theory. The alarm time and premovement times should be agreed upon as part of the Fire
Engineering Brief (FEB) process before calculating the RSET. Proulx has carried out a number of studies
quantifying the evacuation times from both trial evacuations and actual fires that are summarized in
reference19. Unfortunately researchers in the area of human behavior are reluctant to suggest numbers
for the premovement times due to the limited research in this area. However, the PD7974-6:200416 does
address the premovement times for occupants and gives guidance for estimating the premovement times.
The suggested times are based on: occupancy classification, alert status of the occupants (awake or
asleep), familiarity with the building, level of management, and type of alarm signal.
Table 1shows the recommended values for the premovement time suggested in PD7974-6:2004. The
following descriptions help to explain the codes in the first column of
Table 1:
Management Level
M1- occupants (staff and residents) should be trained to a high level of fire safety management
M2- similar to M1 but lower staff ratio and floor wardens not always present
M3-basic management with minimum fire safety management
Alarm Level
A1-automatic detection throughout the building activating an immediate general alarm to all
occupants
A2-automatic detection throughout the building providing a prealarm to management or security
with a manually activated general alarm
A3-local automatic detection and alarm only near location of the fire or no automatic detection
with manually activated general alarm
Building complexity
B1- simple rectangular single story building with one or few enclosures and simple layout
B2-simple multi-enclosure (usually multi-story) building and simple internal layout
B3-large complex building internal layout and enclosures involve often large and complex spaces
such that occupants may have wayfinding difficulties.
Columns 2 and 3 in Table 1 give the time from alarm to the movement of the first few occupants and the
distribution times for the populations of occupants to start their evacuation. For additional details
regarding the values given in Table 1 the reader should consult PD7974-6:200416. The premovement
times shown in Table 1 demonstrate the wide range of values that might be expected in a building.
Clearly the biggest influence is the level of management within the building. For example, in Table 1, for
8

office buildings the values range from 0.5 to >15 minutes for first occupants to start moving based on the
quality of the management. The alarm type dependence is less significant than management but is a
major factor when only a manual alarm is available. For complex building a fixed amount of time is
added to the premovement times ranging from 0.2 to 1.0 minutes.
First occupants
tpre (1st percentile)
(Minutes)

Occupant distribution
tpre (99th percentile) a
(Minutes)

0.5
1
>15

1.0
2
>15

0.5
1.0
>15

2
3
>15

5
10

5
>20

Cii: Managed Occupancy (e.g. serviced apartments, halls of residence)

M1 B2 A1 A2
M2 B2 A1 A2
M3 B2 A1 A3

10
15
>20

20
25
>20

Ciii: Sleeping and Unfamiliar (e.g. hotel, boarding house)

M1 B2 A1 A2
M2 B2 A1 A2
M3 B2 A1 A3
For B3 add 1.0 for wayfinding
M1 would normally require voice alarm/PA
D: Medical Care.
Awake and Unfamiliar (e.g. day centre, clinic, surgery, dentist)
M1 B1 A1 A2
M2 B1 A1 A2
M3 B1 A1 A3
For B2 add 1.0 for wayfinding. For B3 add 1.0 for wayfinding
M1 would normally require voice alarm/PA

15
20
>20

15
20
>20

0.5
1.0
>15

2
3
>15

SCENARIO CATEGORY AND MODIFIER

A: Awake and Familiar (office or industrial)
M1 B1 B2 A1 A2
M2 B1 B2 A1 A2
M3 B1 B2 A1 A3
For B3, add 0.5 for wayfinding
M1 would normally require voice alarm/PA if unfamiliar visitors likely to be present
B: Aawake and Unfamiliar (shop, restaurant, circulation space)
M1 B1 A1 A2
M2 B1 A1 A2
M3 B1 A1 A3
For B2 (Cinema, theater) add 0.2 for wayfinding. For B3 add 1.0 for wayfinding
M1 would normally require voice alarm/PA
Asleep
Ci: Sleeping and Familiar (e.g. dwellings individual occupancy)
M2 B1 A1
M3 B1 A3
For other units in block assume one hour

Sleeping and Unfamiliar (e.g. hospital ward, nursing home, old peoples home)
M1 B2 A1 A2
M2 B2 A1 A2
10 b
5b
10 b
20 b
M3 B2 A1 A3
b
>10
>20 b
For B3 add 1.0 for wayfinding
M1 would normally require voice alarm/PA
E: Transportation. Awake and Unfamiliar (e.g. railway, bus station or airport)
M1 B3 A1 A2
1.5
4
M2 B3 A1 A2
2.0
5
M3 B3 A1 A3
>15
>15
M1 and M2 would normally require voice alarm/PA
a. Total pre-movement time = tpre (1st percentile) + tpre (99th percentile). Figures with greater levels of uncertainty are italicized.
b. These times depend upon the presence of sufficient staff to assist evacuation of handicapped occupants.

