External Fire Spread

EXTERNAL FIRE SPREAD TO
ADJOINING BUILDINGS
- A review of fire safety design
guidance and related research
Emil Carlsson
Department of Fire Safety Engineering
Lund University, Sweden
Brandteknik
Lunds tekniska hgskola
Lunds universitet
Report 5051, Lund 1999
EXTERNAL FIRE SPREAD TO ADJOINING BUILDINGS

- A review of fire safety design guidance and related
research
Emil Carlsson
Lund 1999
-I-
-II-
External fire spread to adjoining buildings A review of fire safety design guidance and related
research
Utvndig brandspridning till intilliggande byggnader En utvrdering av riktlinjer fr
brandteknisk dimensionering samt relevant forskning
Emil Carlsson
Report 5051
ISSN: 1402-3504
ISRN: LUTVDG/TVBB5051--SE
Number of pages: 125
Illustrations: Emil Carlsson
Keywords:
Building Code, enclosing rectangle, external fire spread, flame projection,

ignition, notional radiator, radiation
Abstract:
This report discusses different parameters that affect external fire spread between
buildings. A comparison between different methods and regulations of
determining safe separation distances between buildings is made. The
importance of external flames projected from openings is investigated with
regards to the amount of emitted and received radiation. Finally, areas of further
research are suggested. (English)
While every effort has been made to ensure that the content within this report is true and correct,
errors can never be guaranteed against, so the reader should show due care in using any of the
reports content.
Copyright: Brandteknik, Lunds tekniska hgskola, Lunds universitet, Lund 1999.
Brandteknik
Lunds tekniska hgskola
Lunds universitet
Box 118
221 00 Lund
Department of Fire Safety Engineering

Lund University
P.O. Box 118
SE-221 00 Lund
Sweden
[email protected]
http://www.brand.lth.se
[email protected]
http://www.brand.lth.se/english
Telefon: 046 - 222 73 60

Telefax: 046 - 222 46 12
Telephone: +46 46 222 73 60

Fax: +46 46 222 46 12
-III-
-IV-
Acknowledgements
ACKNOWLEDGEMENTS
This report was conducted as the final requirement for a degree of Bachelor of
Science in Fire Protection Engineering at the Department of Fire Safety
Engineering, Lund University, Sweden.
I would like to thank and show my greatest appreciation to a number of people.
These people listed below have helped me with the necessary information and
support to perform this project and to obtain my degree.
Robert Jnsson, my supervisor, for always finding time to help me when

I was stuck in the wrong track and for answering all my strange
questions.
Per-ke Olsson, my second supervisor and a Fire Engineer at Arup Fire

in Melbourne, for initiating this project and providing me with essential
information and ideas.
Peter Johnson, Arup Fire in Melbourne, for giving me the opportunity to

come to Australia, to do work experience and perform the initial part of
my research.
Margaret Law, Arup Fire in London, for useful help regarding radiation
from external flames.
Chris Marrion, Arup Fire in New York, for providing me with

information and answers about the fire regulations in the United States
of America.
Greg North, my dear friend from New Zealand, for proof-reading my

report and giving me new ideas.
My family, both my parents Cristina and Leif, and my sister Lisa, for
great help and support during all my years of studies. Without you all, I
would never be where I am!
The Fire Engineering Class of 1996, for help and many good times, both
in and out of school, during our four years at Lund University. I will
miss you all!
Emil Carlsson
17th December, 1999
-V-
Acknowledgements
-VI-
Executive summary
EXECUTIVE SUMMARY
At the present, there is no internationally accepted method to design buildings against
external fire spread between buildings. National Building Codes are generally based
on more or less prescriptive provisions and usually no background is given on how
the provisions have been obtained. Where engineering methods can be used, i.e. in a
performance based Building Code environment, the guidance is not globally
consistent.
This project was initiated in order to clarify what parameters that should be
considered when designing buildings against external fire spread and what factors that
are not necessary to take into account. Additionally, the report should also try to make
indications of what methods that should be used to determine the safe separation
distances between buildings.
THE OBJECTIVE of this report is to summarise and review research with regards to
external fire spread between buildings. Furthermore, available fire engineering
guidance literature, calculation methods and Building Regulations addressing safe
separation distances between buildings will be investigated.
The report will also investigate whether the levels of radiation emitted from projected
flames are large enough for it to be necessary to account for when determining safe
separation distances between buildings. The alternative is to use the openings as the
only radiators with an appropriately determined temperature.
THE METHOD used to achieve the results in this report was a combination of an
extensive literature study and assessments of different calculation methods. The
regulations set out in different Building Codes have been compared to each other. A
number of calculations of separation distances between buildings and acceptable
unprotected areas have been performed for three different building types, using
calculation methods referred to by different Building Codes and other methods set out
in the literature.
The importance from externally projected flames was investigated in two ways. A
method derived in Sweden of how to determine the total amount of received radiation
by an adjoining building, was compared with the radiation received if only the
window was assumed to radiate at an appropriate determined temperature. This
radiation from the window was also compared with calculations performed with
equations for flame projections and flame temperatures that are commonly used in the
fire engineering discipline.
THE CONCLUSIONS that have been made based on the findings in this report are listed
below.
Since the radiation intensity is dependent on the temperature to the power

of four, the fire temperature is a very important factor when determining
the emitted radiation from a burning building. The fire temperature is
dependent on several factor such as compartment dimensions, opening
factors, size and distribution of the fire load and type of interior linings.
-VII-
Executive summary
The value of 12.5 kW/m2 for piloted ignition of an adjoining facade may
be slightly low and thus too conservative. Ignition due to exposure from
full scale fires may occur at higher levels of radiation, i.e. 15-18 kW/m2.
The prescriptive Building Regulations studied in this report use values of

84 and 168 kW/m2 (2 and 4 cal/cm2s) as the maximum values of
radiation that can be estimated in a fire.
The Building Regulations in Australia, England and Wales, New Zealand

and Sweden are performance based, which gives greater flexibility to the
fire engineer in the design process of buildings.
The calculation methods studied in this report predicted approximately

the same separation/boundary distance for simple building shapes. The
differing factor was that various flame projection distances are used in
the methods. Complicated building shapes, e.g. irregular shaped facades,
generated a larger distribution of separation distances, which is believed
to be due to the variation in applicability of the different methods.
The calculations performed with a method by Fredlund et al. (1976) as

well as with a method set up by the author indicated that the percentage
of radiation contributed by externally projected flames to the total
emitted and received radiation was in the order of 12-18 %. This
contribution is not very large, which implies that the simpler approach of
assuming that the openings, with an appropriate determined temperature,
are the only radiators, is a method that can be used for design purposes.
Areas that need further research are; radiation from projected flames,
ignition criteria of materials exposed to full scale fires, temperatures in
fires in modern buildings and wind effects on the levels of emitted
radiation. Another area that should be investigated more thoroughly is
whether the whole or only a portion of the facade should be counted as
radiating when determining separation distances for large warehouses.
-VIII-
Table of contents
TABLE OF CONTENTS
1
INTRODUCTION ........................................................................................................................ 1
1.1
1.2
1.3
1.4
BACKGROUND ............................................................................................................................ 1
OBJECTIVES ................................................................................................................................ 1
METHOD ..................................................................................................................................... 2
LIMITATIONS .............................................................................................................................. 2
PARAMETERS AFFECTING FIRE SPREAD BETWEEN BUILDINGS............................ 3

2.1 FIRE SPREAD BETWEEN BUILDINGS ............................................................................................. 3
2.1.1
Flying brands ................................................................................................................... 3
2.1.2
Flame contact................................................................................................................... 3
2.1.3
Convective heat transfer .................................................................................................. 3
2.1.4
Radiative heat transfer..................................................................................................... 3
2.2 EMITTED RADIATION .................................................................................................................. 4
2.2.1
Fire temperature .............................................................................................................. 4
2.2.2
Fuel .................................................................................................................................. 8
2.2.3
Properties of compartment............................................................................................... 9
2.3 RADIATION TRANSFER ................................................................................................................ 9
2.3.1
Flame projection .............................................................................................................. 9
2.3.2
Emissivity ....................................................................................................................... 12
2.3.3
Configuration factor ...................................................................................................... 12
2.3.4
Fire Service Intervention ............................................................................................... 14
2.4 CRITICAL RECEIVED RADIATION ............................................................................................... 14
2.4.1
Ignition criteria .............................................................................................................. 14
2.4.2
Time to ignition .............................................................................................................. 15
BUILDING CODES ................................................................................................................... 17

3.1 AUSTRALIA .............................................................................................................................. 17
3.1.1
Deemed-to-Satisfy Provisions ........................................................................................ 17
3.1.2
Alternative Solution ....................................................................................................... 17
3.1.3
Objective ........................................................................................................................ 17
3.1.4
Performance Requirements............................................................................................ 18
3.1.5
Verification Methods...................................................................................................... 18
3.1.6
Summary ........................................................................................................................ 19
3.2 CANADA ................................................................................................................................... 19
3.3 ENGLAND & WALES................................................................................................................. 20
3.3.1
Requirement ................................................................................................................... 20
3.3.2
Performance................................................................................................................... 20
3.3.3
Space separation ............................................................................................................ 20
3.3.4
Calculation methods ...................................................................................................... 22
3.3.5
Summary ........................................................................................................................ 25
3.4 NEW ZEALAND ......................................................................................................................... 25
3.4.1
Objective ........................................................................................................................ 25
3.4.2
Functional Requirements ............................................................................................... 25
3.4.3
Performance................................................................................................................... 26
3.4.4
Building Separation ....................................................................................................... 26
3.4.5
Summary ........................................................................................................................ 29
3.5 SWEDEN ................................................................................................................................... 29
3.5.1
Regulation ...................................................................................................................... 29
3.5.2
Advice............................................................................................................................. 30
3.5.3
Objective ........................................................................................................................ 30
3.5.4
Requirements.................................................................................................................. 30
3.5.5
Performance based design ............................................................................................. 30
3.5.6
Summary ........................................................................................................................ 31
3.6 THE UNITED STATES OF AMERICA ........................................................................................... 32
-IX-
Table of contents
CALCULATION METHODS................................................................................................... 33
4.1
4.2
4.3
4.4
4.5
4.6
4.7
COMPARISON OF BOUNDARY REQUIREMENTS .......................................................... 51

5.1
5.2
5.3
5.4
ENCLOSING RECTANGLES ......................................................................................................... 33

AGGREGATE NOTIONAL AREAS................................................................................................. 35
PETER COLLIER ........................................................................................................................ 36
C.R. BARNETT.......................................................................................................................... 40
WILLIAMS-LEIR ....................................................................................................................... 42
J.H. MCGUIRE.......................................................................................................................... 43
NFPA 80A ............................................................................................................................... 44
WAREHOUSE ............................................................................................................................ 51
OFFICE BUILDING ..................................................................................................................... 52
RESIDENTIAL BUILDING............................................................................................................ 53
DISCUSSION OF RESULTS .......................................................................................................... 55
IMPORTANCE OF RADIATION FROM PROJECTED FLAMES.................................... 57

6.1 FULL SCALE FIRE TESTS ............................................................................................................ 57
6.2 SWEDISH METHOD .................................................................................................................... 57
6.2.1
Tests ............................................................................................................................... 58
6.2.2
Limitations ..................................................................................................................... 58
6.2.3
Theory ............................................................................................................................ 59
6.2.4
Calculation methodology ............................................................................................... 62
6.3 PROJECTED FLAME OR NOTIONAL RADIATOR ............................................................................ 63
6.3.1
Swedish method.............................................................................................................. 63
6.3.2
Hand calculation model ................................................................................................. 70
6.3.3
Discussion of results ...................................................................................................... 75
DISCUSSION AND CONCLUSIONS...................................................................................... 77

7.1
7.2
7.3
7.4
RADIATION ............................................................................................................................... 77
COMPARISON OF BUILDING CODES .......................................................................................... 78
COMPARISON OF CALCULATION METHODS ............................................................................... 78
PROJECTED FLAME OR NOTIONAL RADIATOR ............................................................................ 79
FURTHER RESEARCH ........................................................................................................... 81
REFERENCES ........................................................................................................................... 83
APPENDIX 1 COMPARISON OF BOUNDARY REQUIREMENTS FOR WAREHOUSE.... 87

APPENDIX 2 COMPARISON OF BOUNDARY REQUIREMENTS FOR OFFICE
BUILDING............................................................................................................... 91
APPENDIX 3 COMPARISON OF BOUNDARY REQUIREMENTS FOR RESIDENTIAL
BUILDING............................................................................................................... 97
APPENDIX 4 IMPORTANCE OF RADIATION FROM PROJECTED FLAMES................ 103
-X-
List of figures
LIST OF FIGURES
Figure 2.1
Figure 2.2
Figure 2.3
Figure 2.4
Figure 3.1
Figure 3.2
Figure 3.3
Figure 4.1
Figure 4.2
Figure 4.3
Figure 4.4
Figure 5.1
Figure 5.2
Figure 5.3
Figure 6.1
Figure 6.2
Figure 6.3
Figure 6.4
Figure 6.5
Figure 6.6
Figure 6.7
Figure 6.8
Figure 6.9
Figure 6.10
Figure 6.11
Figure 6.12
Figure 6.13
Figure 6.14
Figure 6.15
Example of the Swedish fire curves for fire loads between 50 and 800 MJ/m2. .............. 5
Time-temperature curves for ASTM E119 and ISO 834 standard furnace tests .............. 6
Flame shape for no through draught conditions............................................................. 10
Determination of configuration factors .......................................................................... 13
External walls facing the relevant boundary .................................................................. 21
The mirror image concept........................................................................................... 23
Principles of Method 1. Adapted from Approved Document B..................................... 24
Enclosing rectangles ...................................................................................................... 33
Projection of openings onto a plane of reference ........................................................... 37
Configuration factor and radiation distance ................................................................... 39
Separation of openings from a specific point at the exposed building ........................... 47
Office building used in the comparison ......................................................................... 52
Simplified layout of the residential building.................................................................. 53
Residential building used in the comparison.................................................................. 54
Orientations used when calculating the received radiation ............................................ 59
Simplified layout of exterior projected flame used in the model ................................... 60
Coefficient C used in the method by Fredlund et al. (1976). ......................................... 60
Flame emissivity as a function of flame thickness for wood based fuel ........................ 61
Comparison between theoretical derived calculation model and full scale fire tests ..... 62
Comparison of received radiation for all projection distances for case 1....................... 65
Received radiation for case 1 ......................................................................................... 65
Received radiation for case 1b ....................................................................................... 66
Flame shape and dimensions used in the calculation procedure .................................... 71
Rectangles used when determining the configuration factor of each flame strip ........... 73
-XI-
List of figures
-XII-
List of tables
LIST OF TABLES
Table 2.1
Table 2.2
Table 2.3
Table 2.4
Table 3.1
Table 3.2
Table 3.3
Table 3.4
Table 3.5
Table 3.6
Table 3.7
Table 3.8
Table 3.9
Table 4.1
Table 4.2
Table 4.3
Table 4.4
Table 4.5
Table 5.1
Table 5.2
Table 5.3
Table 6.1
Net calorific value and heat release rates for different material............................................ 8
Absorption coefficient for flames from different materials. ............................................... 12
Critical levels of radiation from full scale and laboratory tests. ......................................... 15
Properties of various timber species obtained with Janssens model. .................................. 16
Required distances for buildings located on different properties........................................ 18
Required distances for buildings located on the same allotment. ....................................... 18
Separation distances and maximum unprotected areas. ...................................................... 23
Separation distance and maximum acceptable unprotected areas....................................... 24
Fire hazard categories and fire load energy densities of buildings. .................................... 26
Building separation for unsprinklered multi-unit residential dwellings and buildings with
temporary accommodation................................................................................................. 27
Building separation for unsprinklered buildings with a fire hazard category of 1. ............. 28
Multiplication factor for various distances between unprotected ....................................... 29
Performance requirements of firewalls.. ............................................................................. 31
Multiplication factor for various distances. ........................................................................ 36
Maximum permitted radiation at the facade of the neighbouring building......................... 39
Classification of fire severity according to NFPA 80A. ..................................................... 45
Guide numbers for the determination of separation distances according to NFPA 80A..... 46
Required separation distance or height of protection.......................................................... 48
Comparison of required boundary distance for the warehouse used in the study ............... 51
Comparison of required boundary distance for the office building used in the study......... 53
Comparison of required boundary distance for the residential ........................................... 54
Properties and results from the sample calculations ........................................................... 75
-XIII-
List of tables
-XIV-
Introduction
1 INTRODUCTION
This introduction will present the reason for the initiation of this project and what the
objectives of the report are. A brief description of the used methods and what
limitations that has been necessary to make for the fulfilment of the report is also
given.
This report was conducted by Emil Carlsson as the final requirement for a degree of
Bachelor of Science in Fire Protection Engineering at the Department of Fire Safety
Engineering, Lund University, Sweden.
1.1
Background
At the present, there is no internationally accepted method to design against external

fire spread between buildings. National Building Codes are generally based on more
or less prescriptive provisions and generally no background is given on how the
provisions have been obtained. Where engineering methods can be used, i.e. in a
performance based Building Code environment, the guidance is not globally
consistent.
Very little published reviews regarding external fire spread between buildings have
been presented during the latest 15 years at conferences or in the literature. This is a
concern given the development in the performance based fire engineering codes
during the last decade.
This project was initiated in order to clarify what parameters that should be
considered when designing buildings against external fire spread and what factors that
are not necessary to take into account. Additionally, the report will also try to make
indications of what methods that should be used to determine the safe separation
distances between buildings.
1.2
Objectives
The objectives of this report are to summarise and review research performed during
the latest 50 years with regards to external fire spread between buildings.
Furthermore, available fire engineering guidance literature and calculation methods
addressing safe separation distances between buildings will be investigated.
The report shall also study the Building Regulations and acceptable solutions of a few
countries. Some of the regulations allow for a performance based design while others
are strictly prescriptive. Differences and similarities between the different regulations
will be compared and discussed.
The report shall also investigate the importance of radiation from projected flames
with regards to determining safe separation distances between buildings. Will a flame
projected from an opening during a fire, contribute with high enough levels of
radiation for it to be necessary to take into account, or is it safe to assume that the
openings or windows are the only radiators?
-1-
Introduction
A very important purpose with this report is to identify areas, regarding external fire
spread between buildings that need further research.
1.3
Method
A combination of an extensive literature study and assessments of different

calculation methods have achieved the results presented in this report. The regulations
set out in different Building Codes have been compared to each other and the scope
discussed. A number of calculations of separation distances between buildings and
acceptable unprotected areas have been performed for three different building types,
using calculation methods referred to by different Building Codes and other methods
set out in the literature.
The importance from externally projected flames was investigated in two ways. A
method derived in Sweden of how to determine the total amount of received radiation
by an adjoining building, was compared with the radiation received if only the
window was assumed to radiate at an appropriate determined temperature. This
radiation from the window is also compared with calculations performed with
equations for flame projections and flame temperatures that are commonly used in the
fire engineering discipline.
Based on findings and ideas that have arisen during the progress of the report, a
number of suggestions were made of areas of further research.
1.4
Limitations
The report will only discuss horizontal fire spread between buildings due to thermal
radiation, i.e. other aspects such as ignition due to convective heat transfer and flying
brands will only be mentioned and not discussed in detail. The different countries
Building Codes that will be compared and investigated are; Australia, Canada,
England and Wales, New Zealand, Sweden and the United States of America.
Furthermore, the building types that will be used in the study are warehouses, office
buildings and residential buildings.
-2-
Parameters affecting horizontal fire spread
2 PARAMETERS AFFECTING FIRE SPREAD BETWEEN BUILDINGS

This chapter aims to point out the most important parameters that influence fire
spread between buildings. A brief description will first be made about how fire may
spread from building to building. Secondly, a more detailed explanation will be made
concerning the factors that affect fire spread due to radiation. The detailed explanation
is divided into three sub-chapters, which addresses the following parameters; emitted
radiation, radiative heat transfer and critical received radiation.
2.1
Fire spread between buildings
The spread of fire from a burning building to an adjoining building can occur in a
number of different ways. It has been found that some ways, or a combination of
ways, are more common and often more hazardous than others.
2.1.1 Flying brands
Ignition of combustible materials may occur due to flying brands emitted from a
building on fire, (Barnett 1988, Jnsson et al. 1994). These brands may travel far
distances and protection against this is possible by fitting external surfaces with
appropriate fire resistant claddings, (McGuire 1965). Flying brands do not represent a
significant hazard by itself with respect to ignition of buildings. They may though act
as an igniting source together with radiation, where the volatiles given off by the
radiation exposed material may ignite.
2.1.2 Flame contact
It is possible that projected flames from an opening may impinge onto an adjoining
building and cause ignition. The projection distance, i.e. the horizontal extension of
the flame from the facade, and the flame length, i.e. the vertical extension of the
flame, are dependent on many factors including geometry and size of the opening as
well as wind conditions and mass burning rate, (Barnett 1988, Law et al. 1981).
2.1.3 Convective heat transfer
Convective heat transfer may also result in the ignition of an adjoining building, given
that the stream of hot gases hitting the building can be several hundreds of degrees
Centigrade, (McGuire 1965). In order for ignition to occur by this phenomenon, the
exposed building has to be very close to the fire source.
2.1.4 Radiative heat transfer
Ignition due to radiation is the most common way for fire to spread between
buildings, (Jnsson et al. 1994), and can happen at much greater distances than by
direct flame contact and convection, (McGuire 1965). Typical values of radiation
where ignition of wood may occur is 33.5 kW/m2 for spontaneous ignition, i.e.
ignition in absence of an ignition source, and 12.5 kW/m2 for piloted ignition, i.e.
ignition in presence of an ignition source such as a spark or a brand, (Barnett 1988).
These values are widely used throughout the world.
-3-
Since ignition by radiative heat transfer is the most common and hazardous way of
fire to spread between buildings, the other means of fire spread explained above will
not be further discussed in this report.
2.2
Emitted radiation
A body with a certain temperature, T, emits energy, P, to the surroundings at all times,
(Jnsson et al. 1994). This transfer of energy can be described by Equation 2.1 below.
The emitted energy is expresses in terms of energy emitted per second an area unit,
i.e. J/sm2 or W/m2.
P = T 4
where:
=
=
=
=
[2.1]
emitted radiation [W/m2]