Table 1 Suggested premovement times from PD7974-6:2004.

A CASE STUDY IN DEFINING THE INPUT FOR ASET VERSUS RSET ANALYSIS
Since August 2006 the New Zealand Department of Building and Housing (DBH) has been developing a
new methodology to demonstrate compliance with the Fire Safety requirements of the New Zealand
Building Code (NZBC) specifically the C clauses. This work was identified as necessary after a
comprehensive review of the existing building code. One of the key outcomes of the review was that the
public feels that the existing code provides an acceptable level of safety. The New Zealand Building
Code will maintain its performance basis for fire safety but inputs for performance-based designs will be
predetermined. This approach still permits flexibility and innovation in design, but ensures consistency
between designs for very similar uses. This provides a mechanism for the regulator to exercise control
over the level of fire safety that must be achieved in buildings, without having to go through a formal
process to calculate expected fire losses on a building-by-building basis. These inputs are analogous to
wind, earthquake, snow loads etc given in a loadings code for structural design. At the time of this paper
(October 2009) the framework is being field tested with a number of practicing engineers that will
conclude in December 2009. The design fire scenarios, design fires, premovement times, and acceptance
criteria are briefly discussed below as an example of where this author believes future of PBD should
lead.

Design Fire Scenarios

Ten fire scenarios are proposed for the use in the new framework loosely based on those in NFPA 50006,
with some modification and in one case (fire spread to neighbouring property and fire service operations)
has been expanded. Other scenarios for external vertical fire spread and interior surface finishes have also
been added to specifically address fire scenarios currently dealt with in the existing prescriptive
compliance documents referred to as C/AS12. All ten scenarios are presented here for completeness
although only four of the scenarios are applicable for ASET versus RESET analysis.
Table 2 summarizes the design fire scenarios being proposed in the framework. Column 1 gives the
scenario number for ease of identification, Column 2 gives a description of the scenario, Column 3
describes the performance objectives for each scenario, Column 4 defines the design event that must be
used in the analysis, and Column 5 describes the methodology expected to demonstrate that the scenario
has been addressed. Details about the design fires and performance criteria are described in the following
sections.

10

#
1

Description
These fires are intended to
represent a credible worse case
scenario that will challenge the
fire protection features of the
building.

Performance Objective
Provide a tenable environment for
occupants in the event of fire while they
egress to a safe place.

Fire is located near the primary

escape route or exit that
prevents occupants from
leaving the building by that
route. Fire originating within an
exitway may be the result of a
deliberately lit fire or be
accidental. Fire originating
within an escape route in the
open path will be considered to
be a severe fire applicable to
the particular building use as
described in Scenario 1.
A fire starting in an unoccupied
space may grow to a significant
size undetected and spread to
other areas where the greatest
number of occupants.

Provide a viable escape route from the

building for occupants in the event of
fire, i.e. provide at least 2 exits of equal
size

A fire that starts in a concealed

space could develop
undetected and spread to
endanger a large number of
occupants in another room.

Design Event
Design fires are characterized
with t-squared rate of heat
release, peak rate of heat release,
and fire load energy density
(FLED). Design values for yields
are specified for CO, CO2 and
soot/smoke.
The design fires are intended to
represent free-burning fires but
they may be modified during an
analysis to account for building
ventilation and fire suppression
effects on the fire.
Fire blocking exit in open or safe
path.
Fire characteristics dont matter
since fire is assumed to physically
block the exit.

Expected Methodology
Calculations of the fire environment in
the escape routes that will be evaluated
using the tenability criteria.

Maintain tenable conditions on escape

routes until the occupants have
evacuated.
Protect against fire spread that could
compromise the retreat of fire-fighters.

Use fire characteristics from

scenario 1 for the applicable
occupancy.

Include fire separations or fire

suppression to confine the fire to room
of origin
Include automatic detection to provide
early warning of the fire in the
unoccupied space
Carry out tenability analysis of escape
routes if fire is able to spread into the
occupied space

Maintain tenable conditions on escape

routes until the occupants have
evacuated.
Protect against fire spread that could
compromise the retreat of firefighters.

Currently unable to identify a

suitable quantitative description of
the design event, and would
expect that traditional solutions
would apply i.e containment,
detection or suppression.

Fire separations or suppression to

confine fire to concealed space
Automatic detection to provide early
warning
Tenability analysis with fire spreading
into the occupied space

11

Provide alternative escape routes that

are tenable or design single escape
routes so that no more than 50 people
are served (for open paths) or 150
people (vertical safe paths). Analysis
not required.
This scenario applies to individual
rooms in the open path, and to corridors
and stairs that are part of an exitway.
Escape routes serving less than 50
persons will be permitted to have a
single exit.

#
5

Description
A slow smouldering fire that
causes a threat to sleeping
occupants.
A large fire within a building
may spread to neighbouring
buildings as a result of heat
transfer (predominantly by
radiation through openings in
external walls). To reduce the
probability of fire spread
between neighbouring
properties, measures to limit
the radiation flux received by
the neighbouring building are
required.

A fire source adjacent to an

external wall such as a fire
plume emerging from a window
opening, or a fire source in
close contact with the faade
that could ignite and spread fire
vertically.
There are two parts to this
scenario:
1. External vertical fire spread
via the faade materials
2. Window fire plumes
spreading fire vertically
through higher openings
A flaming fire source located in
a wall-corner junction that
ignites room surface lining
materials and which then
subsequently leads to
untenable conditions on an
escape route.

Performance Objective
Maintain tenable conditions on
escape routes until the occupants
have evacuated.
1. External walls shall be
designed to limit the radiation
received on the neighbouring
property to:
a. no more than 30 kW/m on the
relevant boundary; and
b. no more than 16 kW/m at 1m
beyond the relevant boundary.
2. External walls of buildings, if
located 1m or closer to a relevant
boundary, and when subjected to
a radiant flux of 30 kW/m shall:
not ignite in 30 min. (PG III, IV)
not ignite in 15 min. (PG I, II)
Prevent fire spread to other
property and spaces where
people sleep (in the same
building) and maintain tenable
conditions on escape routes
until the occupants have
evacuated.
Protect against external
vertical fire spread that could
compromise the safety of firefighters working in or around
the building.

Tenable conditions on escape

routes shall be maintained
while occupants evacuate.
Protect against rapid fire
spread that could compromise
the retreat of firefighters.

Design Event
Refer to fire characteristics for a
smouldering fire.
Emitted Radiation flux from
unprotected areas in external
walls (assuming no intervention)
shall be taken as:
88 kW/m for FHC = 1
108 kW/m for FHC = 2
152 kW/m for FHC = 3 or 4

For 7A
Radiant flux of 50 kW/m
impinging on the faade for 15
minutes (for PG II and PG III)
Radiant flux of 90 kW/m
impinging on the faade for 15
minutes (for PGIV)
For 7B
Window plume projecting from
opening in external wall, with
characteristics determined
from design fire for Scenario 1.
Fire source of output 100 kW in
contact with a wall-corner element
for 10 minutes followed by 300
kW for 10 minutes in accordance
with ISO 9705.

12

Expected Methodology
Provide automatic smoke detection in
sleeping rooms and no further analysis is
required.

C/AS1 tabulated data for boundary

distances are acceptable,

Unprotected areas can be calculated

using the given emitted and received
radiation levels, boundaries distances and
configuration factors.

Fire tests of external cladding systems

using the cone calorimeter apparatus (ISO
5660) or similar are needed to demonstrate
that performance measure 2 above is met.

1. Follow existing C/AS1 and use:

a. Large or medium-scale faade type fire
tests (eg NFPA 285, ISO 13785, VCT)
b. Small-scale testing using ISO 5660 or
AS/NZS 3837 (cone calorimeter) for
homogeneous materials.
2. Use non-combustible materials.
3. Validated flame spread models could be
used for some materials.
4. Construction features such as aprons
and/or spandrels or sprinklers could be
used to meet performance measure 3
above. Window plume
characteristics/geometry may be derived
from Scenario 1 design fires.
1. ISO 9705 room corner fire
2. ISO 5660 cone calorimeter test at 50
kW/m (e.g. correlated to a full-scale
result)
3. Use non-combustible materials to AS
1530.1.
4. Use calculations from validated flame
spread models (if available for the
material and configuration of interest)

#
9

10

Description
Mitigation of risk on the
fireground on the part of the
officer requires the ability to
predict both fire and building
behaviour. What compromises
this ability is the occurrence of
events that are sudden,
unexpected or disproportionate
to the change that caused
them. It is the broad
predictability of the building
behaviour and the fire
environment that is
encapsulated in the concept of
reasonable expectations of
firefighters to be safe.