emissivity of the radiating surface
Stefan-Boltzmann constant = 5.6710-12 W/m2K4
temperature of object [K]
How this equation is used for fire situations, where flames and a hot gaseous mass are
assumed to be an object, and the importance of the different variables are discussed
later in this report. However, since the temperature is taken to the fourth power, it is
obvious that the temperature was the greatest impact on the total amount of emitted
radiation.
2.2.1 Fire temperature
There are several methods available on how to predict the temperature of a fire in a
compartment. This report will not discuss these methods in detail but will give a brief
summary of some of the methods and factors influencing the temperature. However,
since the emitted radiation from a burning building and hence the building separation
distance is dependent on the fire temperature to the power of four, this factor should
be investigated thoroughly in a performance based design.
A compartment fire can be either fuel controlled or ventilation controlled, (Law
1963). A fuel controlled fire has an excess of oxygen and not sufficient fuel, while a
ventilation controlled fire has an excess of fuel and the limiting factor is oxygen. Law
found that in ventilation controlled fires, the rate of heat release rate and hence the fire
temperature is dependent on the area of the opening multiplied by the square root of
the opening height, i.e. A h . This factor is often referred to as the ventilation factor.
The maximum compartment temperature found in the tests reviewed by Law (1963)
was 1100 C, which represents a maximum radiation intensity of 168 kW/m2 (4
cal/cm2s). In the tests where the ventilation factor was a non-limiting factor, i.e. where
the window area is comparable to the floor area, approximately the same temperatures
and levels of radiation were found. For fires with a relatively low fire load, i.e. less
than 25 kg/m2 (the amount of combustible material per unit floor area), temperatures
of 800 C were found in the compartments. This temperature represents a radiation of
84 kW/m2 (2 cal/cm2s). These values have been and are still used in several
prescriptive Building Regulations as the maximum levels of radiation that can be
expected in a fire.
-4-
Drysdale (1985) reports on a method to predict fire temperatures in compartments

developed in Sweden by Pettersson et al. (1976). The results of the method are often
referred to as the Swedish fire curves. The method uses heat balance equations for the
fully developed compartment fire to predict the temperature.
The following assumptions has been made in order to simplify the method:
combustion is complete and takes place entirely within the confines of the
compartment;
the temperature is uniform within the compartment at all times;
a single surface heat transfer coefficient may be used for the entire inner
surface of the compartment; and
the heat flow to and through the compartment boundaries is
unidimensional, i.e. corners and edges are ignored and the boundaries
are assumed to be infinite slabs.
The full derivation of the Swedish fire curves will not be presented here. The
interested reader is referred to Drysdale (1985) or the original document by Pettersson
et al. (1976). The result is an equation that can be solved by numerical integration to
give a time-temperature relationship for various values of the opening factor
A h / A tot , where A is the area of the opening, h is the height of the opening and Atot
is the total area of all interior surfaces in the enclosure. A number of time-temperature
relationships have been established for different compartments, window
configurations and fire loads. Figure 2.1 below is an example of Swedish fire curves.
Each curve represents a specific fire load.
1200
1000
800
50 (MJ/m^2)
150 (MJ/m^2)
Temperature ( C) 600
400 (MJ/m^2)
800 (MJ/m^2)
400
200
0
0
0,5
1,5
2,5
Time (h)
Figure 2.1
Example of the Swedish fire curves for fire loads between 50 and 800 MJ/m2. The
above example is for A h /Atot =0.08m1/2, wall thickness = 0.2 m, k = 0.8 W/mK and
c = 1700 kJ/m3K and is derived from data in Drysdale (1985).
-5-
This theoretical model has been compared with full scale fire tests, (Drysdale 1985).
The comparison showed that the model can be used with satisfactory result to
estimate the fire temperature in a compartment.
Another method that is used to determine the fire temperature is with the use of
standard time-temperature curves. Building elements are tested against fire exposures
in large furnaces that can be programmed to produce certain temperatures, (Clarke
1998). In order to make tests conducted in different places and in different furnaces
comparable, time-temperature curves have been developed that the furnaces should
obey. The two most common tests are ASTM E119 and ISO 834 and expressions for
the temperature, T, determined by these tests are given in Equation 2.2a-b and
presented graphically in Figure 2.2.
)+ 170.41
ASTM E119:
T = 750 1 e 3.79553
ISO 834:
T = 345 log(8t + 1) + To
where:
To
t
=
=
[2.2a ]
t + To
[2.2b]
temperature of ambient air [C]

time in hours for ASTM E119 and time in minutes
for ISO 834
1400
1200
1000
800
ASTM E119
Temperature ( C)
ISO 834
600
400
200
0
0
0,5
1,5
2,5
3,5
Time (hours)
Figure 2.2
Time-temperature curves for ASTM E119 and ISO 834 standard furnace tests
It should be noted that the time-temperature curve of the furnace tests is different
from those resulting from a real fire. The growth phases are more rapid and the
furnace tests lack any decay phases. Barnett (1988) proposes that the ISO 834
standard furnace time-temperature curve can be used for the purpose of building
separation. Exposure to the ISO 834 fire for a duration of 30 and 120 minutes would
result in the same levels of emitted radiation as is used in the British and New Zealand
Building Regulations, (Clarke 1998). Drysdale (1985) quotes work on Ingberg (1928),
who proposed that if the areas under two time-temperature curves were equal, the
fires could be assumed to have equal fire severity. The standard furnace relationship
-6-
could therefore be used if the fire severity was equal to the severity from an expected
real fire. Ingberg also derived tables for fire loads with belonging values of fire
resistance requirements that could be used for building design. Drysdale was doubtful
to this method mainly due to three reasons;
1. the fire load in the building has to stay the same during the entire lifetime
of the building,
2. the data on which the method was based, was conducted in old buildings,
and
3. the equal area concept is not totally true, e.g. the effect of a 10 minutes
exposure to 900 C will differ from that of a 20 minutes exposure to
600 C.
Drysdale (1985) concludes that the method by Ingberg (1928) may not be applicable
in the present day.
Law et al. (1981) give an expression, see Equation 2.3, for determining the fire
temperature Tf in a compartment. The equation was derived from experiments and
uses fire load, ventilation configuration and compartment dimensions as inputs.
T f Ta = 6000
where:
Ta
AT
=
=
Aw
h
=
=
(1 e
0.10
1/ 2
)(1 e
0.05
[2.3]
AT
[m-1/2]
A w h 1/2
L
[kg/m2]
1/2
(A w A T )
temperature of ambient air [K]
total area of floor, ceiling and walls minus total window
area [m2]
sum of window areas on all walls [m2]
window height or weighted average of window heights
on all walls [m]
fire load [kg]
This method will however not give a complete time-temperature relationship for the
fire. The output will be an average temperature of the fully developed fire that can be
expected for the specific compartment, fire load and window configuration.
In 1958, the Division of Building Research, National Research Council (Canada)
carried out a series of full scale fire tests that are commonly known as the St.
Lawrence Burns (McGuire 1965). The main findings from the tests are summarised
below.
The type of exterior cladding did not influence the levels of emitted and
received radiation.
The maximum levels of received radiation at some distance from the
buildings was found to be comparable to those that would result if the
-7-
window openings were assumed to be the only radiators radiating with an

appropriate temperature.
The maximum levels of radiation were twice as high for buildings with
highly flammable linings as for buildings with non-combustible linings.
The levels of radiation were found to be higher on the leeward side of the
building than on the windward side.
Another finding from the tests was that the levels of radiation measured after 16
minutes and onwards were much greater than expected and what would be practical
for the purpose of building separation design. However, since fire fighting activities
generally is in progress well before that time, the extraordinary high levels of
radiation were neglected when configuration factors for building separation design
were established. The findings from the St. Lawrence Burns has been incorporated in
the National Building Code of Canada, where it also is regulated that if Fire Service
intervention cannot be guaranteed within ten minutes, the building separation distance
should be doubled.
2.2.2 Fuel
The temperature in a compartment during a fire is also dependent on the type of fuel
that is burning. Plastics and other synthetic materials generally have a higher calorific
value than for example wood, which can cause a higher fire temperature (Buchanan
1994). Examples on calorific values are 16.7 MJ/kg for wood and 39.9 MJ/kg for
polystyrene. Table 2.1 below sets out calorific values and heat release rates per square
metre of some liquids, plastics and different shapes of wood.
Table 2.1
Material
Liquids
Petrol
Light oil
Wood
Flat wood
1 m cube
Furniture
25 mm in crib
Plastics
PMMA
Polyethylene
Polystyrene
Net calorific value and heat release rates for different material.
Adapted from Buchanan (1994).
Net calorific value (MJ/kg)
Heat release rate (MW/m2)
43.5
41.9
3.27
1.75
16.7
16.7
16.7
16.7
0.10
0.61
6.63
15.3
24.9
43.8
39.9
1.34
1.36
1.40
The distribution of the fuel within the compartment is also an important factor. An
example can be made for wooden pallets, which will cause a higher maximum heat
release rate than the same amount (in weight) of solid cubical shaped wood would do.
This is due to the greater surface area of the wood pallets. The wood pallets will
therefore achieve a higher maximum heat release rate but it will at the same time not
burn for as long time as the solid wood will do.
The type of fuel can also have an impact on the amount of emitted radiation from
flames projected from an opening. Plastics tend to cause a higher emissivity and
-8-
hence higher levels of emitted radiation than wood, (Karlsson et al. 1999). The
emissivity of projected flames is discussed in chapter 2.3.2.
2.2.3 Properties of compartment
In chapter 2.2.1 it was shown that there are a number of factors that affect the fire
temperature in a compartment and hence the emitted radiation. The thermal properties
of walls, ceilings and floors have an impact on the heat transfer through the
boundaries of the compartment. The geometry, i.e. width, depth and height, of the
compartment as well as the number and dimensions of windows have a great
influence on the fire temperature. The opening factor does also influence the
temperature since it can be a limiting factor with regards to combustion efficiency and
supplying the fire with oxygen. Furthermore, large openings can allow hot gases to
escape from the compartment, which thereby can result in lower temperatures within
the compartment. The opposite is valid for small openings, i.e. they can result in
higher temperatures in the compartment. It will also be shown later in this report that
the size of the windows influence the temperature and size of flames projected from
the window, which in some cases can result in higher levels of emitted radiation.
2.3
Radiation transfer
Parameters that influence the transfer of radiation between a burning building and a
receiving surface are for example; projections of flames from openings, the emissivity
of the flame, the configuration factor and the Fire Service intervention. These
parameters will be discussed below.
2.3.1 Flame projection
Flames projected out of openings in a burning compartment may cause ignition either
by emitting radiation to combustible objects or by direct flame contact with nearby
objects. The effect of the radiation contributed by external flames during a fire with
regards to the total amount of radiation received by an adjacent building is not totally
agreed on in the fire engineering discipline today. The National Building Code of
Canada uses a flame projection distance of 1.2 m for unprotected openings (Clarke
1998), while the other Building Codes reviewed in this report disregard the effect of
flame projection. The effect of flame projection will be investigated in a separate
chapter in this report and it is not the purpose of this section to further discuss the
matter. This section will only discuss methods of determining the properties of
projected flames.
Law et al. (1974) have reviewed work done by Webster (1959, 1961, 1964), Yokoi
(1960) and Seigel (1969) on flame projections from buildings on fire. Equation 2.4 is
given for the correlation between flame height above the compartment floor and the
ratio between mass burning rate and width of the opening, which was obtained from
test data for no wall conditions.
R
z1 + H = 18.6
W
where:
z1
H
=
=
2/3
flame length above top of window [m]

height of window [m]
-9-
[2.4]
W
R
=
=
width of window [m]

rate of weight loss of fuel [kg/s]
Figure 2.3 shows the layout and dimensions of projected flames according to Law et
al. (1981). Note that the method is for no through draught conditions.
2H/3
x
z1
2H/3
H
Flame axis
Figure 2.3
Flame shape for no through draught conditions according to method by Law et al.
(1981)
The flame tip is defined to be located at the point where the flame temperature is
540 C, which is where the illuminous zone of the flame ends. The rate of burning, R
(in kg/min), can be estimated by the well known expression presented in Equation 2.5,
which was originally derived by Kawagoe, (Drysdale 1985). The correlation below
was derived from a series of both full and small scale tests with wood cribs as fuel
and compartments with different ventilation openings.
R = 5.5 Ao H
where:
Ao
H
=
=
[2.5]
area of opening [m2]

height of opening [m]
The report (Law et al. 1981) also states that flames projected from narrow windows
tend to be longer and not lie against the facade, while flames projected from wide
windows will be shorter and cling to the wall above the window, which is also
described in Jnsson et al. (1994). Extra long flames can emerge where there is;
flammable linings in the fire compartment,

fires on floors below the window from where flames are projected,
resulting in a chimney effect and,
there is a wind across the fire.
-10-
Equation 2.4 for the flame height has later been changed to Equation 2.6 (Law et al.
1981), which is used in several documents, e.g. (Buchanan 1994, Jnsson et al. 1994,
Drysdale 1985).
R
z1 + H = 12.8
W
2/3
[2.6]
Law et al. (1981) presents a method of how to estimate flame shapes and determine
emitted radiation from projected flames. The method was derived in order to design
external steel members. The horizontal flame projection of the flame axis away from
the building facade can be determined with Equations 2.7a-b.
Wall above window:
x = 0.312 H 1.54W 0.54
No wall above window:
x = 0.60 H 2 / 3 z1
1/ 3
[2.7a]
[2.7b]
The dimensions are the same as explained above. The flame is stated to have a
thickness of d = 2H/3 , which means that the flame front is located at distance 2H/3
from the facade.
Fredlund et al. (1976) has conducted a series of full and small scale fire tests with
small houses made in light-weight concrete. In those tests, air was blown into the fire
compartment in order to simulate natural wind conditions. The projected flames
observed in these tests had a slightly different shape than those reported by Law et al.
(1981). The flame shape was triangular and projecting from the upper two thirds of
the window. These tests and the method of approximating the flame shape will be
further discussed in Chapter 6, where also a picture of the flame shape is shown.
In full scale tests performed in Sweden by Ondrus et al. (1986), the temperature in
projected flames was measured. The aim of the tests was to investigate the effect of
externally applied additional thermal insulation. The tests were conducted in a three
storey building and the fuel consisted of 184 kg of wood cribs. The maximum
recorded temperature recorded just outside the window was approximately 900950 C. The flame temperature then decreased with both horizontal and vertical
distance away from the window. Drysdale (1985) presents results from three small
scale tests performed by Bullen and Thomas (1979). The maximal temperature
recorded in the external flames was in the same order and distribution as found by
Ondrus et al. (1986).
Law et al. 1981 gives a theoretical expression of the temperature distribution along
the axis of a projected flame, see Equation 2.8.
Tz Ta
lw
= 1 0.027
To Ta
R
where:
Tz
Ta
=
=
temperature at distance l along flame axis [K]

-11-
[2.8]
To
l
w
R
=
=
=
=
temperature at the window [K]

distance along flame axis [m]
window width [m]
rate of burning [kg/s]
The temperature just outside the window may be found to be higher than the fire
temperature. This is due to unburned gases that are being combusted outside the
compartment.
2.3.2 Emissivity
The emissivity can be said to express how efficient a surface or an object is as a
radiator, (Jnsson et al. 1994). The emissivity is dependent on factors such as
temperature and properties of the object. An ideal black body is assumed to have an
emissivity of 1.
When it comes to flames, the emissivity is dependent on the thickness of the flame
and contents of soot particles. An emissivity of 0.3-0.7 is often used for ordinary
flames, while it can be as low as 0.066 for flames of burning alcohol, (Jnsson et al.
1994). Equation 2.9 below is often used to determine the emissivity of a flame
projected out from an opening in a burning compartment, (Law et al. 1981, Jnsson et
al. 1994, Drysdale 1985).
= 1 e a
where:
=
=
[2.9]
absorption coefficient
flame thickness [m]
The absorption coefficient is dependent on the type of burning material and Table 2.2
sets out values for flames of various fuels. According to Law et al. 1981 and the CEN
document Eurocode 3, Design of Steel Structures, Part 1.2 Structural Fire Design, a
value of a = 0.3 is appropriate for design purposes of external structural steel.
Table 2.2
Absorption coefficient for flames from different materials.

Adapted from Drysdale (1985).
Material
Diesel oil
PMMA
Polystyrene
Wood
Furniture
Coefficient a
0.43
0.50
1.20
0.50-0.80
1.13
An emissivity of 1 is often used for window radiation from compartments, (Barnett

1988, Law 1968), which is a conservative approach.
2.3.3 Configuration factor
The intensity of emitted radiation from a hot surface will decrease when the distance
from the surface increases. A measure of this decrease is the so called configuration
factor, . Buchanan (1994) gives an expression for the configuration factor of a
rectangular radiator and a remote receiver, see Equation 2.10 below.
-12-
=
where:
y
1 x
tan 1
2
360 1 + x 2
1+ x
x
y
Hf
Wf
r
=
=
=
=
=
y
1
+
tan
1+ y2
1+ y2
[2.10]
Hf/r
Wf/r
height of rectangle [m]
width of rectangle [m]
distance between radiating and receiving surface [m]
The above equation determines the configuration factor in one of the corners of the
rectangle. Several documents, e.g. (Drysdale 1985, Jnsson et al. 1994), give the
configuration factor in tables or graphs where different combinations of height, width
and distance has been used to calculate the configuration factor. Note that when
determining the configuration factor with this method, r must be at right angles to the
rectangle. Configuration factors are additive given that the configuration factors of
each contributory part are calculated from the same receiver, P, (Law 1963). Figure
2.4a-b shows the orientations of the variables used to determine the configuration
factor.
Hf
(a)
Figure 2.4
Wf
r
P
(b)
Determination of configuration factors
In Figure 2.4b, the total configuration factor tot is the sum of the configuration
factors of each rectangle as shown in Equation 2.11.
tot = 1 + 2 + 3 + 4
[2.11]
The same method can also be applied when any of the rectangles should not be
included. In this case the configuration factor of this part should be subtracted from
the total configuration factor, which is shown in Law (1963) and Jnsson et al.
(1994).
When it comes to determining separation distances between buildings, critical values
of configuration factors that are widely used in Building Regulations and other
calculation methods are 0.15 for buildings where a low intensity fire can be expected
and 0.075 for higher intensity fires, (Law 1963). McGuire (1965) and NFPA 80A
(1996) use 0.035 for severe hazard levels (buildings with highly flammable linings
according to McGuire 1965), 0.07 for moderate hazard levels and 0.14 for light
hazard levels. As explained above, the configuration factor expresses the decrease in
radiation transfer between the radiator and receiver. Therefore, the critical
-13-
configuration factor for different building occupancies can be determined by dividing

the critical radiation intensity that would cause ignition of the receiving surface by the
maximum radiation intensity predicted in a fire in the relevant building. The
configuration factor then achieved is the maximum allowable in order for ignition not
to occur.
2.3.4 Fire Service Intervention
As reported earlier in chapter 2.2.2, the St. Lawrence Burns generated very high
radiation intensities after 16 minutes, i.e. much higher than what would be practical
for the purpose of determining separation distances. However, since the Fire Service
generally will respond within in this time, only the lower values measured before that
time were adopted in regulations. It is required by both the Canadian and English and
Welsh regulations that where Fire Service intervention cannot be guaranteed within
reasonable time, the separation distance should be increased.
2.4
Critical received radiation
Whether ignition of a building facade exposed to radiation will occur depends on the
ignition criteria and physical properties of the facade material and how long time it
will take for the material to ignite at that specific exposing radiation.
2.4.1 Ignition criteria
Ignition of combustible material due to radiative exposure can occur either
spontaneously or piloted, i.e. in presence of an igniting source such as a spark or a
flame that can ignite combustible volatiles given off by the exposed surface, (Law
1963). Much higher levels of radiation is required to cause spontaneous ignition and
Law (1963) states a value of approximately 33 kW/m2 (0.8 cal/cm2s) for spontaneous
ignition of wood. However, since igniting sources will be present in a fire situation,
the value for piloted ignition must be used for the purpose of building separation
design. A value of 12.5 kW/m2 (0.3 cal/cm2s) is used in most Building Codes and
calculation methods as the maximum tolerable level of radiation at the exposed
facade, (NFPA 80A 1996). This value is used as the lowest value at which piloted
ignition of dry wood can occur and has been derived by the Joint Fire Research
Organization in the United Kingdom. According to McGuire (1965), unfinished and
untreated fibre board can ignite at lower radiation intensities, but no account needs to
be taken to this since such material is unlikely to be a material located in an area
where it would be exposed to radiation from a fire in an adjoining building.
Clarke (1998) has reviewed a number of experiments regarding ignition of solid
materials due to radiant heating. He reports on work completed by Law and Simms in
1977 where they investigated the effect of moisture content in wood for piloted and
spontaneous ignition. The conclusion from the work was that higher moisture content
in wood resulted in increased minimum ignition radiation and time to ignition, which
is also shown in Law (1968). Clarke (1998) also reports on work done by Janssens in
1991 where the critical radiant heat flux was established for different oven dried
timber species. The critical radiation was found to vary between approximately 10-14
kW/m2.
-14-
In a series of tests conducted in Sweden and Finland in the 1970s (Nordiska

industrigruppen, 1975), the minimum radiation intensity causing piloted ignition was
determined. The critical radiation was determined both in full scale fire tests and in
laboratory tests for painted and unpainted wooden walls. One conclusion that was
made based on the results from these tests was that higher levels of radiation was
required in order to cause ignition in the full scale tests than in the smaller laboratory
tests. Table 2.3 sets out the minimum radiation intensity that caused piloted ignition
for the different tests and exposed surfaces.
Table 2.3
Critical levels of radiation from full scale and laboratory tests causing piloted
ignition of painted and unpainted wooden walls. Adapted from Nordiska
industrigruppen (1975).
Levels of radiation causing

piloted ignition (kW/m2)
Full scale test
Laboratory test
Unpainted surface
Painted surface
18-19
10
26-30
15
The small scale laboratory tests were based on 15 minutes continuous radiation
exposure.
The Swedish Building Regulation (Boverket 1995) uses a value of 15 kW/m2 at where
piloted ignition can occur.
2.4.2 Time to ignition
A surface will not ignite immediately when being exposed to the critical levels of
radiation presented in the previous section. The time to ignition depends on factors
such as moisture content and thermal inertia (kc) of the exposed material, (Drysdale
1985). Law (1968) states that it would take approximately ten minutes for oven dried
wood ( = 500 kg/m3) to ignite when exposed to a radiation of 15.8 kW/m2, while it
would take approximately 65 minutes for wood ( = 800 kg/m3) with a moisture
content of 15 % (by weight of dry wood) to ignite when exposed to the same levels of
radiation.
In the same review by Clarke (1998) as was mentioned above, a model by Janssens
(1991) to predict time to ignition for piloted ignition of timber is presented. Equation
2.12 explains the model.
kc
q = qcr 1 + 0.73 2
hig t ig
where:
q
qcr
kc
hig2
tig
=
=
=
=
=
0.547
[2.12]
exposing radiation [kW/m2]

critical radiation [kW/m2]
thermal inertia [kW2s/m4K2]
heat transfer coefficient [W/m2K]
time to ignition [s]
Janssens has also obtained values of these properties for different oven dry timbers
which is shown in Table 2.4.
-15-

Table 2.4
Species
Western Red Cedar
Redwood
Radiata Pine
Douglas Fir
Victorian Ash
Blackbutt
Properties of various timber species obtained with Janssens model.