Performance Objective
In order that the officer in charge
may make a risk-informed
judgement about how to tackle
firefighting and rescue operations

The robustness of the design

will be tested by considering the
design fire with each key fire
safety system rendered
ineffective in turn.

Provide a tenable environment for

occupants in the event of fire
while they escape to a safe place.

Information must be available

to the crew on arrival to enable
them to rapidly size-up the
situation
Access to all floors of the
building must provide
firefighter protection
i. firefighting water must be
available in the vicinity of the
fire

Design Event
Firefighter tenability must be
established for large (>1500m2)
FHC 4 buildings, where fire
growth rate is very rapid, or for
unsprinklered building layouts
where the distance from the safe
path access to any point on a
floor exceeds 75m. The
firefighting design fire is 50MW,
unless the fire is sprinkler,
ventilation or fuel limited at some
lower value by the time the fire
service arrives.

Expected Methodology
1. Features that facilitate rapid sizeup of the situation

Design event is the same as

scenario 1 above.

Calculations of the fire environment in

the escape routes that will be evaluated
with one of the key fire safety systems
rendered ineffective. Only the FED
narcotic criterion is to be met.

Table 2 Design Fire Scenarios for the Conceptual Framework being field tested in New Zealand.

13

Hazardous substance signage

Fire detection system
Panel location and information
Firefighter control of building fire
safety systems
Limitation of fire size by sprinklers
or firecell size
2. Features that facilitate safe access
for rescue and firefighting
Firefighter access around building
Sprinklers in buildings higher than
fire service ladder appliances
Access through tall buildings
Protected from structural collapse
3. Features that facilitate adequate
firefighting water
External hydrants plus fire
appliance access to building
Internal risers, hydrants and
hosereels
Sprinklers

Design Fires
Quantifying the design fire is one of the most challenging requirements for PBD. Resolving the issue of
defining the design fire has resulted in some reflection on the existing compliance documents which have
been considered to provide a societal accepted level of safety. Indeed if the design fires required for use in
PBD are significantly more severe than the inherent fires within the compliance documents2, than there is a
disincentive for PBD that would suppress innovation in building design. Thus choosing an appropriately
rigorous design fire to provide an acceptable level of safety without being too onerous to stifle PBD required
a great deal of effort. Ultimately the following design fire was chosen (the few exceptional cases are
discussed below):
For all buildings except for the buildings explicitly discussed below, the fire is assumed to grow as a
fast t2 fire up to flashover and is then limited by the available ventilation assuming all windows are
broken out.
For sprinklered buildings the fire is assumed to be controlled, i.e. constant heat release rate, after the
sprinkler activates based on RTI and activation temperature.
Species yield for soot (Ysoot) is equal to 0.07 kg/kgfuel.
Species yield for carbon monoxide (YCO) is equal to 0.04 kg/kgfuel.
Net Heat of Combustion (HC) 20 MJ/kg
Radiative fraction from fire 0.35
Exceptions to the fast t2 fire
Building use

Fire Growth
rate ( q )
0.0117t2

Carparks
Rack Storage Group 1(Polystyrene chip in single wall
cardboard cartons)

0.0088t3 H

Rack Storage Group 2 (FMRC Standard Plastic commodity,

upholstery cushions)

0.0025 t3 H

Rack Storage Group 3 (FMRC Class II Double triwall

cardboard cartons)

0.00068t3 H

Species
Ysoot=0.07
YCO=0.04
HC= 20 MJ/kg
Ysoot=0.07
YCO=0.04
HC= 20 MJ/kg
Ysoot=0.07
YCO=0.04
HC= 17 MJ/kg
Ysoot=0.07
YCO=0.04
HC= 15 MJ/kg

Performance Criteria

The performance criteria have been taken primarily from PD7974-6:200416. These values are consistent with
the values found in the literature. Two exceptions are applied to the criteria, first is the relaxed values
allowed for sprinklered buildings. In New Zealand sprinkler systems have a rigorous inspection and
maintenance regime that helps to ensure that the system will function as designed when required. In addition
the current level of modeling does not adequately take into account the positive effect sprinklers can have so
the relaxation of the performance criteria is necessary to promote the use of sprinklers. The second
relaxation is that the performance criteria are not assessed within the household unit of origin.
Two performance criteria are suggested; the simple criteria are used when the smoke layer is not expected to
impact the egressing occupants and greatly simplifies the analysis. The second more detailed criteria are
used whenever the occupants are expected to have to egress through the smoke.

14

Occupant life safety - simple criteria

The simple criteria are used when the smoke layer is not allowed to reach the occupants.
1. minimum clear smoke layer height of 2.5 m
2. maximum upper layer temperature of 200C
Obviously, this method will not be suitable for spaces with low ceilings or where a distinct layer interface
cannot be determined.
Occupant life safety - detailed criteria
The detailed criteria are applied when the occupants are assumed to be egressing through the smoke. Three
criteria, all must be achieved. Calculations should be in accordance with ISO/TS 1357120. FEDs and visibility
may be determined at a height of 2.0 m above floor level using upper/lower layer properties as applicable, or
else can be based on upper layer properties alone.
1. Fractional Effective Dose (FED) for narcotic (toxic) gases. This accounts for the cumulative effects
of CO, O2 depletion and CO2 effects on respiration rate.
FED 0.3 (suitable for most general occupancies)
2. Fractional Effective Dose (FED) for radiant and convective heat. This accounts for cumulative
exposure to skin to radiant heat (2nd degree burns) and to convective heat from air.
FED 0.3 (suitable for most general occupancies)
3. Visibility
Visibility not less than 5 m, for rooms/spaces 100 m
Visibility not less than 10 m, for rooms/spaces > 100 m
4. Sprinklered buildings (System must be installed according to NZS454121 or NZS451522)
Visibility criteria does not apply
FED thermal does not apply
FED Narcotic < 0.3
5. Within household unit of fire origin analysis tenability criteria is not assessed.

Premovement Times
In New Zealand, there exist the evacuation regulations which require most commercial buildings open to the
public to have an approved evacuation scheme. As a result there is a widespread culture of evacuating a
building when the fire alarm sounds. Therefore shorter times then are typically found in the literature have
been suggested:
Description of building use

Premovement
Time (s)
Buildings where the occupants are considered awake alert and familiar with the building. Such
as offices, warehouse not open to the public, etc
Fire Cell of Origin
30
Remote from the Fire Cell of Origin
60
Buildings where the occupants are considered awake, alert and unfamiliar with the building.
Such as retail shops, exhibition space, restaurants,
Fire Cell of Origin (Standard Alarm Signal)
60
Remote from the Fire Cell of Origin (Standard Alarm Signal)
120
Fire Cell of Origin (Voice Alarm Signal)
30
Remote from the Fire Cell of Origin (Voice Alarm Signal)
60
Buildings where the occupants are considered sleeping and familiar with the building. Such as
Sleeping Residential
Fire Cell of Origin (Standard Alarm Signal)
60
Remote from the Fire Cell of Origin (Standard Alarm Signal)
300

15

Buildings where the occupants are considered sleeping and unfamiliar with the building. Such as
Sleeping Accommodation
Fire Cell of Origin
60
Remote from the Fire Cell of Origin (Standard Alarm Signal)
600
Remote from the Fire Cell of Origin (Voice Alarm Signal)
300
Buildings where the occupants are considered awake and under the care of trained staff and
unfamiliar with the building. Such as day care, dental office, clinic
Fire Cell of Origin (independent of alarm signal)
60
Remote from the Fire Cell of Origin (independent of alarm signal)
120
Buildings where the occupants are considered to be asleep, under the care of trained staff. Such
as hospitals and rest homes. (PG3 & PG4)
Room of Origin (independent of alarm signal)
180
Fire Cell of Origin
300
Remote from the Fire Cell of Origin (independent of alarm signal)
1800
Spaces which have only focused activities such cinemas, theatres, stadiums, etc
Evacuation starts when fire reaches 500 kW or 60s after detection which
0
ever is first.
Table 3 - Premovement times for proposed in the New Zealand Performance Based Design Framework

SUMMARY
Performance based design can provide for cost effective and innovative solutions to fire safety challenges.
However, allowing the designer to specify the design fire scenarios, design fires, premovement times, and
acceptance criteria can result in inconsistent levels of safety in building designs and can make it difficult for
the AHJ. Unfortunately the literature is primarily focused on qualitative guidance and is reluctant to give
quantitative guidance for use in PBD. Thus it is up to the regulating authority to specify the input values and
acceptance criteria for PBD. The specified input should include the design fire scenarios, design fires,
premovement times, and acceptance criteria which address societys tolerable risk to life safety from fire.