Adapted from Clarke (1998).
Tig [C]
354
364
349
350
311
300
qcr [kW/m2]
13.3
14.0
12.9
13.0
10.4
9.7
hig [W/m2K]
34.9
35.9
34.6
34.6
31.5
30.6
k
c [kJ2s/m4K2]
0.087
0.141
0.156
0.158
0.260
0.393
The report Standard Test Method for Determining Material Ignition and Flame
Spread Properties (ASTM 1990) sets out a method of how to theoretically determine
the time to ignition of a surface exposed to radiation. The surface is assumed to be a
semi-infinite slab. The outcome of the method is explained in Equation 2.13.
q o,ig
= 1 exp( )erfc
q e
where:
( )]
[2.13]
h 2 t ig
q o,ig
kc
critical flux for ignition [kW/m2]
q e
h
t ig
=
=
=
measured incident flux [kW/m2]

heat loss coefficient [kW/m2K]
ignition time under incident flux [s]
kc
thermal heating property (thermal inertia) [kW2s/m4K2]
-16-
Building Codes
3 BUILDING CODES
This chapter will describe and discuss what is regulated in the Building Codes of a
number of countries with regards to external fire spread between buildings. The
countries that will be discussed are; Australia, Canada, England and Wales, New
Zealand, Sweden and the United States of America. A brief comparison between the
different codes will also be made in chapter 7.
3.1
Australia
In order to comply with the Building Code of Australia (BCA 1996), a building
solution must satisfy the performance requirements. Satisfaction of the performance
requirements can be achieved in either of three ways;
1.
2.
3.
complying with the Deemed-to-Satisfy Provisions; or

formulating an alternative solution which(i)
complies with the Performance Requirements; or
(ii) is shown to be at least equivalent to the Deemed-to-Satisfy
Provisions; or
a combination of (1) and (2).
3.1.1 Deemed-to-Satisfy Provisions

If a building design fulfils the Deemed-to-Satisfy Provisions, it complies with the
Performance Requirements.
3.1.2 Alternative Solution
An Alternative Solution has to be judged in accordance with at least one of the
Assessment Methods described below. The Assessment Method used must show that
compliance with the Performance Requirements has been achieved and thus also
compliance with the BCA.
One or more Assessment Methods can be used to determine whether a building design
agrees with the Performance Requirements or not. The Assessment Methods that can
be used are briefly described below.
1. Prove that the design, type of construction and the materials used comply
with the Performance Requirements or Deemed-to-Satisfy Provisions.
2. Verification Methods either set out in the BCA or accepted by the
appropriate authority.
3. Comparison with the deemed-to-satisfy provision.
4. Expert judgement.
3.1.3 Objective
The objective of the BCA with regards to fire safety and fire resistance is to protect
occupants from injuries which may arise due to a fire within a building and while the
occupants are evacuating the building. Fire fighting and rescuing activities should also
be allowed for. Furthermore, fire spread between adjoining buildings must be
prevented and other properties protected from structural damage caused by structural
failure due to a fire in a neighbouring building.
-17-
Building Codes
3.1.4 Performance Requirements

A building must be constructed in a way that guarantees a maintained structural
stability while exposed to a fire. Spread of fire must also be avoided, both within the
building and to adjoining buildings. In order to comply with these performance
requirements, several factors have to be taken into account. Examples are:
use and layout of building,

fire load,
estimated extent and severity of fire,
fire safety measures,
evacuation time, and
intervention of rescue personnel.
3.1.5 Verification Methods

To comply with the performance requirement regarding prevention of fire spread
between buildings, the BCA sets out two tables with distances between buildings or to
the boundary and the maximum radiation allowed at that location.
Table 3.1 (adapted from Table CV1 in the BCA) should be used when determining
the distance between buildings located on different properties. The table should be
read as; a building will not cause a heat flux greater than what is stated by in Table
3.1 at the belonging distances inside the boundary of an adjoining property. The table
can also be reversed and be read as; a building located at distances within the property
boundary set out in the table will be able to withstand the stated radiation levels.
Table 3.1
Required distances and maximum radiative heat flux between buildings on different
properties. Adapted from the BCA (1996)
Location
On boundary
1 m from boundary
3 m from boundary
6 m from boundary
Allowable radiation (kW/m2)

80
40
20
10
Table 3.2 (adapted from Table CV2 in the BCA) should be used for buildings located
on the same property. Fire spread between buildings will be prevented if the building
will not cause a greater radiant heat flux onto the other building than given by Table
3.2 for belonging values of the distance between the buildings. It can also be
expressed as; when the distance set out in the table separates the buildings, the
buildings are able to withstand the belonging heat flux set out in the table.
Table 3.2
Required distances and maximum radiative heat flux between buildings on the same
allotment. Adapted from the BCA (1996).
Distance between buildings

0m
2m
6m
12 m
Allowable radiation (kW/m2)

80
40
20
10
-18-
Building Codes
Table 3.2 is equivalent to Table 3.1 if it is assumed that the boundary is located in the
middle of the separating distance.
As explained previously in this chapter, the designer can chose to use Table 3.1 and
3.2 to make sure that the design complies with the performance requirements. The
deemed-to-satisfy provisions can also be used, as well as other methods of verifying
that performance requirements with regards to building separation are satisfied.
The BCA does not give an absolute figure of at which radiant heat flux ignition will
occur. The values set out in the BCA ranges between 10-35 kW/m2, depending on
type of material and whether piloted ignition is likely to occur or not. However, no
background information is provided from where these values and how Tables 3.1 and
3.2 were obtained.
3.1.6 Summary
The BCA relies on two tables to verify that the performance requirements are met. In
general, a building should not be able to cause a radiant heat flux in excess of 80
kW/m2 at the boundary and a building should be able to withstand a radiation varying
between 10-80 kw/m2, depending on distance to the boundary or neighbouring
building. The designer can use performance based engineering methods to show that
the performance requirements are fulfilled. The objective of the BCA with regards to
external fire spread is to prevent fire spread between buildings, i.e. to protect both
buildings independent of which is the fire source.
3.2
Canada
Unfortunately, a current version the National Building Code of Canada has not been
available to the author. This description will therefore be based on what is written in
other reports about the National Building Code of Canada and only point out the most
relevant knowledge with regards to building separation.
The National Building Code of Canada sets out tables of building separation that are
based on the same criteria of critical received radiation, i.e. 12.5 kW/m2, as the
Approved Document B (1991), which is the document used in England and Wales,
(Clarke 1998). Canada uses higher values of emitted radiation and a flame projection
distance from openings of 1.2 m.
The used values of emitted radiation was obtained from the St. Lawrence Burns
reported on in chapter 2.2.1. The Canadian code uses configuration factors of 0.07 for
normal buildings and 0.035 for buildings with combustible linings, which is expected
to burn extra vigorously. These configuration factors are the same as those set out by
McGuire (1965) and results in expected levels of radiation of 180 and 360 kW/m2
respectively. The St. Lawrence Burns showed extraordinary high levels of radiation
after 16 minutes, much higher than would practical to use when determining building
separation. The National Building Code of Canada therefore requires that the
separation distance should be doubled in areas where Fire Service intervention cannot
be guaranteed within 10 minutes.
-19-
Building Codes
3.3
England & Wales
3.3.1 Requirement
The requirements in England and Wales regarding external fire spread are set out in
the Building Regulations 1991, Part B of Schedule 1. The requirements are given in
the Approved Document B and are set out below.
(1) The external walls of the building shall resist the spread of fire over the
walls and from one building to another, having regard to the height, use
and position of the building.
(2) The roof of the building shall resist the spread of fire over the roof and
from one building to another, having regard to the use and position of
the building.
3.3.2 Performance
In order to meet with the B4 requirements and to prevent a fire from spreading
between buildings on different sides of the relevant boundary, the Approved
Document B requires that the following measures have to be taken:
1) The risk of ignition of external walls caused by an external fire source
and fire spread over the wall surfaces must be limited. This is achieved
by making provisions for the external walls to be built of material with
low heat release rates.
2) The amount of radiant heat flux that is able to pass through an external
wall must be limited. This is achieved by limiting the amount of
unprotected openings in the wall and by taking into account for the
distance to the relevant boundary.
3) The risk of flame spread over and/or penetration of the roof, caused by an
external fire source, must be limited. This is achieved with an appropriate
roof construction.
The extents to which these measures have to be taken are dependent on factors such
as use of building, distance to the relevant boundary and height of building.
According to the Approved Document B, fire spread between buildings and the
expected consequences if it does occur, are dependent on the fire severity, distance
between buildings, fire resistance of external walls and the risk that occupants in the
adjoining building are exposed to.
3.3.3 Space separation
The provisions in the Approved Document B4 regarding space separation are based
on six main assumptions. The assumptions are listed below.
(a) The fire size is dependent on the fire compartmentation in the building.
(b) The fire intensity is dependent on the use of the building, i.e. purpose
groups. An automatic sprinkler system can also decrease the fire
intensity.
(c) Residential, Assembly and Recreation purpose groups are associated by a
greater risk to life than other purpose groups.
-20-
Building Codes
(d) Fire spread between buildings located on the same property do not
represent a great risk to life and can be disregarded, unless the building
contains Residential, Assembly or Recreation purpose groups.
(e) Another building with a similar height to the building of concern is
located on the other side of the relevant boundary. The buildings are
located at equal distance from the common boundary.
(f) The radiant heat flux that passes through a fire resistant wall is negligible
and can be disregarded.
Smaller fire compartments are recommended when a shorter separation distance or an
increased amount of unprotected wall area is wanted.
Boundaries
The Approved Document B4 sets out the distance to the relevant boundary, i.e. actual
boundary or an assumed boundary located in the middle in the space between two
buildings, should be used when determining the separation distance. This enable the
designer to calculate the total amount of unprotected area in the external wall without
having to take into account to any building located on the other side of the boundary.
An external wall is counted as facing the relevant boundary if any of the three cases
shown in Figure 3.1 is fulfilled. In order to be considered as a relevant boundary, the
boundary should:
1. coincide with,
2. be parallel to, or
3. not form an angle of more than 80 with the external wall of the building.
<80
2.
1.
Figure 3.1
3.
External walls facing the relevant boundary
The separation distance between buildings located on the same site is often decreased,
unless the buildings of concern are the Residential, Assembly and recreation purpose
groups. If the building is classified in any of these purpose groups, a notional
boundary positioned in the space between the buildings should be assumed. The
notional boundary should be located in a position that complies with the provisions
for space separation for one of the buildings. The other building can then be used to
certify that its location also complies with the provisions.
The relevant boundary is the boundary that a wall faces and can be either a notional
boundary or the actual property boundary.
-21-
Building Codes
Unprotected areas
An area in the external wall with less fire resistance than what is required for the wall
shall be regarded as an unprotected area. A fire resistant wall that is covered with
combustible material with a thickness greater than 1 mm shall also be treated as an
unprotected area with a total area of half the actual area of the combustible material.
Small unprotected areas, i.e. < 1.0 m2, can be neglected when determining the
separation distance since the risk of fire spread caused by these small areas are very
small. This assumption is valid when the separation distances of the unprotected areas
in an external wall are;
unprotected areas forming an area not more than 1 m2 a minimum

distance of 4.0 m to other unprotected areas of the same size is required,
unprotected areas not greater than 0.1 m2 a minimum distance of 1.5 m
to other unprotected areas is required,
unprotected areas positioned in different fire compartments no
restrictions, and
unprotected external areas of a stairway forming a protected shaft no
restrictions.
The unprotected areas of an uncompartmented building that are more than 30 m above
ground level may be neglected with regards to separation distance. External walls
located within 1.0 m of the relevant boundary must not have unprotected areas in
excess of what is described above and the walls must be fire resistant on both sides.
3.3.4 Calculation methods
The Approved Document B4 sets out two methods of calculating the acceptable
unprotected area in an external wall. The methods are obtained from the Fire Research
Technical Paper No. 5, 1963 and are useable for buildings located more than 1.0 m
from the relevant boundary. The objective of the calculation methods is to make sure
that the building is separated from the relevant boundary by at least half the distance
at which the total radiant heat flux received from all unprotected areas in the external
wall would be 12.6 kW/m2. This is based on the assumption that the emitted radiation
from the unprotected areas in the wall is as set out below.
1. 84 kW/m2 for buildings in the Residential, Office, Assembly and
Recreation purpose groups.
2. 168 kW/m2 for buildings in the Commercial, Industrial, Storage or Other
non-residential purpose groups.
The method is also called the mirror image concept. The concept of the mirror
image is shown in Figure 3.2.
-22-
Building Codes
D/2
D/2
84 kW/m2
168 kW/m2
Figure 3.2
12.6 kW/m2
The mirror image concept
With a properly designed sprinkler system installed in the building, the separation
distance D/2 from the relevant boundary may be halved. However, the distance to the
boundary is not allowed to be less than 1.0 m. Other calculation methods may be used
instead of the two methods set out above. The alternative methods are explained in
External Fire Spread: Building separation and boundary distances, (Fire Research
Station 1991).
Method 1
Method 1 should be used for dwelling houses, flats and other residential buildings.
Furthermore, the external walls should not be longer than 24 m and the building
height no more than three levels. The minimum distance from the relevant boundary
to the sides of the building and the maximum acceptable unprotected areas are shown
in Table 3.3. Figure 3.3 shows the principles of Method 1 and the distance to the
relevant boundary. Areas of the external wall that are in excess of the values given by
Table 3.3 should be fire resisting.
Table 3.3
Separation distances and maximum unprotected areas.

Adapted from Approved Document B, (1992).
Minimum distance X Maximum acceptable

between external wall and unprotected area (m2)
relevant boundary (m)
1
5.6
2
12
3
18
4
24
5
30
6
No limit
-23-
Building Codes
24m maximum
Boundary
Figure 3.3
Principles of Method 1. Adapted from Approved Document B.
Method 2
Method 2 can be used for any building, regardless of the purpose group of the
building. However, buildings should not be higher than 10 m, except for open-sided
car parks. The distance from the relevant boundary to the side of the building and the
amount of acceptable unprotected areas are set out in Table 3.4. Areas of the external
wall that are in excess of the values given by Table 3.4 should be fire resisting.
Table 3.4
Separation distance and maximum acceptable unprotected areas.

Adapted from Approved Document B
Separation distance between external Maximum unprotected area as a

wall and relevant boundary (m)
percentage of total wall area (%)
Purpose groups
Residential, Office,
Shop &
Assembly and
Commercial,
Recreation
Industrial, storage
& other nonresidential
1
4
1.0
2
8
2.5
5
20
5.0
10
40
7.5
15
60
10.0
20
80
12.5
25
100
If an automatic sprinkler system is installed in the building, the separation distance
may be halved, provided that the distance does not become less than 1.0 m. For
buildings with a height in excess of 10 m, the calculation methods described in
External Fire Spread: Building separation and boundary distances, (Fire Research
station 1991) can be used.
-24-
Building Codes
External walls
The necessary degree of fire resistance of the external walls is dependent on usage,
height and size of the building. A wall located more than 1.0 m from the relevant
boundary is allowed to have a lower level of fire resistance. In order to limit an
external walls ability to ignite when exposed to an external fire source and to prevent
fire from spreading upwards the external wall, provisions are made in the Approved
Document B4 to limit the combustibility of the external wall. These provisions
concern buildings located closer to the relevant boundary than 1.0 m, and buildings in
the Assembly and Recreation purpose groups.
3.3.5 Summary
Approved Document B uses a radiation intensity of 12.6 kW/m2 as the critical value
of when ignition may occur. The emitted radiation from a building during a fire is
assumed to be either 84 or 168 kW/m2, depending on the building purpose group, so it
is not necessary to look at the actual fire load. Two methods are presented of how to
determine the amount of acceptable unprotected areas of external walls. The building
separation distances are based on the mirror image concept, which means that
another building is located on the other side of the relevant boundary. The building of
concern should be separated from the relevant boundary by at least half the distance at
which the total radiant heat flux would be 12.6 kW/m2.
3.4
New Zealand
The New Zealand Building Code, has been performance based since 1992, i.e. the
clauses say what has to be achieved but not the exact way of how the objectives
should be fulfilled. As a complement to the Building Code, sets of Acceptable
Solutions have been developed. The Acceptable Solutions provide one mean of how
to meet with the performance requirements of the Building Code, (Building Industry
Authority 1995).
3.4.1 Objective
The objective of the New Zealand Building Code with respect to spread of fire can be
divided into four major areas. The objectives are summarised below.
1. Protect occupants from injuries during evacuating a building on fire.
2. Allow a safe enough environment for fire fighting and rescuing activities
within a building during fire.
3. Prevent adjoining properties and neighbouring household units from
being affected by a fire.
4. Protect noxious effects to the environment due to fire in a building.
3.4.2 Functional Requirements
The functional requirements regulate that buildings shall be constructed with fire
safety measures that counteract the spread of fire. A building shall in a fire situation
provide:
1) enough time for occupants to evacuate the building without being
injured,
-25-
Building Codes
2) an environment that enables fire fighters to perform rescuing operations

and protecting of property,
3) damage protection of adjoining properties and neighbouring household
units, and
4) means of limiting the amount of hazardous substances released into the
environment.
3.4.3 Performance
The performance provisions set out how the functional requirements shall be fulfilled
and the objectives achieved. In order to prevent spread of fire and smoke to adjacent
buildings, a building shall be provided with fire compartmentation within the
building. The estimated fire load within a building and the spatial distance to other
buildings shall determine the fire resistance of external walls and roofs.
According to the Acceptable Solution, external walls and roofs must be constructed in
a way to prevent horizontal fire spread due to radiation and building collapse to other
properties, neighbouring buildings with sleeping occupancies and to external safe
paths. Measures achieving the required protection are by providing enough separation
between buildings, fire resisting ratings of building elements and using less
combustible finishes. Other measures are limiting the amount of unprotected areas in
external walls and installing a sprinkler system. The requirements of external walls
and roofs are adapted from the Building Regulations 1985 (England and Wales) and
the Approved Document B4 External Fire Spread 1990.
3.4.4 Building Separation
According to Appendix A of the Acceptable Solution, buildings can be divided into
purpose groups with a specific fire hazard category. Appropriate purpose groups can
be determined based on the types of activities carried out within in the building. The
fire hazard categories set out a range of values of fire load energy densities (FLED)
and a fixed value that can be used for design purposes. Table 3.5 below shows fire
hazard categories and fire load energy densities.
Table 3.5
Fire hazard categories and fire load energy densities of buildings.

Adapted from Approved Document C3: Spread of Fire
Fire Hazard Category

1
2
3
4
Fire Load Energy

Density (MJ/m2)
0-500
501-1000
1001-1500
>1500
Design value of Fire Load

Energy Density (MJ/m2)
400
800
1200
-
Appendix C of the Acceptable Solution sets out five methods of determining the
separation between buildings. They will be discussed in detail in Chapter 4. The
separation is dependent on factors such as unprotected areas in external walls, purpose
groups in the building of concern as well as in adjoining buildings, size of fire
compartments and whether the building is sprinkler protected or not. The methods can
also be reversed and used to determine the maximum allowed unprotected areas in an
external wall, given that the separation distance is known from the beginning.
-26-
Building Codes
In order to prevent fire spread between buildings, buildings should be separated from
the relevant boundary by at least half the distance at which the total radiant heat flux
would be 12.6 kW/m2. It is assumed that the emitted heat flux from a building during
a fire will be:
a) 84 kW/m2 for crowd or office purpose groups, or
b) 168 kW/m2 for sleeping, commercial and industrial purpose groups.
It is further assumed that the emitted radiation can be reduced if the building is fitted
with an automatic sprinkler system. In this case, the unprotected areas may be
doubled or the separation distance halved but still not less than 1.0 m.
As said above, the five methods of calculating the separation distance and the amount
of acceptable unprotected areas will be reviewed later in this report. They are obtained
from BRE report External fire spread: Building separation and boundary distances
(Fire Research Station 1991), which is based on Fire Research Technical Paper No. 5
Heat radiation from fires and building separation (Law 1963). A brief summary of
the methods is presented below.
1)
Walls closer to the relevant boundary than 1.0 m.

External walls must have a fire resisting rating according to the Acceptable
Solution and unprotected areas can range between 0.1-1 m2. Unprotected
areas in the walls of the same fire compartment have to be separated from
each other by 1.5 or 4.0 m depending on the area. Furthermore, the
unprotected areas must have a fire resistance rating for integrity of at least 30
minutes.
2)
Walls located more than 1.0 m from the relevant boundary.

This method is valid for occupancies containing attached and multi-unit
residential dwellings and buildings with temporary accommodation (purpose
groups SA and SR). The external walls must not be longer than 24 m and the
building height not exceed 7.0 m. Table 3.6 (Table C1 in Appendix C of the
Acceptable Solution) below sets out distances to the boundary and the
maximum accepted unprotected area.
Table 3.6
Building separation and maximum unprotected areas for unsprinklered multi-unit

residential dwellings and buildings with temporary accommodation.
Adapted from Appendix C, Acceptable Solution.
Minimum distance between external Maximum acceptable

unprotected area (m2)
1
5.6
2
12
3
18
4
24
5
30
6
No limit
3)
External walls more than 1.0 m from the boundary. This method is valid
for buildings no higher than 7.0 m and with fire hazard category of 1.
-27-
Building Codes
Minimum separation distances and acceptable unprotected areas are shown

in Table 3.7 (Table C2 in Appendix C of the Acceptable Solution).
Table 3.7
Building separation and maximum unprotected areas as a percentage of total wall

area for unsprinklered buildings with a fire hazard category of 1.
Adapted from Appendix C, Acceptable Solution
Separation distance between external Maximum unprotected area as a

percentage of total wall area (%)
Purpose groups
Other
Sleeping activities
1
4
1.0
2
8
2.5
5
20
5.0
10
40
7.5
15
60
10.0
20
80
12.5
25
100
4)
Enclosing rectangles. This method provides means of calculating the

maximum unprotected area allowed as a percentage of the external wall of a
fire compartment. The calculated value is then used together with the
purpose group of the building to determine the minimum required distance to
the relevant boundary, which is set out in a number of different tables
(Tables C3 in Appendix C of the Acceptable Solution).
5)
Aggregate notional areas. This method is used to calculate the effect on the
relevant boundary of unprotected areas in external walls. Different assumed
reference points on the boundary are to be investigated and the one resulting
in the worst case scenario to be chosen for design purposes. For each of these
reference points, the distance to and the size of unprotected openings are
determined. The unprotected area is then multiplied by a factor, which is
dependent on the horizontal distance and is set out in Table 3.8 (Table C4 in
Appendix C of the Acceptable Solution). The notional areas of all
unprotected areas facing the relevant boundary in the fire compartment are
summarised and compared with the given criteria. In order to comply with
the criteria, the aggregate notional area for each fire compartment must not
exceed either 100 or 200 m2 depending on purpose group and extent of the
fire compartment. The procedure has to be redone until a separation distance
resulting in a satisfactory aggregate notional area.
-28-
Building Codes
Table 3.8
Multiplication factor for various distances between unprotected

areas and reference points on the relevant boundary.
Adapted from Appendix C, Acceptable Solution.
Horizontal distance between

unprotected area and
reference point, (m)
1.0 - 1.2
1.2 - 1.8
1.8 - 2.7
2.7 - 4.3
4.3 - 6.0
6.0 - 8.5
8.5 - 12.0
12.0 - 18.5
18.5 - 27.5
27.5 - 50.0
>50
Multiplication factor
80.0
40.0
20.0
10.0
4.0
2.0
1.0
0.5
0.25
0.1
0.0
3.4.5 Summary
The New Zealand Building Code and Acceptable Solution uses a radiation intensity of
12.6 kW/m2 as the critical value of when ignition may occur. The emitted radiation
from a building during a fire is assumed to be either 84 or 168 kW/m2, depending on
the type of occupancy. Five methods are presented of how to calculate minimum
separation distances between buildings and maximum acceptable unprotected areas of
external walls. In general, buildings should be separated from the relevant boundary
by half the distance at which the total radiant heat flux would be 12.6 kW/m2. This
principle is sometimes called the mirror image concept. The objective of the New
Zealand Building Code with regards to external fire spread is to protect adjacent
buildings and other properties from being affected by a fire in the own building.
Nothing is said about preventing the owners building from being affected by a fire in
an adjoining building.
3.5
Sweden
The Swedish design guide contains regulations and general advice of how compliance
with the Building Code and other essential directions that has to be followed in the
building process is achieved (Boverket 1995).
3.5.1 Regulation
The regulation is applicable in either of the following cases:
construction of new buildings,

renovation and addition of existing buildings,
ground and demolishing work, and
properties that are claimed to be built on.
-29-
Building Codes
3.5.2 Advice
The regulations are complemented with a set of general guidelines on how the
requirements of the regulation can be achieved. The advice states how a designer
should or could act in order to make sure that the regulations are followed. However,
each designer can chose to use other methods and solutions than suggested by the
advice, as long as it is done in accordance with the regulations.
3.5.3 Objective
According to the Swedish design guide, fire spread between buildings should be
prevented by limiting the radiant heat flux either emitted by a burning building or
received by an adjacent building. This can be achieved in a number of ways.
Providing a large enough distance between buildings.