ACKNOWLEDGEMENTS
The author wishes to acknowledge the other members of the Department of Building and Housing - Fire
Safety Design Framework Project workgroup who have helped the author to form the ideas presented in this
paper and have worked diligently to develop the proposed framework described here.
Paula Beever
Colleen Wade
Dennis Pau
Ian Miller
Nick Saunders
Peter Thorby
Rob Hallows

New Zealand Fire Service

QStar Solutions
University of Canterbury
Heimdall Consulting Ltd
Department of Building and Housing
Department of Building and Housing
Department of Building and Housing

16

REFERENCES
1

NFPA101, Life Safety Code, 2009 Edition, National Fire Protection Association, 2009.

Compliance Document for New Zealand Building Code Clauses C1, C2, C3, C4, Fire Safety, Department of Building
and Housing, http://www.dbh.govt.nz/UserFiles/File/Publications/Building/Compliance-documents/clause-c1-c2-c3c41nov-2008.pdf, 2008.
3

SFPE Guide to Performance-Based Fire Protection, 2nd Edition, Society of Fire Protection Engineers and National
Fire Protection Association, Boston, MA, USA, 2007
4

Custer, R L P, Meacham, B J, Introduction to Performance Based Fire Safety, Society of Fire Protection Engineers
and National Fire Protection Association, 1997.

ISO/TS16733 Fire safety engineering Selection of design fire scenarios and design fires, International Organization
for Standardization, 2006.

NFPA5000, Building Construction and Safety Code, 2009 Edition, National Fire Protection Association, 2009.

Babrauskas, V. and Peacock, R. D., "Heat Release Rate: The Single Most Important Variable in Fire Hazard," Fire
Safety Journal, 18, 255-272, 1992.
8

NPFA72 National Fire Alarm Code, National Fire Protection Association, Boston, MA, USA, 2002.

NFPA204 Guide to Smoke and Heat Venting, National Fire Protection Association, 1998.

10

Babrauskas, V., Fire Modeling Tools for FSE: Are They Good Enough? Journal of Fire Engineering, 8 (2) 87-96,
1996.

11

Kawagoe, K, Fire Behaviour in Rooms, Report of the Building Research Institute, No. 27, Ibaraki-ken, Japan, 1958.

12

Fleischmann, C M and Parkes, A R, Effects of Ventilation on the Compartment Enhanced Mass Loss Rate, Fire
Safety Science Proceedings of the fifth International Symposium, International Association for Fire Safety Science,
Interscience Communications, London, UK, 415-426, 1997.

13

Purser, D A, Assessment of Hazards to Occupants from Smoke, Toxic Gases, and Heat, The SFPE Handbook for
Fire Protection Engineering, Section 2 Chapter 6, 4th Edition, 2-96 to 2-193, 2008.
14

Gann, R G, and Bryner, N P, Combustion Products and Their Effects on Life Safety, Fire Protection Handbook,
20th Edition, Section 6 Chapter 2, 6-11 to 6-35, 2008.

15

ISO/TS113571, Life-threatening components of fire- Guidelines for the estimation of time available for escape using
fire data, International Organization for Standardization, 2007.

16

PD7974-6:2004 The application of fire safety engineering principles to fire safety design of buildings Part 6: Human
factors: Life Safety strategies Occupant evacuation, behaviour and condition, British Standards, 2004

17

Gwynne, S M V and Rosenbaum, E R, Employing the Hydraulic Model in Assessing Emergency Movement, The
SFPE Handbook for Fire Protection Engineering, Section 3 Chapter 13, 4th Edition, 3-373 to 3-396, 2008.

18

Custer, R L P, Meacham, B J, Schifiliti, R P, Design of Detection Systems, The SFPE Handbook for Fire
Protection Engineering, Section 4 Chapter 1, 4th Edition, 4-1 to 444, 2008.

19

Proulx, G, Evacuation Time, The SFPE Handbook for Fire Protection Engineering, Section 4 Chapter 1, 4th
Edition, 4-1 to 444, 2008.
20

ISO/TS 13571 life-threatening components of fire guidelines for the estimation of time available for escape using
fire data. International Organization for Standardization.

21

NZS4541:2007, Automatic Fire Sprinkler Systems, Standards New Zealand, 2007.

22

NZS4515:2003, Fire Sprinkler Systems for Residential Occupancies, Standards New Zealand, 2003.

17