Limiting the size of unprotected building sections.
Limiting the heat content of facade materials.
Limiting the extent of the fire and thereby limit the radiant heat flux
emitted from a fire. This can be done by e.g. smoke ventilation or an
automatic sprinkler system.
3.5.4 Requirements
The main requirement with regards to separation between buildings is that a building
should not be closer to the boundary than 4.0 m. If the building is located closer to the
boundary than 4.0 m, the neighbouring building has to be located at a position that
ensures a minimum separation of 8.0 m. In cases where a total separation distance of
8.0 m is not achieved, the building has to be constructed in a way that limits the risk
of fire spread to adjacent buildings.
It is also said that fire spread should be made more difficult by limiting the radiant
heat flux and preventing flames from affecting neighbouring buildings. This is
achieved by providing a safe distance between the buildings, by a fire partition or a
combination of both methods. Buildings located at the boundary must have a firewall,
e.g. fire partition, facing the neighbouring property. The class of the firewall is
dependent on the building classification.
The advice says that radiation received by neighbouring buildings should not exceed
15 kW/m2 during a period of 30 minutes. Buildings with more than two levels are
preferable constructed with a firewall facing the boundary. No background is given
for where these figures are obtained from or what they are based on.
Extraordinary fire safety measures could be demanded by the Fire Service where
evacuation and prevention of fire spread are dependent on Fire Service intervention
and where fire fighting activities can not be guaranteed to begin within reasonable
time.
3.5.5 Performance based design
The fire safety design may be performed in a different way than what is regulated by
the code, provided that an investigation based on performance based engineering
methods proves that the design is as good as if all demands in the regulation are
fulfilled.
-30-
Building Codes
Residential dwellings in small houses should be separated to prevent fire spread for at
least 60 minutes. Large buildings should be divided into compartments of appropriate
sizes by firewalls. This would enable intervention by the fire service and thereby
preventing or make it more difficult for the fire to spread to adjoining buildings.
Factors that have to be considered are e.g. spatial distance, fire load, smoke
ventilation, automatic fire alarm and automatic extinguishing system.
A firewall should contain a fire without any intervention of the fire service for a
certain period of time. The wall should also be provided with enough stability to
withstand any mechanical influences caused by a fire. The construction of a firewall is
dependent on building type and estimated fire load in the building. Table 3.9 outlines
the performance requirements of firewalls.
Table 3.9
Performance requirements of firewalls. Adapted from Boverket (1995).
Building class
1. Br1
2. Br2 and Br3
Fire class of wall at different fire load,

f (MJ/m2)
f 200
F 400
f > 400
REI-M90 REI-M120
REI-M240
REI-M60 REI-M90
REI-M120
REI-M90 means that the wall has load bearing capacity (R), integrity (E), isolation (I)
and can withstand mechanical impact (M) in a fire for 90 minutes. Buildings can be
divided into classes according to the following rules:
Br1-
Buildings associated with a large risk of injuries to occupants in a fire

situation, e.g. buildings with
three or more levels,
sleeping occupants without good knowledge of the building,
occupants without ability to evacuate on their own, and
an assembly on the second level.
Br2-
Building with moderate risk for injuries to occupants in a fire situation, e.g.
buildings with
more than two dwellings and where living area or working room is
located in the attic,
an assembly on the ground level, and
2
compartments larger than 200m .
Br3-
Other buildings than listed above.
3.5.6 Summary
In Sweden, buildings have to be located at least 4.0 m from the boundary or at least
8.0 m from any building at the neighbouring property. If a building is closer to the
boundary or another building than stated above, it has to be shown with best
engineering principles that fire spread between the buildings will not occur. A
building should be able to withstand a radiant heat flux of 15 kW/m2 during a time
period of 30 minutes.
-31-
Building Codes
3.6
The United States of America
There is no single Building Code that is valid for the whole country and different
Building Codes can be used in different states. The three most commonly used
Building Codes are; the National Building Code (BOCA 1996), the Uniform Building
Code (1994) and the Standard Building Code (1997). They are strictly prescriptive
codes with no allowance for performance based design.
One document that is used in the US to determine separation distances between
buildings is the NFPA 80A Recommended Practice for Protection of Buildings from
Exterior Fire Exposures. This method will be further discussed in chapter 4.
-32-
Calculation methods
4 CALCULATION METHODS
This chapter will describe six methods of determining separation distances between
buildings and the maximum allowable amount of unprotected areas. Some of them are
referred to by Building Regulations, and others are set out in the fire engineering
literature. The description of the methods presented in this report should not be used
as a complete design guide. If anyone wants to use the methods, the original reports
should be consulted.
4.1
Enclosing rectangles
This method called the Enclosing rectangle method is set out in the BRE report
External fire spread: building separation and boundary distance (Fire Research
Station 1991). The method is derived from work performed by Law (1963) and is
incorporated in the Approved Document B2/3/4 (HMSO, 1985). The method
determine the boundary distance based on rectangles that enclose unprotected areas in
the facade, see Figure 4.1. Tables of boundary distances for different types of
buildings and for different dimensions of the buildings are set out, which enable a
person not familiar with the fire engineering knowledge to perform the design.
Figure 4.1
Enclosing rectangles
Step 1.
Identification of unprotected areas

The Building Regulations of the country of concern should be
consulted with regards to which areas in the facade that should be
regarded as unprotected. This may vary between different countries
and different regulations.
Step 2.
Establishing of a plane of reference

A plane of reference should be established so that the plane complies
with or touches the side of the building. Furthermore, it should not pass
through any parts of the building (except for balconies and coping), nor
the relevant boundary. If any part of the facade is set back no more
than 1.5 m, this area may be disregarded and the plane of reference
determined as explained above. Those unprotected areas that are at an
angle of 80 or less to the plane of reference should be regarded in the
design process.
-33-
Calculation methods
Step 3.
Determination of the enclosing rectangle

The enclosing rectangle should be chosen as the smallest rectangle
with dimensions as set out in Table 1 in the original document, (Fire
Research Station 1991) that encloses all unprotected areas projected
onto the plane of reference. The extent of the enclosing rectangle
depends on factors such as the horizontal and vertical dimensions of
the fire compartmentation, type of cladding fitted on the exterior wall
and if the wall has an approved fire resistance rating. The unprotected
areas should be expressed as a percentage of the area of the enclosing
rectangle.
Step 4.
Determination of the boundary distance or maximum unprotected

areas
Once the width, height and percentage of unprotected areas of the
enclosing rectangle are known, Table 1 in the original document may
be used to determine the minimum required distance between the plane
of reference and the relevant boundary. The separation distance is also
dependent on the fire load density within the building. The fire load
density in Table 1 either classed as normal or low depending on for
what purpose(s) the building is used. If the separation distance already
is given, Table 1 can also be used to determine the maximum allowed
unprotected areas. This latter process may require extraordinary work
in order to determine an appropriate enclosing rectangle.
It may be needed to investigate more than one enclosing rectangle and
the largest separation distance of these rectangles should be chosen.
The process should be repeated for all sides of the building.
Step 5.
Consideration of areas of special exposure

In some situations, unprotected areas may be concentrated to certain
parts of the building, thereby representing a greater local hazard than if
they were evenly distributed throughout the building. This area should
be considered separately and it may be found that a greater separation
distance is required.
The method explained above may also be used where groups of
unprotected areas are widely spaced from each other. If the groups of
unprotected areas are spaced from each other by more than four times
the boundary distance determined by Step 1-4, they should the be
treated separately since the amount of received radiation from the other
group may be neglected.
Where a part of the building is set back more than 1.5 m, a reduction of
the boundary distance is allowed. This is accomplished by creating an
equivalent radiator positioned between the further most opening of
the set back and the edge of the original facade and then determining
the boundary distance in relation to this equivalent radiator.
In a recess with unprotected areas on all three sides, the total area, i.e.
even the unprotected areas in the side of the recess, should be
-34-
Calculation methods
expressed as a percentage of the enclosing rectangle. If the recess only

contains unprotected areas in the rear wall, it may in some cases be
possible to reduce the boundary distance. A further description of how
this is done is explained in detail in the original report.
4.2
Aggregate notional areas
This method is also set out in the BRE report External fire spread: building
separation and boundary distance (Fire Research Station 1991). The method is
derived from work performed by Law (1963) and is incorporated in the Approved
Document B2/3/4 (HMSO, 1985).
In this method, a number of points on the relevant boundary are chosen and the
amount of unprotected areas that are visible from the point is calculated. These areas
are called effective or notional areas and are calculated by multiplying the actual
unprotected area by a distance dependent factor. The aggregate notional areas must be
less or equal to certain predetermined values in order to fulfil safe separation
requirements. This method, which also is called the Protractor method, may be
favourable for complicated building shapes and in combination with the Enclosing
rectangle method. In this method, a protractor in the same scale as the building and
with arcs representing fixed distances from the datum point of the protractor are
made. A datum line perpendicular to the base line of the protractor is also established.
The area between each arc represents a specific factor, which later is used to calculate
the aggregate notional areas.
Step 1.
Identification of unprotected areas

The Building Regulations of the country of concern should be
consulted with regards to which areas in the facade that should be
regarded as unprotected. This may vary between different countries
and different regulations.
Step 2.
Determination of points on the boundary to be tested

Points to be tested should be chosen along the relevant boundary at
3-m intervals. If a boundary distance has been determined with the
Enclosing rectangle method, the only points on the relevant boundary
that need to be tested are those that are inside the boundary set by the
Enclosing rectangle method.
Step 3.
Determination of which unprotected areas that may be ignored

A vertical datum of unlimited height should be set at the point to be
tested and a datum line drawn between this point and the nearest point
on the building. The protractor should be placed so that its datum point
and datum line coincide with the datum and datum line of the point on
the boundary to be tested. Unprotected areas that could be ignored are
those that cannot be seen from the vertical datum. These are areas:
positioned beyond the base line or the 50 m arc of the
protractor,
facing another direction (i.e. away from) than towards the
vertical datum,
-35-
Calculation methods
Step 4.
shielded by other parts of the building, and

making an angle of 10 or less with a line between the vertical
datum and the unprotected area of consideration.
Calculation of aggregate notional areas

To find the aggregate notional area, the actual unprotected area should
be multiplied by a factor for the particular zone in which the area falls.
The factor is given by the protractor as explained above. The
multiplication factor of the different arcs is given in Table 4.1. This
procedure should be made for all relevant unprotected areas for each
point and vertical datum of concern. The sum of all notional areas will
make the aggregate notional area. The aggregate notional area should
not exceed 210 m2 for Residential, Office and Assembly recreation
use, or 90 m2 for Shop and commercial, Industrial, Storage and Other
non-residential recreation use.
Table 4.1
Multiplication factor for various distances between unprotected areas and
reference points on the relevant boundary. Adapted from Fire Research Station (1991).
Horizontal distance between

unprotected area and
reference point, (m)
1.0 - 1.2
1.2 - 1.8
1.8 - 2.7
2.7 - 4.3
4.3 - 6.0
6.0 - 8.5
8.5 - 12.0
12.0 - 18.5
18.5 - 27.5
27.5 - 50.0
>50
4.3
Multiplication factor
80.0
40.0
20.0
10.0
4.0
2.0
1.0
0.5
0.25
0.1
0.0
Peter Collier
This method is presented in Collier (1996) and allows the user to specify the fire
intensity depending on the actual building and its properties. The method can mainly
be used for five situations.
(A) Calculate the required separation distance between buildings in order to
prevent fire spread.
(B) Calculate the required distance to a boundary in situations where
neighbouring buildings do not exist.
(C) Calculate the incident radiation received by neighbouring buildings.
(D) Calculate the maximum allowed area of unprotected openings in the
facade of the owners building.
(E) Check if fire spread can occur in an existing building situation.
The method is aimed to prevent fire spread between both the adjoining buildings; i.e.
it is necessary to calculate the radiation in both directions.
-36-
Calculation methods
The method is very user-friendly and allows the user to specify or calculate the
temperature in the fire compartment. Furthermore, the critical radiant flux for ignition
may be varied depending on the external cladding of the neighbouring building.
In situations where neighbouring buildings are not parallel to each other, the facades
of the buildings can be assumed to be parallel which will thereby create a
conservative separation distance. If a neighbouring building does not exist, the
required distance to the boundary should be taken as half the distance of what would
be required if the neighbouring building was a mirror image of the owners building.
It may be necessary to consider one or several openings separately if the openings are
not distributed evenly throughout the facade.
The calculation procedure of how to determine the required separation distance
between buildings is set out below. If the separation distance already is given, the
methodology can be reversed to determine the incident radiant flux or maximum
allowed unprotected openings.
Step 1.
Set up enclosing rectangles.

This should be based on the extent of the fire compartment and the
amount of unprotected openings. With irregular shaped facades the
unprotected areas are projected onto a plane of reference, which is
assumed to touch the outer parts of the building see Figure 4.2.
Plane of reference
Figure 4.2
Step 2.
Projection of openings onto a plane of reference
Calculate the area of the enclosing rectangle, Ae.

Ae = H W
where:
Step 3.
H = height of enclosing rectangle [m]

W = width of enclosing rectangle [m]
Calculate the width-to-height or height-to-width ratio, AR, of the

enclosing rectangle.
AR = H/W (or W/H)
Step 4.
[m2]
AR 1
Determine the radiation intensity, Is, in the fire compartment.

-37-
Calculation methods
A fire temperature of 1000C, which represents a radiation intensity of

149 kW/m2, can be used for most cases. If a lower or higher
temperature and radiation intensity can be expected, appropriate
chosen or calculated values may be used instead.
The radiation intensity can be calculated with the following equation:
Is = 5.6710-11 T4
where:
[kW/m2]
T = fire temperature [K]
The fire temperature can be determined by using the ISO 834 standard
fire curve and the fire resistance rating of the fire compartment.
T = 345 log10 (8t+1) [C]
where:
Step 5.
t = fire resistance time [min]
The whole facade of the owners building should be counted as the

radiating source if the external wall is not fire resistance rated. Where
the wall is fire resistance rated the radiation intensity may be reduced
by a reduction factor, Rf.
Rf = Ao / Ae
where:
Step 6.
Ao = unprotected areas [m2]

Ae = area of enclosing rectangle [m2]
Calculate the emitted radiant flux, Ie. If fire resistance glazing are fitted
in the unprotected openings of the owners building, the emitted
radiation may be reduced by an additional 50%.
Ie = Is Rf
Step 7.
[kW/m2]
Determine at which received radiant flux the neighbouring building

will ignite. This value may vary depending on external cladding and if
the openings are fitted with fire resistant glazing. In the latter situation
the incident radiation that passes through windows may be reduced by
50%. Non fire-rated glazing is assumed to break when exposed to high
levels of radiation. Examples on values of critical incident radiation,
Icr, received by the neighbouring building are set out in Table 4.2.
-38-
Calculation methods
Table 4.2
Maximum permitted radiation at the facade of the neighbouring building.

Adapted from Collier (1996).
Critical incident radiation, Icr [kW/m2]

External cladding
No
Usage of non fire- Usage of fire
openings
rated glazing
rated glazing
(1)
EIFS
9.0
9.0
9.0
Timber
12.5
12.5
12.5
Fibre-cement board
25.0
12.5
25.0
Non-combustible(2)
No limit
12.5
50.0
(1)
Exterior insulation and finishing system.
(2)
Concrete, brick, steel or aluminium.
Step 8.
Calculate the acceptable configuration factor, .

= Icr / Ie
Step 9.
Calculate the separation distance, S, between the buildings.

S=R+P
Where:
[m]
R = radiation distance obtained from Figure 4.3 [m]

P = projection distance [m]
The projection distance can be set to 2 m if the windows in the owners

building are fitted with non fire-rated glazing and 0 m if fire-rated
glazing is used. Values other than 2 m may be used if it is proven by
adequate calculations.
Figure 4.3
Configuration factor and radiation distance (Collier, 1996)
-39-
Calculation methods
4.4
C.R. Barnett
This method by Barnett (1988) can be used in the design process of new buildings and
to check critical situations for existing buildings. The method does not rely on the use
of water or the intervention by the fire brigade and can be linked to the fire resistance
rating of the building.
The method can be used in two ways. Either to calculate the maximum allowed
unprotected permissible openings or the incident radiation intensity onto the
neighbouring building.
Maximum allowed permissible openings
The area of the maximum allowed permissible openings can be determined with
Equation 4.1.
AV =
AE I RC
n I EC
[4.1]
where:
AE =
area of enclosing rectangle [m2].

A number of different rectangles that enclose a combination of openings (both
big and small) in the facade should be chosen.
IRC = critical value of received radiation [kW/m2].

For design purposes this value should be chosen 12.5 kW/m2. However, if the
neighbouring building is fitted with non-combustible cladding and fire
resistance glazing, the received radiation can be 25 kW/m2 since no ignition
source (e.g. spark) is present inside the window and the presumed ignition
therefore has to be spontaneous. A fire window that remains in place for the
whole duration of the fire is assumed to decrease the passing radiation
intensity by 50 %, i.e. the radiation outside the fire window can be 50 kW/m2
if the external walls are made of non-combustible material.
n =
configuration factor of the enclosing rectangle.

The enclosing rectangle can be divided into four minor rectangles of equal size
and the configuration factor calculated as the sum of the configuration factors
of each of these rectangles see Equation 4.2.
[4.2]
n = A + B + C + D = 4 A
The configuration of the smaller rectangles can be calculated with Equation

4.3.
A =
where:
Y
1 X
tan 1
2
360 1 + X 2
1+ X
X
H/2R
-40-
X
Y
+
tan 1
2
1+ Y 2
1+ Y
[4.3]
Calculation methods
Y
R
H
W
=
=
=
=
W/2R
radiating distance [m]
height of enclosing rectangle [m]
width of enclosing rectangle [m]
In situations where flame projection out of the opening is expected, an extra

two metres (or other appropriate distance if it is shown by adequate
calculations) should be added to the radiating distance.
IEC =
critical value of emitted radiation [kW/m2].

The emitted radiation can be calculated with the Equation 4.4.
I EC = T2 T1
[4.4]
= emissivity of radiator, generally 1.0

= Stefan-Boltzmann constant = 56.710-12 [kW/m2K4]
T2 = temperature in fire compartment [K]
T1 = ambient temperature [K]
where:
The temperature in the fire compartment can be determined by using the ISO
834 time-temperature curve as follows (Equation 4.5):
T2 = 345 log10 (8t m + 1) + T1
where:
[4.5]
tm = time [min]
T1 and T2 are expressed in degrees Centigrade.
If fire resistance glazing is fitted in the fire compartment, the radiation

intensity that passes through the window can be reduced by 50 %.
Level of received radiation by the neighbouring building
The level of received radiation by the neighbouring building must not exceed the
critical design value of received radiation IRC. The level of received radiation intensity
can be determined with Equation 4.6.
I R = kV n I EC
where:
kV
IEC
I RC
[4.6]
ratio between the area of vertical openings and the area

of enclosing rectangle, AV/AE.
configuration factor of enclosing rectangle, see
explanation above.
critical value of emitted radiation [kW/m2], see
explanation above.
This latter method can be used to check whether an existing building situation is
critical or not.
-41-
Calculation methods
4.5
Williams-Leir
This method by Williams-Leir presents a way to determine safe separation distances

between buildings where the results always are conservative. The method takes into
account for a flame projection of 3 ft (1 m) out from the windows. Two different
configuration factors are incorporated in the equations. They are 0.035 for high hazard
occupancies and 0.070 for moderate hazard occupancies.
The derivation of the methodology will not be presented in this report. The interested
reader are referred to the original report Another Approximation for Spatial
Separation by Williams-Leir (1970).
Before the determination method is presented, a number of expressions and equations
must be defined. They are:
EBF
L
K
J
BAUO
Exposing Building Face

The section of the facade that is situated within the boundaries of a fire
compartment.
Limiting distance [ft]
The closest distance between the exposing building face (EBF) and the
boundary/lot line.
Floating-square distance [ft]
K = 3.54(L-3)
Floating-zone distance [ft]
J = 3.14(L-3)
Basic Allowance of Unprotected Openings [ft2]
BAUO = 0.44(L-3)2 for high hazard occupancies
BAUO = 0.88(L-3)2 for moderate hazard occupancies
The implication of these expressions and equations is determined by a set of rules.

Rule 1 is valid where it is needed to determine the separation distance or maximum
allowed unprotected areas for as much as 90 % of all cases. Rule 2 may be used for
facades with a low amount of unprotected areas (approximately less than 15 % of the
total area of the facade). Rule 3 is applicable for elongated facades where one of the
dimensions is eight times longer than the other.
Rule 1:
Every exposing building face (EBF) may include one or more basic
allowance of unprotected openings (BAUO). The openings may be
distributed in whatever way that is most suitable to the owner, given
that the EBF includes a maximum of one BAUO.
Rule 2:
Where one of the sides of the EBF is greater than K, a floating square
of side K that is parallel to and placed on the EBF can only contain a
maximum of one BAUO. The openings shall be distributed to fulfil
this demand.
Rule 3:
Where one of the sides of the EBF is greater than J, a suitable

direction, generally perpendicular to the longest side of the EBF shall
be chosen. A floating zone of side J, unlimited length and that is
parallel to the chosen direction can only contain a maximum of one
BAUO. The openings shall be distributed to fulfil this demand.
-42-
Calculation methods
4.6
J.H. McGuire
This method by McGuire (1965) can briefly be described by the following three steps:
1.
2.
3.
Determine the height and width of the relevant fire compartment.

Determine the percentage of window openings in the facade of the fire
compartment. All parts of the wall that are not provided with sufficient
integrity to meet with the performance requirements should be treated as a
window opening.
McGuire sets out two tables, one for hazardous conditions and one for
normal conditions, with values in feet of required building separations. Once
the terms in steps 1. and 2. are determined, the distance can be easily read
from the tables.
The two tables can be found in the original report Fire and the Spatial Separation of
Buildings by McGuire (1965). In this method, piloted ignition of facade material is
said to occur at an incident radiation level of 12.5 kW/m2.
McGuire (1965) defines the configuration factor as;
The ratio of the radiant intensity at the receiving surface to that at the
(one or more) radiating surfaces.
For design purposes, configuration factors of 0.035 for hazardous cases, i.e. buildings
with highly flammable linings, and 0.07 for normal cases, i.e. buildings with noncombustible linings, are used. These values are derived by dividing the value for
piloted ignition by the maximum value of radiation measured in a series of tests in
1958 called the St. Lawrence Burns. However, much higher values of radiation was
measured during the tests than what is used to derive the above configuration factors.
The configuration factors are still justified since it was noted that much lower
radiation intensities, approximately one fifth of the maximum radiation intensities,
was measured during the first 16 minutes of the tests. After this time it can be
assumed that the fire brigade has responded and has thereby prevented fire spread to
adjoining the buildings.
The tables also take into account horizontal flame projection out of windows. 7 ft is
added to the separation distance for hazardous cases and 5 ft is added for normal
cases. The projections are justified by the St. Lawrence Burns tests.
Since the method uses a percentage of openings, the openings need to be distributed
evenly throughout the facade in reality. If most of the windows are accumulated in
one end of the building, a greater separation distance would be required for that part
of the building. This is also valid where some of the windows are much larger than
the other windows. When determining the separation distance for irregular shaped
buildings, a line should be drawn between the farthest points of the facade. If all parts
of the building are contained behind the line, the facade and openings can be
projected onto the line and the separation distance determined based on this new
imaginary facade. If any part of the building extends beyond this line, this part should
be considered separately.
-43-
Calculation methods
A dangerous situation may occur where neighbouring buildings are of different shape
and size. Generally, the smaller building will be exposed to the greatest hazard. The
reason for this is that the separation distance is often determined in relation to the
property boundary rather than to the other building. Then a smaller building could be
situated closer to the boundary than a larger building would be and thereby be
exposed to too high levels of radiation from a fire in a larger building.
4.7
NFPA 80A
The scope of the NFPA 80A Recommended Practice for Protection of Buildings
from Exterior Fire Exposures (NFPA 80A 1996) is to protect combustible material
on the outside as well as the inside of a building exposed to an external fire source.
The document is intended as a guide for assurance of property protection of buildings
exposed to an external fire. Below is a description of how to determine the separation
distance for buildings of greater or equal height as well as for building of lesser
height.
Buildings of greater or equal height
The thermal radiation is the only means of exposure that needs to be considered when
a building is exposed by a fire in a building of at least the same height. The separation
distance between the buildings should be determined in a way to protect the exposed
building and its contents from igniting due to piloted ignition. Determination of
separation distances should be based on the assumption that neither of the buildings is
fitted with means of protection against fire by, for example sprinklers and fire
resistant glazing.
The width of the exposing wall that needs to be considered is the width between
existing fire separations within the building. If no separations exist within the
building, the width between the external walls should be used. The height of the
exposing wall is dependent on factors such as number of stories involved in the fire,
type of construction, vertical openings in the wall and fire resistance of floors.
Openings
Openings in an external and exposing wall that have to be considered are doors,
windows and other openings that may contribute to the emitted thermal radiation.
Other openings that should be considered are:
Walls with less resistance against fire penetration than 20 minutes should
be treated as being a 100% opening.
Walls that can withstand fire penetration for more than 20 minutes but
not longer than the estimated fire duration should be treated as being a
75% opening.
Severity
The severity of a fire will influence the amount of thermal radiation emitted from an
exposing building. The fire severity is highly dependent on two factors; the average
fire load per unit floor area within the building and the characteristics of interior
finishes. Based on these factors, the fire severity can be divided into three major
groups, namely light, moderate and severe. Table 4.3 (Table 2-2.4(a)-(b) in NFPA 80
-44-
Calculation methods
A, 1996) shows the fire severity groups and the criteria after which they should be
determined.
Table 4.3
Classification of fire severity. Adapted from the NFPA 80A.
Fire severity group Fire load per unit Average flame spread rating of
floor area (kg/m2) interior wall and ceiling finish
Light
0-34
0-25
Moderate
35-73
26-75
Severe
74
76
The factor resulting in the highest fire severity group should be used for design
purposes. Consideration should be taken if the whole building or only a portion of the
building is fitted with combustible interior finishes. The flame spread ratings are
explained in the NFPA 255 Standard Method of Test of Surface Burning
Characteristics of Building Materials.
According to the explanatory part of the NFPA 80A, exposure fire severity can be
defined as; the intensity of the exposing fire. Hence, the fire severity is the amount of
thermal radiation passing through openings or from flames projected outside the
building. The fire severity is dependent on several factors such as the ventilation
conditions of the compartment in concern, fire load, characteristics of the fuel,
geometry of the room and properties of the interior finishes.
Separation distances
It is assumed that the facade is made of cellulosic materials, with the ability to
withstand ignition when exposed to a maximum radiation level, Icrit, of 12.5 kW/m2.
Table 4.4 (Table 2-3 in NFPA 80A, 1996) below, sets out guide numbers that should
be used when determining the separation distances between buildings. To be able to
determine a guide number from the table, the fire severity, percentage of openings in
the wall and the ratio of the width-to-height or height-to-width of the compartment or
the enclosing rectangle must be known. To calculate the required separation distance,
D, the guide number, g, should be multiplied with the lesser dimension of the width,
w, and height, h, of the exposing fire and then added by 1.52, which is given by
Equation 4.7. The extra 1.52 m is added in order to take into account for flame
projections out of openings and prevent ignition due to flame impingement on the
exposed building.
D = gZ + 1.52
where:
g
Z
=
=
guide number from Table 4.4

lesser dimension of w and h [m]
-45-
[4.7]
Calculation methods
Table 4.4
Guide numbers for the determination of separation distances.

Adapted from NFPA 80A (1996).
Fire severity
Percent openings
Guide number
Ratio Width-to-Height or Height-to-Width
Light Moderate Severe 1.0

20
30
40
50
60
80
100
-
10
15
20
25
30
40
50
60
80
100
-
5
7.5
10
12.5
15
20
5
30
40
50
60
80
100
0.36
0.60
0.76
0.90
1.02
1.22
1.39
1.55
1.82
2.05
2.26
2.63
2.96
1.3
1.6
2.0
2.5
3.2
10
13
0.40
0.66
0.85
1.00
1.14
1.37
1.56
1.73
2.04
2.30
2.54
2.95
3.32
0.44
0.73
0.94
1.11
1.26
1.52
1.74
1.94
2.28
2.57
2.84
3.31
3.72
0.46
0.79
1.02
1.22
1.39
1.68
1.93
2.15
2.54
2.87
3.17
3.70
4.16
0.48
0.84
1.10
1.33
1.52
1.85
2.13
2.38
2.82
3.20
3.54
4.13
4.65
0.49
0.88
1.17
1.42
1.64
2.02
2.34
2.63
3.12
3.55
3.93
4.61
5.19
0.50
0.90
1.23
1.51
1.76
2.18
2.55
2.88
3.44
3.93
4.36
5.12
5.78
0.51
0.92
1.27
1.58
1.85
2.34
2.76
3.13
3.77
4.33
4.82
5.68
6.43
0.51
0.93
1.30
1.63
1.93
2.48
2.95
3.37
4.11
4.74
5.30
6.28
7.13
0.51
0.94
1.32
1.66
1.99
2.59
3.12
3.60
4.43
5.16
5.80
6.91
7.88
0.51
0.94
1.33
1.69
2.03
2.67
3.26
3.79
4.74
5.56
6.30
7.57
8.67
0.51
0.95
1.33
1.70
2.05
2.73
3.36
3.95
5.01
5.95
6.78
8.24
9.50
16
20
25
32
In cases where the facade consists of material with a different critical radiant heat
flux, i.e. Icrit12.5 kW/m2, and where no openings exist in the external wall of the
exposed building, the percentage of openings in the exposing building should be
adjusted. The new percentage of openings, Onew, should be obtained by multiplying
the old percentage openings of the external wall, Oold, with the ratio between
12.5 kW/m2 and the critical radiant heat flux for the particular facade material. This
procedure is explained in Equation 4.8.
Onew = Oold
12.5
I crit
[4.8]
It is very important to note that the method of determining separation distances

contemplates Fire Brigade intervention. At locations where Fire Brigade intervention
can not be guaranteed within reasonable time, the separation distance should be
extended by multiplying by a factor of up to three. The values of emitted radiation
used are obtained from a series of full-scale tests called the St. Lawrence Burns.
These tests indicated much higher levels of radiation for fire duration of more than 20
minutes. Other findings from the St. Lawrence Burns were that the amount of
emitted radiation was affected by the percentage of openings in the external walls.
Furthermore, high amount of combustible interior finishes resulted in very high levels
of emitted radiation outside the building. The external covering of the wall did not
contribute to significantly higher levels of radiation.
Table 4.4 has been derived by using the methodology presented in Fire and the
Spatial Separation of Buildings by McGuire (1965). NFPA specifically states that no
consideration has been taken to the risk of ignition by convective heat transfer and
flying brands. The critical level of thermal radiation, i.e. 12.5 kW/m2, where piloted
ignition of cellulosic material may occur has been obtained from work done in the UK
by the Joint Fire Research Station.
The amount of thermal radiation received by the exposed building can be determined
with Equation 4.9 below.
-46-
40
0.51 0.51 0.51 0.51 0.51

0.95 0.95 0.95 0.95 0.95
1.34 1.34 1.34 1.34 1.34
1.71 1.71 1.71 1.71 1.71
2.07 2.08 2.08 2.08 2.08
2.77 2.79 2.80 2.81 2.81
3.43 3.48 3.51 3.52 3.53
4.07 4.15 4.20 4.22 4.24
5.24 5.41 5.52 5.60 5.64
6.29 6.56 6.77 6.92 7.01
7.23 7.63 7.94 8.18 8.34
8.89 9.51 10.05 10.50 10.84
10.33 11.15 11.91 12.59 13.15
Calculation methods
[4.9]
I = I 0
where:
I0
=
=
received radiation at the surface of the exposed building

[kW/m2]
emitted radiation from the exposing building [kW/m2]
configuration factor
The configuration factor is dependent on the size of the radiating surface and the
separation distance. The configuration factor can be determined by using Equation
4.10, assuming the radiation surface being of rectangular shape.
=
where:
X
Z
2
arctan
2
2
2
2
X + Y
X +Y
X
Y
Z
=
=
=
Z
X
+
arctan
2
2
2
2
Y +Z
Y +Z
[4.10]
half of the width of the radiating surface [m]

half of the height of the radiating surface [m]
distance between the radiating surface and the exposed
surface [m]
Table 4.4 has been developed based on evenly distributed openings, which are not
separated from each other by more than one third of the total separation between the
buildings. Where this is not the case, additional calculations are required. The
following methods should be considered:
1. A calculation should be made for the smallest area including all
openings, which in some cases may be a single window. The largest
calculated separation distance for this area should be used.
2. A calculation should be made for a single opening in cases where the
openings are separated from each other by more than one third of the
total separation distance between the buildings.
According to the NFPA 80A (1996), the amount of received radiation at a specific
point opposite the exposing building is barely influenced by radiation from openings
separated from the point by more than twice the total separation distance between the
buildings, as shown in Figure 4.4. Then calculations for a single window should be
made and considered valid.
2D
Figure 4.4
2D
Separation of openings from a specific point at the exposed building
-47-
Calculation methods
Buildings of lesser height

The separation distance for an exposing building of lesser height than the building
that are being exposed should first be determined with the same method as explained
above for buildings of the same height or higher. In cases where the roof of the
exposing building consists of combustible material without a fire resistance rating,
Table 4.5 (Table 2-4 in NFPA 80A, 1996) should be used in order to determine the
required separation distance and means of protection necessary above roof level.
Table 4.5
Required separation distance or height of protection for buildings with combustible

and roof with no fire resistance rating. Adapted from NFPA 80A (1996).
Number of stories likely to contribute

to flaming through the roof
1
2
3
4
Horizontal separation distance or height

of protection above exposing fire (m)
7.6
9.8
12.2
14.3
When the distances determined from Table 4.5 are greater than those determined from
Table 4.4, the exposed building should be fitted with means of protection. The
protection should have a height above roof level of the exposing building that is equal
to the required separation distance. Where the roof of the building is not provided
with enough fire resistance to contain a fire, Table 4.5 should be applied. The top
storey and those stories directly underneath that are not fire separated from each other,
should be counted as the number of stories likely to contribute to flaming through the
roof.
Means of protection
According to the NFPA 80A (1996), several means of protection of a building from
an external fire source can be applied. The means of protection are listed below.
Spatial separation between buildings

An automatic sprinkler system
External walls of non-combustible material
Extended external masonry walls to form protective wings
Automatic external water curtains
Removal of openings
Openings fitted with:
glass block panels
wired glass in steel sash
automatic or deluge sprinklers
fire shutters
fire doors
fire dampers
If means of protection are installed, the separation distances can be decreased in

accordance to Chapter 4 in NFPA 80A (1996). Appropriate standards should be
followed and the means of protection installed in a building should be approved for its
particular application.
-48-
Calculation methods
No exposure hazard is assumed to exist if the exposing building is fitted throughout

with a properly installed and maintained automatic sprinkler system. The exposure
hazard to the exposed building is assumed to be greatly reduced if the exposed
building is fitted throughout with an automatic sprinkler system.
-49-
Calculation methods
-50-
Comparison of boundary requirements
5 COMPARISON OF BOUNDARY REQUIREMENTS

In this chapter a comparison of the different calculation methods presented in chapter
4 will be made by applying the methods on fictitious buildings. Three different types
of buildings will be investigated; a warehouse, an office building and a residential
building. Conclusions with regards to the applicability of the methods will be made
based on the findings from the comparison.
5.1
Warehouse
A fictitious warehouse has been used for the comparison between different methods
of determining separation distances between buildings. The building is assumed to be
50 m wide and 50 m long and have a height of eight metres. Furthermore, the external
walls are not provided with enough fire resistance to contain a fire for its whole
duration, i.e. the whole facade is considered to be an opening. The interior finishes are
non-combustible. The fire load within the building is assumed to be of the same size
as is generally found in warehouses and storage facilities. It is estimated that the
duration of the fire will be 120 minutes. This estimation is made for comparison
purposes. A fire will according to the ISO 834 time-temperature curve have then
reached a temperature of about 1050 C. This fire temperature represents a radiation
intensity of 173 kW/m2, which is approximately the same levels of radiation used in
those methods with fixed radiation values, (Law 1963).
In the comparison, the distance to the relevant boundary will be determined. It is
assumed that an identical building is positioned at the same distance from, but on the
other side the boundary, i.e. the buildings are mirror images of each other. The critical
received radiation is in all cases set to be 12.5 kW/m2.
The boundary distance that is required by the different determination methods is
shown below in Table 5.1. The calculations that have been performed in order to
achieve the results below are presented in detail in Appendix 1.
Table 5.1
Comparison of required boundary distance for the warehouse used in the study
Method
Peter Collier
C.R. Barnett
J.H. McGuire
G. Williams-Leir
NFPA 80A
Enclosing rectangle
Required boundary distance (m)

19.4
20.4
19.2
22.2
19.9
18.3
As can be seen from the table, the methods predict approximately the same boundary
distance. If the enclosing rectangle method, which does not take into account for
flame projections, is used as a basis for the comparison, it can be seen that the
methods will compare even better if the flame projection of each method is subtracted
from the overall boundary distance. Therefore, the methods predict approximately the
same boundary distance for relatively simple building shapes and opening
configurations. One exception is the method by Williams-Leir, which predicts a
boundary distance slightly higher than the other methods.
-51-
The methods by Collier and Barnett are almost similar and the difference in boundary
distance is most probably due to the lack in precision when obtaining values from the
graph of configuration factors, aspect ratios and constants used in Colliers method.
Another factor that may have a small impact is that Colliers method does not take
into account for the temperature of the receiving surface and ambient air when
determining the levels of emitted radiation. The radiation intensity is therefore slightly
lower than what is used in Barnetts method.
5.2
Office building
A fictitious office building has been used in this comparison and Figure 5.1 shows the
building layout and relevant measurements. The building is 40 m wide and consists of
five storeys, each with an individual height of 3.8 m. Each storey is a separate fire
compartment, i.e. the floors will resist a fire for its whole duration. Seven windows,
two metres high and three metres wide, are distributed evenly along each storey. A
distance of three metres separates the windows from each other. The external wall is
fitted with non-combustible cladding.
0.6m
19m
2m
1.2m
3m
3m
3m
3m
3m
3m
3m
3m
3m
3m
3m
3m
3m
40m
Figure 5.1
Office building used in the comparison
The fire is assumed to last for a period of 0.5 h (30 min), resulting in a fire
temperature of 842 C according to the ISO 834 time-temperature curve. This
temperature will cause a radiation intensity inside the fire compartment of 87 kW/m2,
which is approximately the same level of radiation used for an office building in the
methods by Read and NFPA 80A and will therefore make the calculations more
comparable. The critical incident radiation is set to be 12.5 kW/m2.
In the comparison, the distance to the relevant boundary was determined. It is
other side the boundary, i.e. the buildings are mirror images of each other.
The boundary distance required by the different calculation methods is shown below
in Table 4.2. The calculations that have been performed in order to achieve the results
below are presented in detail in Appendix 2.
-52-

Table 5.2
Comparison of required boundary distance for the office building used in the study
Method
Peter Collier
C.R. Barnett
J.H. McGuire
NFPA 80A
Enclosing rectangle

2.9
3.0
4.9
2.7
2.0
The enclosing rectangle method by Fire Research Station (1991) is the method
predicting the shortest boundary distance, but it should be noted that this is the only
method that does not take into account for flame projection. The method by McGuire
(1965) is not totally comparable with the others since the table for normal cases uses a
configuration factor of 0.07 while the other methods use a configuration factor for
office buildings of 0.14 or 0.15. This is the reason why the method by McGuire
(1965) predicts a boundary distance that is about two metres higher than predicted by
the other methods.
Once again, the methods predict similar boundary distances for a building with rather
simple layout. A difference can be noted between methods that use flame projection
and methods that do not.
The complete results of each method presented in Appendix 2 shows that the
enclosing rectangles that cause the largest boundary distance is the rectangles that
enclose two, three or four of the windows in the storey. If a fire duration of 120
minutes and hence a higher radiation intensity had been chosen for the methods by
Collier (1996) and Barnett (1988), the predicted boundary distance would be
approximately 1.7 m longer. The distance would in that case be caused by the
rectangle that encloses all seven windows in the storey.
5.3
Residential building
The fictitious residential building that has been used in this comparison is two storeys
high which equates to a total height of seven metres. No fire compartmentation exists
within the building or between the storeys, i.e. a fire is assumed to involve the whole
building. The building is 20 m wide and has a recess located in the right corner of the
building. The recess is 1.5 m deep and five metres wide. Figure 5.2 shows a
simplified layout of the building.
Recess
Figure 5.2
Simplified layout of the residential building
There are seven openings in the facade and they are distributed according to Figure
5.3. The openings consist of five windows, one door and one carport. The carport and
-53-
one large window are positioned in the recess while the other openings are positioned
in the other part of the facade. The external wall is fitted with non-combustible
cladding.
0.5m
3m
1m
3m
1m
0.8m
0.8m
1.5m
1.9m
1.5m
1.2m
1.2m
1.2m
0.5m
7m
1.4m
1.9m
2.1m
3m
1.2m
0.5m 1.5m
1.5m
1.9m
7.2m
1.9m 0.5m 0.5m
4m
0.5m
20m
Figure 5.3
Residential building used in the comparison
The fire is assumed to last for a period of 0.5 h (30 min), resulting in a fire
temperature of 842 C according to the ISO 834 time-temperature curve. This
temperature will cause a radiation intensity inside the fire compartment of 87 kW/m2,
which is approximately the same level of radiation used for an office building in the
methods by Read and NFPA 80A and will therefore make the calculations more
comparable. The critical incident radiation is set to be 12.5 kW/m2.
The boundary distance required by the different determination methods is shown
below in Table 5.3. The calculations that have been performed in order to achieve the
results below are presented in detail in Appendix 3.
Table 5.3
Comparison of required boundary distance for the residential

building used in the study
Method
Peter Collier
C.R. Barnett
J.H. McGuire
NFPA 80A
Enclosing rectangle

With recess
Without recess
3.5*
3.4
3.0
3.6
5.8*
n.a.
2.7
3.3
2.2
2.2
* Boundary distance counted from the plane of reference.

n.a. = not applicable
With this more complex building shape, the boundary distance required by the
different calculation methods varies more and are harder to compare. Table 5.3 is
divided into two columns. The column labelled With recess is for the actual
-54-
building layout and the methods are applied exactly according what is described in
each method, e.g. where nothing is said about a recess, no plane of reference is
established that touches the outer parts of the building and makes an angle to the
facade. The column labelled Without recess sets out the boundary distances when
the recess is assumed not to exist.
The method by The Fire Research Station (1991) once again predicts the shortest
boundary distance. One reason is that no account is taken to flame projection, but
another contributing factor may be that the enclosing rectangles have to have
dimensions according to the tables. This will make the area of the enclosing rectangle
larger than the smallest rectangle that encloses the same openings in the other
calculation methods. Since the area of unprotected openings is the same but the area is
larger, the unprotected percentage will be smaller.
The other methods vary more in relation to each other for this building. It should be
noted that the distance predicted from the methods by Collier and McGuire is counted
from the plane of reference.
The method by McGuire (1965) is not totally comparable with the others for this
building either. The reasons for this are that a higher configuration factor is used and
that the method is made for evenly distributed openings.
Where the recess is assumed not to exist, approximately the same boundary distance
is predicted by the different methods. Once again the method that does not take into
account for flame projection predicts the shortest distance.
5.4
Discussion of results
The performed comparison has drawn the attention to some interesting and important
matters when it comes to determining boundary distances with different methods.
There is no consistency with regards to if account for flame projections should be
taken and in that case, which distance that should be used.
As has been seen in the comparison, the methods predict approximately the same
boundary distance for simple building shapes. However, for more complex building
shapes there are greater differences between the different methods. Since some
methods do not cover all situations, differences will occur due to variation in
judgements of the user.
The methods by McGuire (1965), Williams-Leir (1970), Fire Research Station (1991)
and NFPA 80A (1996) use predetermined values of radiation, configuration factor and
flame projections (where such are used), and no performance-based design is
possible. In the methods by Barnett and Collier, there is a greater flexibility with
regards to determining the radiation intensity, flame projections and hence the
boundary distances. This is a great advantage today with all new building materials,
furnishings and more complex building shapes. One disadvantage is that it requires an
experienced engineer in the field of fire protection to perform a design according to
these methods. An advantage with the methods using predetermined values is that
they can be used and applied by persons without any particular knowledge in fire
protection engineering.
-55-
McGuires (1965) method is restricted in the way that it does not have a table for
buildings with low fire load density such as office and residential buildings. It is
though stated that such tables can be available on request. Furthermore, the method is
only applicable for buildings with evenly distributed openings or windows.
The enclosing rectangle method presented by the Fire Research Station (1991) gives a
good and systematic description of how to determine the boundary distance. Detailed
explanations are given of how complicated building shapes should be treated. The
aggregate notional area method that is presented in the same document can be used as
a complement to the enclosing rectangle method, which may be especially favourable
for complicated building shapes. An example is buildings with recesses where a plane
of reference is used to determine the boundary distance. The distance may then be
overestimated at some locations and the aggregate notional area method can be used
to find the proper boundary distance at those specific locations.
It has been shown that for higher intensity fires, more openings and openings
separated further apart will contribute with radiation and thus increase the boundary
distance. For very wide buildings such as the warehouse used in this study, this will
lead to very large separation distances. One question that arises from the above
discussion is whether it is reasonable to assume that the whole facade of a warehouse
will radiate at the same time? It may well be a reasonable assumption for smaller
facades, but since it takes time for the fire to spread horizontally within the building it
may lead to an overestimation of the boundary distance for wider buildings.
-56-
Importance of radiation from projected flames
6 IMPORTANCE OF RADIATION FROM PROJECTED FLAMES

In the comparison of the different calculation methods it was found that one
difference between the methods was whether account should be taken for flame
projection or not. This chapter will investigate if a projected flame from an opening
contributes with enough radiation for it to be necessary to take into account when
determining safe separation distances between buildings.
6.1
Full scale fire tests
Law (1968) reports on a series of full scale fire tests that were conducted to
investigate the levels of radiation from fires in compartments. The compartments used
were brick-walled with dimensions 7.73.73 m (widthdepthheight). Two different
window areas, 5.6 and 11.2 m2, were used and the total fire load, fire load distribution
and wall linings were varied. The fuel used in the tests was 45 mm thick wood cribs.
In order to measure the radiation contributed by the projected flames, two radiometers
were placed adjacent to the burning building and facing the midpoint of the windows.
One of the radiometers was shielded from the flames above the windows, i.e. one
radiometer only measured the radiation from the windows while the other radiometer
measured the total radiation from both the window and the projected flames.
From these tests, Law (1968) concluded that the fire was cooler and hence the
radiation less for the larger windows, independent of the fire load. The window area
and fire load had a significant effect on the window radiation. The tests also showed
that the radiation from flames was not significant for the larger windows and
significant only at the 20 percent level for the smaller windows. Law (1968) therefore
came with the conclusion that:
Thus the analysis of these eleven tests indicates that there is a small but not
highly significant increase in intensity due to flames from the smaller
window; it can therefore be neglected.
and that this small increase in intensity from the smaller windows does not justify an
increase in building separation recommendations.
Law (1968) also reports on another full scale test carried out by Webster and Smith
(1964) and, where one side of the 2.4 m cubical compartment was totally opened. In
this test it was estimated that the radiation contributed by the flames above the
opening was only two percent of the total amount of received radiation.
6.2
Swedish method
Fredlund et al. (1976) have developed a method of determining separation distances

between buildings which is based on both full scale and small scale fire tests, as well
as theoretical derivation. The method uses fire load density, opening factor, dimension
of opening areas and the area of the compartments interior surfaces as inputs. The
method has been approved by the Swedish building authorities to be used in the
design process of small houses.
-57-
6.2.1 Tests
Calculation of time-temperature curves for the fire compartment was performed with
heat and mass balance equations according to methods by Magnusson and
Thelandersson (1970, 1971, 1974). This is the same method that has been used to
derive the so-called Swedish fire curves. Note that the time-temperature curves used
in this model were derived for the specific test configuration and therefore differ from
the original Swedish curves. The validity of the results calculated with the theoretical
model has been checked with the time-temperature curves obtained in the tests. The
fire compartments used in the tests were made of light-weight concrete with a density
of 500 kg/m3.
A number of assumptions have been necessary in deriving the curves.
combustion is complete and takes place entirely within the confines of the
compartment;
the temperature is uniform within the compartment at all times;
a single surface heat transfer coefficient may be used for the entire inner
surface of the compartment; and
the heat flow to and through the compartment boundaries is unidimensional,
i.e. corners and edges are ignored and the boundaries are assumed to be
infinite slabs.
In the same tests by Fredlund et al. (1976), the radiation received at different points
on the facade of an adjoining house was measured. These measures have been the
basis behind this method to predict incident radiation and separation distances. A
strength with the model is that it takes into account for both the radiation from the fire
compartment through openings as well as the radiation from externally projected
flames. The calculated levels of radiation can also be combined with ignition
characteristics of the receiving surface to predict time to ignition. Three sets of tables
are set out for design purposes; one that gives the time-temperature curve of the fire
compartment, one that gives the time-radiation curve at various distances and
locations on the neighbouring building and finally the sidewise radiation from
external flames. All these tables are given for different fire loads and different
opening factors. In the tables of received radiation at various distances and locations
on the receiving surface, expressions for emissivity and configuration factors are
already incorporated. The tables also set out a correlation between time of the burning
period and received radiation. The radiation is expressed as a fraction of the total
received radiation and allows the user to establish a time-radiation curve for the
receiving point. Below is a brief description of which equations and assumptions that
has been used to obtain the tabulated values.
6.2.2 Limitations
The method is derived for one-storey buildings with walls and supporting members
made of light-weight concrete. The roof should be constructed in a way that has little
ability to ignite and thereby starting a secondary fire, i.e. the roof could extend a
distance no greater than 0.5 m from the facade. If the roof extends further than this,
the extra distance should be added to the building separation distance. Ceilings and
internal walls should be constructed so that they are difficult to ignite. The method is
-58-
derived for houses made of light-weight concrete but can also be used for houses
made of concrete, with a slightly conservative separation distance as a result.
Another restriction in the model is that the width of the opening, from which radiation
is to be calculated, must not be greater than the separation distance. In this case, the
opening should be divided into smaller parts so that the above requirement is fulfilled
and the radiation from each part summarised.
6.2.3 Theory
The radiative contribution from the fire, i.e. the flame neglected, has been calculated
with Equation 6.1 and 6.2. Equation 6.1 gives the radiative heat exchange dQ12
between two surfaces and Equation 6.2 gives an expression for the radiation dP12
received by a certain point at a distance, r, from the radiator.
dQ12 = rTa4 dA1dA2
where:
Ta
r
1
2
=
=
=
=
=
=
cos 1 cos 2
r 2
[6.1]
resulting emissivity of both elements

Stefan-Boltzmanns constant = 1.3710-12 cal/cm2sK4
temperature of radiating surface dA1, [K]
distance between surfaces [m]
angle between r and the normal to surface dA1
angle between r and the normal to surface dA2
dP12 =
dQ12
dA2
[6.2]
The orientations used in the equations above are explained in Figure 6.1.
dA1
1
r
2
dA2
Figure 6.1
Orientations used when calculating the received radiation
In the tests, air with a velocity of 10 m/s was blown into the fire compartment, which
is claimed by Thuresson (1973) to represent natural wind conditions with a velocity of
20 m/s. The projected flames were noticed to have a triangular shape, see Figure 6.2
below.
-59-
h/3
Flame axis
h/3
h/3
Lflame
Figure 6.2
Simplified layout of exterior projected flame used in the model

by Fredlund et al. (1976)
The angle, , between the flame axis and the horizontal plane was estimated to be
30, which also can be used for design purposes. The maximum flame length was
determined with Equation 6.3.
[6.3]
1/ 3
L flame ,max = CR80
30
where:
R8030 =
C
average burning rate for period when the fuel loses its
weight from 80% to 30% of its initial value, [kg/min]
coefficient obtained from full scale fire tests,
[min1/3m/kg1/3], see Figure 6.3,
1,6
1,2
C
0,8
(min1/3m /kg1/3)
0,4
0
0
50
100
1/2
M/Ah
Figure 6.3
150
5/2
(kg/m )
The coefficient Cs dependence on amount of mass, M, in fire compartment, area, A,

of window and height, h, of window.
-60-
The total radiation from the flame to a remote receiver has been calculated by
dividing the flame into several smaller units and then adding the radiative contribution
from each unit.
The resulting emissivity, r, for the radiation from the window was calculated with
Equation 6.4.
1
1
1
=
+
1
r w m
where:
[6.4]
emissivity of the window opening,

0.55 is used in the model for the entire burning
emissivity of the receiving surface,
1.0 is used in the model
The emissivity of the flame is dependent on flame thickness and type of burnt fuel.
Figure 6.4 gives the correlation used in the model.
Emissivity
1
0,9
0,8
0,7
0,6
0,5
0,4
0,3
0,2
0,1
0
0
0,5
1,5
2,5
Flame thickness (m)
Figure 6.4
Flame emissivity as a function of flame thickness for wood based fuel
Fredlund et al. (1976) shows in a number of examples how well the time-temperature
curves of the theoretical derived method compares with the full scale tests. The
comparison in Figure 6.5 below is for an opening factor A h /Atot = 0.04. The upper
two curves are for a fire load f = 30 Mcal/m2 and a separation distance c = 4 m, while
the lower two curves are for a fire load f = 20 Mcal/m2 and a separation distance c = 6
m.
-61-
Figure 6.5
Comparison between theoretical derived calculation model and full scale fire tests.
The dashed curve is for the theoretical derived time-radiation relationship
and the solid curve is for full scale fire tests.
6.2.4 Calculation methodology

Calculations of separation distances or received radiation at various locations can be
divided into five main steps, which are briefly explained below.
1. Determination of the total fire load f within the fire compartment.
The fire load should be expressed as energy content (Mcal/m2 or MJ/m2) per unit
of total interior surface area (openings included and internal walls excluded). Note
that the fire load in this method is not, like in many other methods, expressed as
energy content per unit floor area.
2. Determination of the opening factor A h / Atot of the fire compartment.
where:
A
=
area of all vertical openings in the fire compartment
[m2]
h
=
average height of all openings [m]
Atot
=
total area of the fire compartments interior surfaces
(openings included and internal walls excluded) [m2]
3. The opening factor and fire load is then used to find the appropriate table for
received levels of radiation.
4. Equation 6.5 explains the correlation between the maximum received radiation at
the point of concern. The radiation set out in the tables. This procedure is
necessary since the tables are derived for Atot = 85 m2.
Pmax, p = p mp
Atot , p
85
-62-
Pmax,t
[6.5]
where:
Pmax,p =
p
mp
Atot,p
=
=
Pmax,t =
maximum radiation intensity at the relevant point

[kW/m2]
ratio between area Aa of the opening of concern and the
total area A of all openings
emissivity of receiving surface = 1
total area of the fire compartments interior surfaces
(openings included and internal walls excluded) [m2]
maximum radiation intensity set out in the tables
[kW/m2]
5. Establishing of a time-radiation curve

When the maximum received radiation has been determined, the correlation
between time and fraction of total radiation can be used to establish a timeradiation curve for the receiving point. Together with ignition characteristics,
which also is set out in the report by Fredlund et al., of the receiving surface, the
time-radiation curve can be used to predict whether ignition will occur and also
the time to ignition.
6.3
Projected flame or notional radiator
As has been shown in previous chapters, there is no consistency in the calculation

methods with regards to whether one should account for flame projection when
determining the separation distance between buildings. Most of the Building
Regulations studied in this report use the openings or unprotected areas as radiators
and neglect flame projection and radiation contributed from external flames. One
exception is Canada where a flame projection of 1.2 m is used , (Barnett, 1988). The
studied calculation methods use various projection distances. For example, Collier
(1996) and Barnett (1988) uses a distance of 2 m, NFPA 80A (1996) uses 1.5 m,
Williams-Leir (1970) uses 0.9 m and McGuire (1965) uses 1.5 m for normal cases and
2.1 m for hazardous cases.
6.3.1 Swedish method
It is the aim of the following chapter to check whether it is necessary to take into
account for flame projections when determining safe separation distances, or if the
assumption that the openings (with emissivity = 1) are the only areas contributing
with radiation is acceptable. This will be done by applying the method by Fredlund et
al. (1976), on a number of representative houses with different dimensions, window
configurations and fire loads. No definite conclusions can be made from the study, but
it can act as an indication to see if the above assumption is safe or if account for the
flame projection needs to be taken.
Six cases were investigated with the model by Fredlund et al. (1976). The basic inputs
and results of the calculations will be presented here and all main steps of the
calculations will be presented in Appendix 4. The emissivity of all receiving surfaces
is assumed to be mp = 1 and the point that receives the highest levels of radiation are
studied. The results from these calculations were compared with the radiation
achieved when only the window is assumed to radiate with emissivity = 1 and
-63-
compartment temperature estimated by the time-temperature curves given in Fredlund

et al. (1976).
Calculations were made for notional radiators, i.e. the window is assumed to be the
only radiator, especially at a projection distance of 0 m but also for distances 1.2, 1.52
and 2 m away from the opening. In case 1-5 comparisons were made at distances
where the received levels of radiation are significant for building separation purposes
while case 6 was investigated to show how the radiation varies with the distance. The
radiation, P, received from the notional radiators was calculated with Equation 6.6.
P = T f Ta
where:
Tf
Ta
=
=
=
=
=
[6.6]
emissivity of radiator = 1
Stefan-Boltzmanns constant = 56.710-12 kW/m2K4
configuration factor of a notional radiator
temperature in the fire compartment [K]
temperature of the receiving surface [K]
Case 1a
Compartment: 54.52.4 m (widthdepthheight)
1 window:
31.5 m (widthheight)
Fire load:
30 Mcal/m2 (125 MJ/m2)
Distance:
5m
Figure 6.6 below shows the results of all projection distances for case 1a and Figure
6.7 shows only the results from the method by Fredlund et al (1976) and a notional
radiator with no projection distance, i.e. 0 m. It can be seen from Figure 6.6, that the
received radiation from the notional radiators with projection distances of 1.2, 1.52
and 2 m was much higher than the levels of radiation resulted from the method by
Fredlund et al. (1976) and a notional radiator without any projection distance at all.
This was also the situation for all the other cases studied in this chapter. Therefore,
presentation of all radiation distances will only be made in Appendix 4.
-64-
25,00
20,00
15,00
N otional ra diator (P =0m )

N otional ra diator (P =1,2m )
Incident radiation
(kW /m ^2)

P rojected flam e
10,00
5,00
0,00
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
Tim e (h)
Figure 6.6
Comparison of received radiation for all projection distances for case 1
12,00
10,00
8,00
Incident radiation
(kW /m ^2)
6,00
P rojected flam e
4,00
2,00
0,00
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
Tim e (h)
Figure 6.7
Comparison of received radiation for method by Fredlund et al.(1976) and a

notional radiator with projection distance 0 m for case 1
The peak levels of radiation predicted by the method by Fredlund et al. (1976) and the
notional radiator are 10.1 and 8.6 kW/m2 respectively, which makes a difference of
1.5 kW/m2.
To check how higher fire load affects the radiation at the same distance, an extra
calculation was made for a compartment with the same dimensions as above but
where the fire load was higher.
Case 1b
1 window:
Fire load:
50 Mcal/m2 (209 MJ/m2)
Distance:
5m
-65-
Figure 6.8 shows a comparison for case 1b between the levels of radiation predicted
by Fredlund et al. (1976) and a notional radiator with a projection distance of 0 m.
12,00
10,00
8,00
Incident radiation
(kW /m ^2)
6,00
P rojected flam e
4,00
2,00
0,00
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
Tim e (h)
Figure 6.8
Received radiation for case 1b
1.8 kW/m2. The higher fire load resulted in increased levels of radiation for both
methods.
Case 2
Compartment: 8102.5 m (widthdepthheight)
3 windows: 21.3 m (widthheight) each
1 window:
4.21.3 m (widthheight)
Fire load:
50 Mcal/m2 (209 MJ/m2)
Distance:
5m
Figure 6.9 shows a comparison for case 2 between the levels of radiation predicted by
Fredlund et al. (1976) and a notional radiator with a projection distance of 0 m.
-66-
14,00
12,00
10,00
8,00
N otional ra diator (P = 0m )
In cid en t rad iation

(kW /m ^2)
P rojec ted flam e

6,00
4,00
2,00
0,00
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
T im e (h )
Figure 6.9
Received radiation for case 2
1.5 kW/m2.
Case 3
Fire load:
40 Mcal/m2 (167 MJ/m2)
Distance:
3m
Figure 6.10 shows a comparison for case 3 between the levels of radiation predicted
18,00
16,00
14,00
12,00
10,00
Incident radiation
(kW /m ^2)
P rojected flam e
8,00
6,00
4,00
2,00
0,00
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
Tim e (h)
Figure 6.10
The peak levels of radiation predicted by the two methods are 16.2 and 14.8 kW/m2
respectively, which makes a difference of 1.4 kW/m2.
-67-
Case 4 This case has been investigated since narrow window tend to produce long
flames that are projected a relatively great distance away from the facade.
Compartment: 6.35.53.0 m (widthdepthheight)
2 windows: 12 m (widthheight) each
Fire load:
30 Mcal/m2 (125 MJ/m2)
Distance:
3
14,00
12,00
10,00
8,00

(kW /m ^2)
P rojec ted flam e

6,00
4,00
2,00
0,00
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
T im e (h )
Figure 6.11
2.0 kW/m2.
Case 5 This case has been investigated since windows with equal sides also tend to
produce flames that are projected far away from the facade, (Barnett 1988).
4 windows: 1.451.45 m (widthheight) each
Fire load:
50 Mcal/m2 (125 MJ/m2)
Distance:
3m
-68-
14,00
12,00
10,00
8,00
Incident radiation
(kW /m ^2)
P rojected flam e
6,00
4,00
2,00
0,00
0,00
0,20
0,40
0,60
0,80
1,00
1,20
Tim e (h)
Figure 6.12
1.4 kW/m2.
Case 6 This case was investigated in order to check how well the method by
Fredlund et al. (1976) compares with the radiation from the window (with no flame
projection) only, when the distance to the receiving point increases.
1 window:
Fire load:
30 Mcal/m2 (125 MJ/m2)
Distance:
3, 4, 5, 6 and 8 m
-69-
14
12
10
8
Projected flame
Incident radiation
(kW/m^2)
Notional radiator (P=0m)

6
0
3
Radiating distance (m)
Figure 6.13
The biggest difference for this case occurs at a radiation distance of 3 m, where the
levels of radiation are 13.1 and 11.3 kW/m2 respectively. However, when the distance
increased, the radiation resulting from the notional radiator became larger in relation
to the method with flame projection. This is due to the shape of the flame. As the
distance becomes larger, the angle between a line connecting the receiver and the
flame front becomes smaller and the effective area of the flame seen from the receiver
will therefore be less. Accordingly, the fraction of received radiation resulting from
the flame was less as the distance increased.
6.3.2 Hand calculation model
The method investigated above can only be applied with accuracy for specific cases.
It would be interesting to check if equations that are more commonly used in the fire
engineering discipline will predict that the projected flame contributes with the same
percentage of radiation as observed in chapter 6.3.1. To do this a calculation
procedure has been established by the author of this report and sample calculations
performed on relatively simple cases. The method is presented for a compartment
with only one opening, but can also be extended to be valid for compartments with
several openings of various dimensions. The procedure is mainly based on Law et al.
(1981), but has been modified in certain ways to giver higher accuracy.
For no-wind conditions Law et al. (1981) uses a flame shape and dimensions
according to Figure 6.14.
-70-
2H/3
x
z1
2H/3
H
Flame axis
Figure 6.14
Flame shape and dimensions used in the calculation procedure,( Law et al. 1981)
The opening is assumed to radiate with the temperature found inside the fire
compartment and with an emissivity =1. One further assumption is that the leaning
part of the flame is assumed not to exist, i.e. the window will radiate with its total area
and is not shielded by any flame. This has been agreed on in correspondence with
Law. One argument against this assumption may be that the radiating distance will be
longer and accordingly the configuration factor smaller. However, the temperature in
the compartment is generally higher than the average flame temperature and the
emissivity in the compartment is much greater than the emissivity of the flame. Two
radiators will therefore be used in the calculations; the window with height, H, and
width, W, and the part of the projected flame that is above the soffit with height, z1,
and width, W. In the sample calculations, the received radiation is calculated in a
point located in the centre of the window at a distance, r, away from the facade. The
maximum received radiation may be found in point located slightly higher up but this
is assumed to have a negligible effect.
Radiation from the window
The radiation from the window is calculated in ordinary way with Equation 6.5.
Pwindow = T f Ta
where:
Tf
Ta
=
=
=
=
=
emissivity of radiator = 1
configuration factor of window
temperature in fire compartment [K]
temperature of receiving surface [K]
-71-
[6.5]
The temperature Tf in the fire compartment is determined with Equation 6.6 (Equation
3 in Law et al. 1981) or with any other appropriate method.
T f Ta = 6000
where:
Ta
AT
=
=
Aw
h
=
=
(1 e
0.10
1/ 2
)(1 e
0.05
[6.6]
AT
[m-1/2]
1/2
Awh
L
[kg/m2]
1/2
(A w A T )
total area of floor, ceiling and walls minus total window
area(s) [m2]
sum of window area(s) on all walls [m2]
window height or weighted average of window heights
on all walls [m]
fire load [kg]
Radiation from the projected flame

The temperature of a flame projected from a window is not constant through the full
extent of the flame. An average flame temperature could be used, but in order to
achieve as precise results as possible, the temperature is varied along the flame axis
and the flame is divided into strips with a height h. The total received radiation Pflame
is calculated with Equation 6.7 as a sum of the radiation Pflame,p from each strip.
[6.7]
Pflame = Pflame , p
The radiation from each strip is calculated with Equation 6.8.
Pflame , p = p T flame , p Ta
where:
p
Tflame,p
Ta
=
=
=
=
=
[6.8]
emissivity of radiator
configuration factor of relevant flame strip
average temperature of flame strip [K]
temperature of receiving surface [K]
The height, z1, of the flame tip is calculated with Equation 6.9 (Equation 4 in Law et
al. 1981).
R
z1 + H = 12.8
W
where:
z1
H
W
R
=
=
=
=
2/3
flame length above top of window [m]

height of window [m]
width of window [m]
rate of weight loss of fuel [kg/s]
-72-
[6.9]
The emissivity of the flame is calculated with Equation 6.10 (Equation 12 in Law et
al. 1981).
[6.10]
= 1 e 0.30
where:
flame thickness, i.e. 2H/3 [m]
The configuration factor tot,p of each flame strip can be determined with Equation
6.11. In order for configuration factors to be additive, they have to be calculated from
the same reference point, O, which is shown in Figure 6.15.
[6.11]
tot , p = 2 p 2 p 1
where:
p-1
configuration factor of the rectangle with height =

H+ph and width = W/2
configuration factor of the rectangle with height =
H+(p-1) h and width = W/2
number of the relevant flame strip counted from the
soffit
p
3
2
1
h
h
h
H/2
W/2
Figure 6.15
W/2
Rectangles used when determining the configuration factor of each flame strip
The configuration factor p should be calculated to the corner of the rectangle with
width W/2 and height depending on the flame strip in concern. This is done with
Equation 6.12.
p =
Where:
x
y
H
W
r
p
y
1 x
tan 1
2
360 1 + x 2
1+ x
=
=
=
=
=
=
y
1
+
tan
1+ y2
1+ y2
[6.12]
(H+ph)/2r
W/2r
height of rectangle [m]
width of rectangle [m]
distance between radiating and receiving surface [m]
number of the relevant flame strip counted from the
soffit
-73-
The distance between the radiator and receiver is equal to the distance between the
facade and receiver, minus the thickness of the flame, i.e. 2H/3.
The flame temperature varies along the flame axis. The temperature is assumed to be
constant through the full width and thickness of the flame, which is a conservative
approach. The temperature in the beginning and end of each flame strip can be
determined to give the average temperature in the strip. The temperature varies along
the flame according to Equation 6.13 (Equation 10 in Law et al. 1981).
Tz Ta
lw
= 1 0.027
To Ta
R
where:
Tz
Ta
To
l
w
R
=
=
=
=
=
=
[6.13]
temperature at distance l along flame axis [C]

temperature of ambient air = 20 C
temperature at the window [C]
distance along flame axis [m]
window width [m]
rate of burning [kg/s]
The rate of burning R (in kg/min) is calculated with Equation 6.14, which was
originally derived by Kawagoe, (Drysdale 1985). Other appropriate methods of
determining the burning rate may also be used.
R = 5.5 Ao H
where:
Ao
H
=
=
[6.14]
area of opening [m2]

height of opening [m]
The temperature at the window can be found since it is known that the temperature at
the flame tip is 540 C and that the total length of the flame axis is approximately
H/2+z1.
Total level of received radiation
The total level of received radiation ,Ptot, can be calculated by adding the radiative
parts of the window and the projected flame , see Equation 6.15.
Ptot = Pwindow + Pflame
[6.15]
Applications
The model presented above has been applied on four relatively simple cases.
Calculations have been made for case 1a-b that was used in chapter 6.3.1. An extra
calculation has been made for case 1a where the window is smaller, i.e. 1.51.5 m
instead of 31.5 m. One calculation has also been made for case 4 but where only one
of the original two windows exists. Table 6.1 below sets out the inputs that have been
used and the calculated values of temperatures and radiation intensities. The same fire
loads, which were given as energy contents per square metre, have been used for each
case as was done in the previous study. The fire load needs to be expressed in
-74-
kilograms when determining the compartment temperature according Equation 6.6,

and therefore has the total area of the enclosure and a calorific value of 18 MJ/kg
(wood) been used to convert the fire load into the right unit. The portion of the total
radiation that is contributed by the projected flame is also given.
Table 6.1
Properties and results from the sample calculations
Case 1a
Case 1b
Compartment (m)
54.52.4
54.52.4
Window (m)
31.5
31.5
Fire load (kg)
629
1052
-1/2
15.62
15.62
(m )
31.96
53.44
(kg/m2)
977
1137
Tf (C)
1072
1072
To (C)
Flame thickness (m)
1
1
Flame height* (m)
2.4
2.4
Radiating distance (m)
5
5
Pwindow (kW/m2)
7.4
11.9
Pflame (kW/m2)
1.6
1.6
Ptotal (kW/m2)
9.0
13.6
Pflame/ Ptotal
0.18
0.12
*The flame height is counted from the soffit.
Case 1c
54.52.4
1.51.5
629
32.06
44.60
927
1072
1
2.4
3
8.6
1.9
10.5
0.18
Case 4
6.35.53
12
972.9
48.83
98.45
865
957
1.3
3.2
3
6.1
1.2
7.3
0.16
It can be seen in the table above that the difference between the total levels of
received radiation and the radiation from the window was found to be in the range
1.2-1.9 kW/m2. The radiative portion from the projected flame was in the order of 1218 %. In increase in fire load resulted in higher temperature in the compartment and
hence an increased level of received radiation. It was also found that the radiative
portion from the flame decreased as the fire load was increased.
6.3.3 Discussion of results
The difference in received radiation predicted by the method by Fredlund et al. (1976)
and a notional radiator with projection distance 0 m for the studied cases varies
between 1.4-2.0 kW/m2. Note that this difference occurred around the area of interest
for building separation and ignition purposes, i.e. the distance where the level of
radiation is approximately 12.5 kW/m2. In this area it was always the method with
projected flames that predicted the highest radiation intensities. The largest difference
in levels of received radiation, i.e. 2.0 kW/m2, occurred for high and narrow windows.
The received radiation from the projected flame decreased with distance and become
less than the radiation for a non-projected notional radiator. This is due to that the
effective flame area viewed from the receiver also decreases with distance, but it also
gives an indication that the flame shields some of the radiation emitted from inside the
compartment.
The assumption of notional radiators at distances 1.2, 1.52 and 2.0 m appears to
overly estimate the levels of received radiation.
-75-
An increase in fire load caused a slightly higher fire temperature with both the method
by Fredlund et al. (1976) and the hand calculation method presented, and accordingly
higher levels of emitted and received radiation. However, this is only valid to a certain
extent since the fire will become ventilation controlled with a high enough increase in
fire load. The hand calculations predicted that the percentage of radiation from the
projected flame decreased when the fire load increased.
The calculations performed using the method by Fredlund et al. (1976) and the
presented hand calculation method, show only a small difference between taking into
account for flame projections and using non-projected notional radiators with the
compartment temperature and an emissivity of = 1. The contribution from the flame
to the total emitted radiation in the area of interest appears to be in the order of 1218 %.
The use of the notional radiator is an acceptable assumption that can be used for
building separation purposes with reasonable accuracy, which also has been
concluded by Law (1968) as presented in Chapter 6.1. It should be noted that all
possible compartment configurations, window dimensions and fire loads have not
been tested and a higher percentage of radiation from projected flames may occur for
a case not identified by the author.
-76-
Discussion and Conclusions
7 DISCUSSION AND CONCLUSIONS

The purpose of this chapter is to discuss findings made in this report and to make
conclusion based on the work performed for the fulfilment of this report. Some
recommendations will be made with regards to how different parameters and methods
should be used.
7.1
Radiation
Ignition due to radiation is the most common and thus hazardous way for a fire to
spread between adjoining buildings. There are other ways for fire spread to occur,
such as flying brands and convective heat transfer, but they are often disregarded in
the design process of spatial separations of buildings.
Two very important parameters with regards to determining safe separation distances
between buildings are;
a) the predicted fire temperature in the compartment, and
b) the levels of emitted radiation from the building.
Generally used values of the expected emitted radiation are 84 and 168 kW/m2 (2 and
4 cal/cm2s) depending on the type of facility. These values have been derived from
several full scale fire tests.
There are a numbers of methods to determine the fire temperature in a burning
compartment. It is up to the designer to determine which method that is best
applicable for the specific building in concern. The fire temperature is highly
dependent on the following factors;
the dimensions of the compartment,

the opening factor,
the size and distribution of the fuel, and
interior linings.
Flames will project out of openings in a compartment during a severe fire. The height,
z1, of the external flame above the top of the window is generally determined by the
use of Equation 9.1, where the flame tip is defined as the point where the flame
temperature is 540 C.
R
z1 + H = 12.8
W
2/3
[9.1]
H is the height of the opening, W is the width of the opening and R is the mass
burning rate in the compartment.
The transfer of radiation between a building on fire and an adjoining building depends
on several factors. Some parameters that have an influence are the dimensions and
shape of the radiating source, emissivity of radiating and receiving surfaces, as well as
the distance between the buildings.
-77-
Ignition of a radiation-exposed building can occur either spontaneously or piloted,

depending on whether an ignition source is present or not. Piloted ignition should be
considered in fire situations since ignition sources, such as sparks and flames, most
certainly are present. The level of received radiation that is commonly used as ignition
criteria for piloted ignition is 12.5 kW/m2. However, this value is adopted from
laboratory tests and findings presented in this report have indicated that the value of
critical received radiation that will cause ignition in full scale fire tests conducted in
an outdoor environment, may be in the range 15-18 kW/m2.
7.2
Comparison of Building Codes
The comparison of different Building Regulations and acceptable solutions showed

that performance based design solutions can be used in Australia, England and Wales,
New Zealand and Sweden. Canada and the United States of America still use strictly
prescriptive regulations with regards to spatial separation of buildings.
In general, the separation requirements are based on findings from the so-called St.
Lawrence Burns or work conducted by Law (1963). Values used for the maximum
levels of radiation that can be expected in a fire are 84 and 168 kW/m2 (2 and 4
cal/cm2s). In opposite to the other Building Codes studied in this report, the National
Building Code of Canada does not have a table with separation values for buildings
classed as having a light fire severity, i.e. a configuration factor of 0.14-0.15.
A widely used method in the regulations is to determine the distance to the relevant
boundary rather than to an adjoining building. Then it is assumed that an identical
building will be located at equal distance from, but on the other side of the relevant
boundary. This method is often referred to as the mirror image concept. However, a
hazardous situation can occur where the mirror image concept has been applied, but
where one of the buildings is greater or classed in a different purpose group.
The National Building Code of Canada and the NFPA 80A (1996) are the only codes
requiring that account should be taken to flames projected from openings. The flame
projection is not only incorporated in these codes with respect to the emitted radiation
from the flames, but also in order to guard against direct flame impingement.
7.3
Comparison of calculation methods
Three of the calculation methods studied in this report are incorporated or referred to
by Building Codes. The Enclosing rectangle method and the Aggregate notional area
method (Fire research Station 1991) are set out in the English and Welsh Approved
Document B (1991) and the New Zealand Approved Document C3 (BIA 1995), while
the method by McGuire (1965) is used in the NFPA 80A (1996) and the National
Building Code of Canada.
Most of the calculation methods studied in this report were developed in order to be
applicable to people without any particular knowledge in fire engineering. However,
the methods by Collier (1996) and Barnett (1988) require some extent of expert
judgement when it comes to determining the fire temperature and the flame
projection. Furthermore, these methods do not specify exactly how every possible
-78-
case should be treated, which enables the designer to treat the building in the most
appropriate way.
Even though the Enclosing rectangle method and the Aggregate notional area method
(Fire Research Station 1991) do not allow the user much flexibility when determining
the separation distance, the methods can be combined and used with great benefit for
complicated building shapes. One disadvantage may be that the Aggregate notional
area method is very time demanding to use.
The method by Williams-Leir (1970) does not require any expert judgement by a fire
engineer in order to perform the design. However, the rules set out are hard to
understand and very difficult and time demanding to apply on complex building
shapes.
The method by McGuire (1965) is easy to use when determining separation distances
for simple building shapes with evenly distribution of openings. One disadvantage is
that no table with separation distances is given for buildings classed as having a light
exposure hazard.
The NFPA 80A is a user-friendly method that can be used for most cases. However,
nothing is said of how buildings with recesses and set backs should be treated. The
method has a great flexibility and account can be taken for many influencing factors,
such as sprinklers and other fire safety measures, Fire Service intervention and actual
fire load in the building.
All the methods use rectangles that enclose unprotected areas when determining the
separation distance between buildings. It was noticed when the methods were applied
on different building types, that the methods predict approximately the same
separation/boundary distance for simple building shapes. The observed differences
could be referred to the various flame projection distances used in the methods. A
more complicated building shape generated larger differences in the separation
distances. This is as well caused by the flame projection distances, but also by the
limitations in applicability in the different methods.
7.4
Projected flame or notional radiator
The comparison between different calculation methods for determining safe

separation distances between buildings showed that there is an inconsistency in
whether to account for flame projections out of openings in a fire compartment or not.
This report has shown, both with performed calculations and in reviews of work
conducted earlier by others, that the levels of radiation contributed by the flames can
be neglected compared to the total amount of radiation received by an adjoining
building.
The calculations were performed with two different methods, which both showed
similar results, i.e. 12-18 % of the total received radiation were contributed by the
flames. Note that this value only is valid in the area of concern for piloted ignition, i.e.
12.5 kW/m2. The increase in total received radiation due to projected flames is not
large enough for it to be necessary to take into account. For the purpose of building
-79-
separation design, it is therefore only needed to consider the openings as radiators

with an appropriate chosen temperature or radiation.
The use of notional radiators located at various distances from the facade tends to
overestimate the levels of received radiation and will thereby cause a separation
distance that is too conservative. However, in the design process it may be necessary
to use flame projection equations to make sure that flame impingement and flame
contact with the adjoining building will not occur.
The hand calculation method presented in this report of how to calculate the emitted
radiation from the fire compartment and the externally projected flames can be used
with a reasonable result. Note that the method is valid for no-wind conditions.
-80-
Further research
8 FURTHER RESEARCH
As this project has progressed, a number of parameters have been identified that
influence horizontal fire spread to adjoining buildings. The importance of and the
significance of some these parameters are however not thoroughly investigated and
further research is needed in certain areas in order to achieve reliable methods of
determining safe separation distances. Below is listed a number of suggested areas to
further investigate in the future.
Even though this and other reports has shown that no account needs to be
taken to projected flames with regards to radiative heat transfer, it should be
investigated more thoroughly. It may well be so, that a situation, e.g. small
and narrow windows, not studied in detail may cause extraordinary high
levels of radiation.
Most of the research on which the Building Regulations and calculation

methods of today are based, has been derived from old fire tests. The
building materials and interior furnishings used in those days are very
different from the material used today. Today, the materials are more
complex and generally contain more energy and it is therefore essential to
investigate what levels of radiation and temperatures that can be expected in
a fire in a modern building.
The sample calculations performed in this report for a warehouse, indicated

that separation distances predicted for large warehouses will be very large
and may be too conservative. It should be investigated how much of the
facade that can be expected to radiate at the same time.
Full scale fire tests reported on in this project have shown that the levels of
radiation required to cause ignition of a neighbouring building actually was
higher than the commonly used value of 12.5 kW/m2. The ignition criteria
should therefore be reinvestigated in full scale fire tests conducted in an
outdoor environment.
Wind effects have not been addressed in this report. An interesting research
project in the future would be to investigate the influences of external wind
on the total amount of emitted and received radiation.
-81-
Further research
-82-
References
9 REFERENCES
Australian Building Codes Board, Building Code of Australia, Volume One, October
1996
Barnett C.R., Fire Separation between External Walls of Buildings, Fire Safety
Science, Proceedings of the 2nd International Symposium (pp. 841-850), 1988
Boverket, Boverkets byggregler, BBR 94:3, Norstedts Tryckeri, Karlskrona, June
1995, In Swedish
Building Industry Authority, Approved Document C3: Spread of Fire, New Zealand,
December 1995
Centre for Advanced Engineering, Fire Engineering Design Guide, A.H. Buchanan
Editor, University of Canterbury, July 1994
Clarke James M.W., A review of the building separation requirements of the New
Zealand Building Code Acceptable Solutions, University of Canterbury, Christchurch,
1998
Department of the Environment and The Welsh Office, Approved Document B, Fire
safety, The Building Regulations 1991
Drysdale Dougal, An Introduction to Fire Dynamics, Department of Fire Safety
Engineering, University of Edinburgh, 1987
Fire Research Station, External Fire Spread: Building separation and boundary
distances, Edited by Read R.E.H., Building Research Establishment Report BR 187,
Borehamwood, 1991
Fredlund Bertil, Magnusson Sven Erik, Nilsson Leif, Pettersson Ove, Strandberg
Sven, Thelandersson Sven, Skydd mot brandspridning inom smhusbebyggelse i
lttbetong, Svenska Brandfrsvarsfreningen, Uppsala, 1976, In Swedish
Jnsson R., Frantzich H., Karlsson B., Magnusson S.E., Ondrus J., Pettersson O.,
Bengtsson S., Osterling T., Thor J., Brandskydd Teori och Praktik,
Brandskyddslaget, LTH-Brandteknik, Stockholm, 1994, In Swedish
Karlsson Bjrn, Quintiere James G., Enclosure Fire Dynamics, Department of Fire
Safety Engineering, Lund University, 1998
Law Margaret, Heat radiation from fires and building separation, Fire Research
Technical Paper No. 5, Joint Fire Research Organization, 1963
Law Margaret, Radiation from fires in a compartment, Fire Research Technical Paper
No. 20, Ministry of Technology and Fire Offices Committee, 1968
Law Margaret, Thomas P.H., The projection of flames from buildings on fire, Fire
Prevention Science and Technology, No. 10, 1974
-83-
References
Law Margaret, OBrien Turlogh, Fire safety of bare external structural steel,
Constrado, Croydon, May 1981
National Fire Protection Association, NFPA 80A Recommended Practice for
Protection of Buildings from Exterior Fire Exposures, Quincy, MA, 1996
Nordiska industrigruppen trhus/brandskydd,
brandspridningsrisker, June 1975, In Swedish
Tt
smhusbebyggelse
och
Ondrus Julia, Pettersson Ove, Fire hazards of facades with externally applied
additional thermal insulation. Full scale experiments, Lund Institute of Technology,
Division of building fire safety and technology, Lund, 1986
Magnusson S.E., Thelandersson S., Temperature-Time Curves for the Complete
Process of Fire Development A Theoretical Study of Wood Fuel Fires in Enclosed
Spaces, Acta Polytechnica Scandinavica, Ci 65, Stockholm, 1970
Magnusson S.E., Thelandersson S., Comments on Rate of Gas Flow and Rate of
Burning of Fires in Enclosures, Division of Structural Mechanics and Concrete
Mechanics, Lund Institute of Technology, Bulletin No. 19, Lund, 1971
Magnusson S.E., Thelandersson S., A Discussion of Compartment Fires, Fire
Technology, Vol. 10, No. 3, August 1974
McGuire J H, Fire and the Spatial Separation of Buildings, Fire Technology Vol. 1
No. 4, 1965
Pettersson O., Magnusson S.E., Thor J., Fire engineering design of structures,
Swedish Institute of Steel Construction, Publication 50
Smith P.G., Webster C.T., The burning of well ventilated compartment fires. Part IV.
Brick compartment, 2.4 m cube, Joint Fire Research Organization, F.R. Note
578/1964
The American Society for Testing and Materials, Standard Test Method for
Determining Material Ignition and Flame Spread Properties, Reprinted with
permission from the Annual Book of ASTM Standards, May 1990
Thuresson S., Vindtunnelstudie i anslutning till brandfrsk gllande
grupphusbebyggelse, Avdelningen fr byggnadskonstruktion, CTH, Gteborg, 1973,
In Swedish
Williams-Leir G., Another Approximation for Spatial Separation, Fire Technology
Vol. 6 No. 3, 1970
-84-
Appendices
APPENDICES
-85-
Appendices
-86-
Appendix 1 Comparison of boundary requirements for warehouse
APPENDIX 1 COMPARISON
OF BOUNDARY REQUIREMENTS FOR
WAREHOUSE
The calculation methods and all used equations and terms are presented earlier in this
report. The following chapter only shows how the methods are used and what
separation distances that is predicted by each method.
C.R. Barnett (1988)
By calculating the configuration factor of the enclosing rectangle, the separation
distance between the buildings can be determined. At first, the fire temperature has to
be determined. The ISO 834 time-temperature curve gives for duration of 120 minutes
the following temperature:
T2 = 345log10(8120+1)+20 = 1049 C
The configuration factor can then be calculated with equation A1.1:
n =
AE I RC
AV I EC
[A1.1]
where:
AE = 580 m2 = 400 m2
AV = 580 m2 = 400 m2
IRC = 12.5 kW/m2
IEC = 156,710-12[(1049+273)4 2934] = 172.79 kW/m2
Which gives:
n = 0.07234
The radiation distance can be found by iteration of Equation 4.3, which is described in
the summary of this method by Barnett (Chapter 4.4). The width of the enclosing
rectangle is 50 m and the height is 8 m, which result in a radiation distance of 36.7 m.
Since the external walls are not fire resistance rated, another two metres should be
added to take into account for flame projection giving a total radiation distance of
38.7 m. The distance to the relevant boundary is half the total radiating distance and
the required boundary distance will in this case be 19.4 m.
J.H. McGuire (1965)
This calculation will follow the three main steps that are explained in the summary of
J.H. McGuires method (Chapter 4.6).
1. Height and width of fire compartment.
H = 8 m (26 ft)
W = 50 m (164 ft)
-87-
2. Percent of window openings in the building facade.

The external walls are not fire resistance rated 100 % opening
3. Since the interior walls are fitted with non-combustible linings and the building
therefore is considered as normal, Table 3 in McGuire, 1965, can be applied.
Interpolation gives a radiating distance of 38.5 m (126.2 ft). Thus, the distance to
the relevant boundary is 19.2 m (63.1 ft). Note that a horizontal flame projection
distance of 1.52 m (5 ft) already is incorporated in the tables.
Peter Collier (1996)
Step 1: Determination of the enclosing rectangle.
The enclosing rectangle equals the whole building facade.
Step 2: Area of enclosing rectangle
H=8m
W = 50 m
Ae = 850 m2 = 400 m2
Step 3: Aspect ratio, AR.
AR = H/W
AR = 8/50 = 0.16
Step 4: Radiation intensity of the fire source, Is.
T = 345log10(8t+1)
T(120) = 345log10(8120+1) = 1029 C
Is = 5.6710-11T4
Is = 5.6710-11 (1029+273)4 = 163.0 kW/m2
Step 5: Reduction factor, Rf.
Rf = Ao/Ae
Rf = 400/400 = 1
Step 6: Determination of the emitted radiation, Ie.
Ie = IsRf
Ie = 163.01 = 163.0 kW/m2
Step 7: Critical incident radiation, Icr.
Icr = 12.5 kW/m2
Step 8: Determination of the permissible configuration factor, .
= Icr/Ie
= 12.5/ 163.0 = 0.07669
Step 9: Determination of the separation distance, S.
S = R+P
P = 2 m (projection distance of flames)
R = CAe
C = 1,94 (obtained from Figure 3)
-88-
R = 1.94400 = 38.75 m
S = 38.75+2 = 40.75 m
Accordingly, the required distance to the relevant boundary with this method is
20.4 m.
G. Williams-Leir (1970)
The dimensions of the exposing building face (EBF) are:
Width = 164.04 ft
Height = 25.25 ft
AEBF = 164.0426.25 = 4306.05 ft2
Applying Rule 1 gives:
20.44(L-3)2 = AEBF
20.44(L-3)2 = 4306.05 ft
L = 72.95 ft
Check Rule 2:
K = 3.54(L-3)
K = 3.54(72.95-3) = 247,62 ft > 164,04 ft Rule 2 does not apply
Check Rule 3:
J = 3.14(L-3)
J = 3.14(72.95-3) = 219.64 ft > 164,04 ft Rule 3 does not apply
The limiting distance or the distance to the relevant boundary is 22.2 m (73.0 ft).
Enclosing rectangle (Fire Research Station 1991)
Step 1: Determination of unprotected areas.
The whole building facade should be considered as an unprotected area.
Step 2: Establishing a plane of reference
The plane of reference should be located so that it touches all parts of the
building facade.
Step 3: Determination of the area of exposure.
The whole facade of the building should be considered as the area of exposure,
i.e. the enclosing rectangle.
Width = 50 m
Height = 9 m
Unprotected area = 508 = 400 m2
Area of enclosing rectangle = 509 = 450 m2
Unprotected percentage = 100400/450 = 88.89 %
Step 4: Determination of boundary distance.
Table 1 in Fire Research Station (1991), gives for Storage purposes a
minimum distance to the relevant boundary of 18.3 m.
-89-
NFPA 80 A (1996)
The width and height of the exposing fire are the same as the facade dimensions, i.e.
164.04 and 26.25 ft respectively. The fire is assumed to penetrate the exterior walls
within 20 minutes, giving 100 % openings in the exposing wall. Warehouses and
storage facilities can be assumed to have a moderate fire severity, which equals the
same configuration factor as used by the other methods for warehouses. In order to
find the appropriate guide number g, from Table 4.4, the ratio between the height and
width of the exposing fire needs to be calculated.
W/H = 164.04/26.25 = 6.25
Interpolation in Table 4.4 gives the guide number g, as follows:
g = 4.7925
Note that Table 4.4 already is based on a critical incident radiation of 12.5 kW/m2 and
therefore, no further modifications needs to be done with regards to ignition criteria.
The separation distance between the buildings can now be determined as explained
below:
D = gZ+5
D = 4.7926.26+5 = 130.80 ft (39.87 m)
Hence, the required distance to the relevant boundary is 19.9 m (65.4 ft) and in this
distance, account is taken for horizontal flame projection of 1.52 m.
-90-
Appendix 2 Comparison of boundary requirements for office building
APPENDIX 2 COMPARISON OF BOUNDARY REQUIREMENTS FOR OFFICE

BUILDING

separation distances and boundary distances that are predicted by each method.
In those methods that use enclosing rectangles, the rectangles have been named
according to the following:
Rectangle 1:
Rectangle 2:
Rectangle 3:
Rectangle 4:
Rectangle 5:
Rectangle 6:
Rectangle 7:
The rectangle that encloses only one window

The rectangle that encloses two windows
The rectangle that encloses three windows
The rectangle that encloses four windows
The rectangle that encloses five windows
The rectangle that encloses six windows
The rectangle that encloses all seven windows
C.R. Barnett (1988)

The ISO 834 time-temperature curve is used to determine the temperature in the fire
compartment.
T2 = 345log10(8tm+1)+T1
T2(30) = 345log10(830+1)+20 = 842 C
Knowing the fire temperature allows the radiant heat flux inside the compartment to
be determined.
IEC = (T24- T14)
IEC = 156,710-12[(842+273)4 2934] = 87.15 kW/m2
Seven different enclosing rectangles are of interest with regards to determining the
minimum required boundary distance. The rectangle resulting in the largest distance
should be chosen. Table A2.1 below sets out relevant properties of the enclosing
rectangles, configuration factors and eventually the boundary distance. The
configuration factors are determined with equation A2.1.
n =
AE I RC
AV I EC
[A2.1]
The radiating distance, i.e. the distance to where the received radiation is 12.5 kW/m2,
has been found by iteration of Equation 4.3. Equation 4.3 is described in the summary
of this method by Barnett (Chapter 4.4).
-91-

Table A2.1
Relevant measures and boundary distances for the enclosing rectangles
Properties
Width (m)
Height (m)
Area of openings,
Av (m2)
Area of enclosing
rectangle, Ae (m2)
Configuration
factor,
Radiating distance,
R (m)
Flame projection,
P (m)
Boundary distance,
B (m)
1
3
2
2
9
2
Enclosing rectangle
3
4
5
15
21
27
2
2
2
12
18
24
30
36
42
18
30
42
54
66
78
0.1434
0.2151
0.2391
0.2510
0.2582
0.2630
0.2664
3.4
3.9
3.9
3.8
3.7
3.7
3.6
2.7
3.0
3.0
2.9
2.9
2.9
2.8
6
33
2
7
39
2
The largest and thus the minimum required boundary distance is 3.0 m. This distance
is caused by the rectangles that enclose two and three windows in the storey.
If the designer had chosen a fire duration time of 120 minutes, like what was done for
the warehouse, rather than 30 minutes, the required boundary distance would have
been 4.7 m. The rectangle that encloses all seven windows in the storey would be the
rectangle resulting in the largest boundary distance.
J.H. McGuire (1965)
1. Determination of height and width of fire compartment
H = 12.47 ft (3.8 m)
W = 131.23 ft (40 m)
2. Percentage of openings in the facade
Av
7 3 2
100 =
100 = 26.18 %
Ae
3.8 40
3. The office is assumed to be fitted with non-combustible linings and therefore, the
building is classed as a normal case building and has a configuration factor of
0.07. Interpolation of Table 3 in McGuire (1965), sets the building separation to
32.06 ft (9.77 m) and accordingly the boundary distance to 4.9 m (16 ft). Note that
a five feet horizontal flame projection distance is included in the boundary
distance.
With this method it is interesting to investigate the same seven enclosing rectangles as
in the method by Barnett (1988). The plane of reference is located so that it touches
the whole length of the building facade. Table A2.2 sets out Step 2-9 for the seven
rectangles.
-92-
Note that in this method by Collier (1996) and in difference to the method by Barnett
(1988), the temperature of ambient air is neglected in the calculations of the fire
temperature and radiation intensity.
Table A2.2
Step
2
3
4
5
6
7
8
9
Determination of boundary distance
Properties
Width (m)
Height (m)
Area
of
enclosing
rectangle, Ae (m2)
Aspect ratio,
AR
Fire temperature,
T (C)
Radiation intensity,
Is (kW/m2)
Reduction factor,
Rf
Emitted radiation,
Ie (kW/m2)
Critical incident radiation,
ICR (kW/m2)
Configuration factor,
Projection distance,
P (m)
Factor from Figure 4.3,
C
Radiating distance,
R (m)
Separation distance,
S (m)
Boundary distance,
B (m)
1
3
2
2
9
2
Enclosing rectangle
3
4
5
15
21
27
2
2
2
6
33
2
7
39
2
18
30
42
54
66
78
0.667
0.222
0.133
0.095
0.074
0.061
0.051
822
822
822
822
822
822
822
81.45
81.45
81.45
81.45
81.45
81.45
81.45
0.667
0.600
0.571
0.556
0.545
0.538
81.45
54.33
48.87
46.51
45.29
44.39
43.82
12.5
12.5
12.5
12.5
12.5
12.5
12.5
0.153
0.230
0.256
0.269
0.276
0.282
0.285
1.34
0.83
0.68
0.53
0.46
0.40
0.35
3.28
3.52
3.72
3.43
3.38
3.25
3.09
5.28
5.52
5.72
5.43
5.82
5.25
5.09
2.6
2.8
2.9
2.7
2.7
2.6
2.5
The minimum required boundary distance is 2.9 m. The largest distance is caused by
the rectangle that encloses three windows in the storey.
If the designer had chosen a fire duration time of 120 minutes, like what was done for
the warehouse, rather than 30 minutes, the required boundary distance would have
been 4.7 m. The rectangle that encloses all seven windows in the storey would be the
rectangle resulting in the largest boundary distance in that case.
Enclosing rectangle (Fire Research Station 1991)
The same enclosing rectangles will be investigated with this method as has been done
with previous methods. Table A2.3 below gives the necessary properties of Step 3-4
-93-
and eventually the boundary distance from Table 1 in Fire Research Station (1991).
Note that the enclosing rectangles must have dimensions according to Table 1 in Fire
Research Station (1991). Since the building is an office building, it belongs to the
Office purpose group, which is considered to have a low fire load density.
Table A2.3
Step
3
Determination of boundary distances for residential building
Properties
Width,
W (m)
Height,
H (m)
Unprotected area,
Av (m2)
Area of enclosing
rectangle, Ae (m2)
Unprotected percentage
Boundary distance,
B (m)
Enclosing rectangle
3
4
5
15
21
27
40
40
12
18
24
30
36
42
27
45
63
81
120
120
66.7
44.4
40
38.1
37.0
30
35
1.5
1.72
2.0
1.90
1.85
1.5
1.75
With this method, the required boundary distance is 2.0 m and is caused by the
rectangle that encloses three windows.
NFPA 80A (1996)
Once again, five enclosing rectangles have been investigated and the calculation
procedure can be followed in Table A2.4. An extra calculation has also been made for
the whole extent of the fire compartment, i.e. not only the smallest rectangle that
encloses all openings. This extra rectangle is labelled Rectangle 8. All windows are
considered to be 100-percents openings. An office building does not generally contain
35 kg of combustible material per metre square or more and the building is therefore
classed as having a light exposure severity and the building will then have the same
configuration factor as is used in the other methods for office buildings. The guide
number g, is obtained from Table 4.4. Since the critical incident radiation already is
set to 12.5 kW/m2, no further compensation needs to be done with regards to this.
-94-

Table A2.4
Boundary distances determined with NFPA 80A (1996)
Properties
Width of exposing fire,
W (m)
Height of exposing fire,
H (m)
Area of openings,
Av (m2)
Area of enclosing
rectangle, Ae (m2)
Percent of openings
Width-height ratio,
W/H
Lesser dimension,
L (m)
Guide number,
g
S (m)
Boundary distance,
B (m)
Enclosing rectangle
3
4
5
6
15
21
27
33
39
40
3.8
12
18
24
30
36
42
42
18
30
42
54
66
78
152
100
66.7
60
57.1
55.6
54.5
53.8
27.6
1.5
4.5
7.5
10.5
13.5
16.5
19.5
10.53
3.8
1.68
1.96
1.98
1.93
1.90
1.87
1.85
0.838
4.88
5.43
5.47
5.39
5.32
5.27
5.22
4.70
2.4
2.7
2.7
2.7
2.7
2.6
2.6
2.4
The largest and hence the required boundary distance predicted by this method is 2.7
m and is caused by the rectangle that encloses three widows in the storey.
-95-
-96-
Appendix 3 Comparison of boundary requirements for residential building
APPENDIX 3 COMPARISON
OF BOUNDARY REQUIREMENTS FOR
RESIDENTIAL BUILDING

separation distances and boundary distances that are predicted by each method.
In those methods that use enclosing rectangles, the rectangles have been named
according to the following:
Rectangle 1:
Rectangle 2:
Rectangle 3:
Rectangle 4:
Rectangle 5:
The smallest rectangle that encloses all openings

The rectangle that encloses both openings in the recess
The rectangle that encloses the carport in the recess
The rectangle that encloses the door and the two windows closest to it
The smallest rectangle that encloses all openings in the front facade,
i.e. the recess not included
Rectangle 6: The rectangle that encloses the two windows to the left of the recess
Rectangle 7: The rectangle that encloses the largest window in the front facade, i.e.
the recess not included
C.R. Barnett (1988)
The ISO 834 time-temperature curve is used to determine the temperature in the fire
compartment.
T2 = 345log10(8tm+1)+T1
T2(120) = 345log10(8120+1)+20 = 842 C
Knowing the fire temperature allows the radiant heat flux inside the compartment to
be determined.
IEC = (T24- T14)
IEC = 156,710-12[(1049+273)4 2934] = 87.15 kW/m2
Seven different enclosing rectangles are of interest with regards to determining the
minimum required boundary distance. The rectangle resulting in the largest distance
should be chosen. Table A3.1 below sets out relevant properties of the enclosing
rectangles, configuration factors and eventually the boundary distance. The
configuration factors are determined by Equation A3.1.
n =
AE I RC
AV I EC
-97-
[A3.1]
The radiating distance, i.e. the distance to where the received radiation is 12.5 kW/m2,
has been found by iteration of Equation 4.3. Equation 4.3 is described in the summary
of this method by Barnett (Chapter 4.4).
Table A3.1
Relevant measures and boundary distances for enclosing rectangles used for the
residential building
Properties
Width (m)
Height (m)
Area of openings,
Av (m2)
Area of enclosing
rectangle, Ae (m2)
Configuration
factor,
Radiating distance,
R (m)
Flame projection,
P (m)
Boundary distance,
B (m)
1
19
6.6
2
4
6.2
Enclosing rectangle
3
4
5
4
4.9
14
3
6.2
6.6
34.98
16.5
12
11.26
18.48
7.22
4.5
125.4
24.8
12
30.38
92.4
10.26
4.5
0.5142
0.2156
0.1434
0.3870
0.7171
0.2038
0.1434
5.1
5.3
4.8
3.9
3.0
3.3
2.9
3.6
3.7
3.4
3.0
2.5
2.7
2.5
6
1.9
5.4
7
3
1.5
The largest distance is 3.7 m and is caused by the rectangle that encloses both
openings in the recess. However, since the recess is set back 1.5 m, the distance
between the relevant boundary and the front facade is instead 3.0 m resulting from the
rectangle that encloses the door and the two windows closest to it. If the recess is
assumed not to exist, the rectangle that encloses all openings would call for a
boundary distance if 3.6 m.
If fire duration of two hours had been used instead, the required boundary distance
would be 5.9 m. This distance is caused by the rectangle that encloses all unprotected
openings in the facade.
J.H. McGuire (1996)
This method is derived for building facades with even distribution of openings.
Therefore, reliable results cannot be obtained when this method is applied to the
building used in the comparison. A sample calculation has even though been made
strictly for comparison purposes. Step 1-3 outlines the calculation procedure. Since
the building is irregular shaped, all openings are assumed to be projected onto a plane
of reference. The plane of reference joins the extremities of the facade.
1. Determination of height and width of fire compartment
H = 22.97 ft (7 m)
W = 65.62 ft (20 m)
-98-
2. Percentage of openings in the facade

Av
3 1.9 2 + 4 3 + 2 3 1.5 + 1.5 2.1
100 =
100 = 25.0 %
Ae
7 20
3. The office is assumed to be fitted with non-combustible linings and therefore, the
building is classed as a normal case building and has a configuration factor of
0.07. Interpolation of Table 3 in McGuire (1965), sets the building separation to
38.26 ft (11.66 m) and accordingly the boundary distance to 5.8 m (19 ft). Note
that a five feet horizontal flame projection distance is included in the boundary
distance and that the boundary distance should be counted from the plane of
reference.
Since some parts of the facade are located in front of the plane of reference, extra
calculations should be made for this section. However, the percentage of openings
within in this section is less than 25 %, which is the lowest value set out in the table.
Hence, no separate distance can be found for this section.
This method allows for unprotected areas to be projected onto a plane of reference for
buildings with irregular shaped facades. In this case, the plane of reference is located
so that it touches but not passes through both corners of the recess. The smallest
rectangle that encloses all projected unprotected areas in the plane of reference is
called Rectangle 8. The plane of reference makes an angle of 17 to the building
facade. The recess is dealt with according to methods outlined by Law (1963), which
means that calculations should be made for both Rectangle 1 and 8, and the boundary
distance be a combination of the two as shown in Figure A3.1. When determining the
boundary distance caused by Rectangle 1, the recess is assumed not to exist. To make
sure that no single opening or combination of openings call for a greater boundary
distance, calculations has also been made for the other six rectangles mentioned
above. Table A3.2 sets out Step 2-9 for the eight rectangles.
Plane of reference
Final boundary position
First boundary position

Second boundary position
Figure A3.1
Boundary distance for a recess according to Law (1963)
-99-
Note that in this method by Collier (1996) and in difference to the method by Barnett
(1988), the temperature of ambient air is neglected in the calculations of the fire
temperature and radiation intensity.
Table A3.2
Step
2
3
4
5
6
7
8
9
Properties
Width (m)
Height (m)
Area of enclosing
rectangle, Ae (m2)
Aspect ratio,
AR
Fire temperature,
T (C)
Radiation intensity,
Is (kW/m2)
Reduction factor,
Rf
Emitted radiation,
Ie (kW/m2)
Critical incident
radiation, ICR (kW/m2)
Configuration factor,
Projection distance,
P (m)
Factor from Figure 4.3,
C
Radiating distance,
R (m)
S (m)
Boundary distance,
B (m)
Enclosing rectangle
4
5
6
19
6.6
4
6.2
4
3
4.9
6.2
14
6.6
1.9
5.4
3
1.5
18.6
6.6
125.4
24.8
12
30.4
92.4
10.3
4.5
122.8
0.347
0.645
0.750
0.790
0.471
0.352
0.500
0.355
822
822
822
822
822
822
822
822
81.45
81.45
81.45
81.45
81.45
81.45
81.45
81.45
0.279
0.665
0.371
0.200
0.704
0.285
22.72
54.14
81.45
30.21
16.28
57.32
81.45
23.21
12.5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
0.550
0.231
0.154
0.414
0.768
0.218
0.154
0.539
0.43
0.98
1.38
0.65
0.25
0.98
1.35
0.45
4.82
4.88
4.78
3.58
2.40
3.15
2.86
4.99
6.82
6.88
6.78
5.58
4.40
5.15
4.86
6.99
3.4
3.4
3.4
2.8
2.2
2.6
2.4
3.5
The minimum required boundary distance is 3.5 m and the shape of and distances to
the boundary is shown in Figure A3.1 above. No single opening or combination of
openings call for a greater separation distance. If the recess were assumed not to exist
the required boundary distance would be 3.4 m
If fire duration of two hours had been used instead, the required boundary distance
would be 5.3 m. This distance is caused by the rectangle that encloses all unprotected
openings in the facade and is also counted from the plane of reference.
-100-
Enclosing rectangle (Fire Research Station, 1996)

The same enclosing rectangles will be investigated with this method as has been done
with previous methods. Since the recess is only 1.5 m wide, the recess can be assumed
not to exist and the plane of reference positioned in a way that it touches the whole
front part of the facade. This method outlines that any parts of the facade that are set
back not greater than 1.5m is assumed not to be a set back at all. Table A3.3 below
gives the necessary measures of Step 3-4 and eventually the boundary distance from
Table 1 in Read, 1991. Note that the enclosing rectangles must have dimensions
according to Table 1 in Fire Research Station (1991). Since the building is a
residential building, it belongs to the Residential purpose group, which is considered
to have a low fire load density.
Table A3.3
Step
3
Properties
Width,
W (m)
Height,
H (m)
Unprotected area,
Av (m2)
Area of enclosing
rectangle, Ae (m2)
Unprotected percentage
Boundary distance,
B (m)
Enclosing rectangle
3
4
5
21
34.98
16.5
12
11.26
18.48
7.22
4.5
189
54
18
54
54
18
18.51
30.56
66.67
20.85
34.22
40.11
50
2.0
2.0
2.0
1.1
2.2
1.5
1.5
With this method, the required boundary distance is 2.2 m and is caused by the
rectangle that encloses all windows in the front facade.
NFPA 80A
Once again, seven enclosing rectangles have been investigated and the calculation
procedure can be followed in Table A3.4. An extra calculation has also been made for
the whole extent of the facade and the enclosing rectangle used for this is named
Rectangle 8. All windows are considered to be 100-percents openings. The building is
classed as having a light exposure severity in order to make the results from the
different calculation methods better comparable. The guide number g, is obtained
from Table 2-3 in NFPA 80A, 1996. Since the critical incident radiation already is set
to 12.5 kW/m2, no further compensation needs to be done with regards to this.
-101-
Table A3.4
Boundary distances for residential building determined with NFPA 80A (1996
Properties
Width of exposing fire,
W (m)
Height of exposing fire,
H (m)
Area of openings,
Av (m2)
Area of enclosing
rectangle, Ae (m2)
Percent of openings
Width-height ratio,
W/H
Lesser dimension,
L (m)
Guide number,
G
S (m)
Boundary distance,
B (m)
Enclosing rectangle
4
5
6
19
4.9
14
1.9
20
6.6
6.2
6.2
6.6
5.4
1.5
34.98
16.5
12
11.26
18.48
7.22
4.5
34.98
125.4
24.8
12
30.38
92.4
10.26
4.6
140
27.89
66.53
100
37.06
20
70.37
100
24.99
2.879
1.550
1.333
1.267
2.121
2.844
2.857
6.6
4.9
6.6
1.9
1.5
0.782
1.323
1.58
0.785
0.465
1.768
1.93
0.672
6.68
6.81
6.26
5.37
4.59
4.88
4.42
6.22
3.3
3.4
3.1
2.7
2.3
2.4
2.2
3.1
The largest distance predicted by this method is 3.4 m and arises from the rectangle
that encloses all the openings in the recess. However, since this rectangle is set back
1.5 m, the largest boundary distance resulting from the front facade is 2.7 m. If the
recess is assumed not to exist, the required boundary distance would be 3.3 m.
A boundary distance of 5.9 m would be required if the building was classed as having
a moderate exposure severity.
-102-
Appendix 4 Importance of radiation from projected flames
APPENDIX 4 IMPORTANCE OF RADIATION FROM PROJECTED FLAMES

This chapter contains the performed calculations for the comparison of radiation from
a compartment with projected flames according to Fredlund et al. (1976) and notional
radiators at various distances. The radiation from the notional radiators has been
calculated with the method discussed in Chapter 6 and the used temperatures are
obtained from Fredlund et al. (1976) for the specific compartment configuration. The
results are presented directly in the figures presented below.
Case 1a
1 window:
Fire load:
30 Mcal/m2 (125 MJ/m2)
Distance:
5m
Calculation procedure
1. Fire load f = 30 Mcal/m2 (125 MJ/m2)
2. A h / Atot = (31.5 1.5 )/(54.52+52.42+4.52.42) = 0.060 m1/2
3. f = 30 Mcal/m2 and A h / Atot = 0.060 m1/2 Table 20a-b Appendix 2 and timetemperature curve from Appendix 1 in Fredlund et al. (1976)
4. Atot,p = 92.5 m2
mp = 1
p = 4.5/4.5 = 1
Pmax,p = 11 (92.5/85) Pmax,t = 1.088 Pmax,t
c = 5 m Pmax,t = 9.3 kW/m2 Pmax,p = 1.0889.3 = 10.1 kW/m2
5. Figure A4.1 shows the time-radiation relationship at the receiving point for the
method by Fredlund et al. (1976) and notional radiators at various projection
distances.
25,00
20,00
15,00
N otional ra diator (P = 1,2m )

(kW /m ^2)

P rojec ted flam e
10,00
5,00
0,00
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
T im e (h )
Figure A4.1
Received radiation for case 1a
Case 1b
1 window:
-103-
50 Mcal/m2 (125 MJ/m2)

5m
Fire load:
Distance:
2. A h / Atot = (31.5 1.5 )/(54.52+52.42+4.52.42) = 0.060 m1/2
4. Atot,p = 92.5 m2
mp = 1
p = 4.5/4.5 = 1
Pmax,p = 11 (92.5/85) Pmax,t = 1.088 Pmax,t
distances.
25,00
20,00
15,00

Incident radiation
(kW /m ^2)

P rojected flam e
10,00
5,00
0,00
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
Tim e (h)
Figure A4.2
Received radiation for case 1b
Case 2
1 window:
Fire load:
50 Mcal/m2 (209 MJ/m2)
Distance:
5m
2. A h / Atot = ([321.3+4.21.3] 1.3 )/(8102+82.52+102.52) = 0.040
m1/2
4. Atot,p = 250 m2
-104-
mp = 1
p = (4.21.3)/(321.3+4.21.3) = 0.411
Pmax,p = 10.411 (250/85) Pmax,t = 1.209 Pmax,t
distances.
40,00
35,00
30,00
25,00

20,00
(kW /m ^2)

P rojec ted flam e
15,00
10,00
5,00
0,00
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
T im e (h )
Figure A4.3
Case 3
Fire load:
40 Mcal/m2 (167 MJ/m2)
Distance:
3m
2. A h / Atot = (621.3 1.3 )/(5102+52.52+102.52) = 0.10 m1/2
4. Atot,p = 175 m2
mp = 1
p = 1/6 = 0.167
distances.
-105-
90,00
80,00
70,00
60,00
50,00
P rojec ted flam e

(kW /m ^2)

40,00
30,00
20,00
10,00
0,00
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
T im e (h )
A4.4
Case 4
Compartment: 6.35.53.0 m (widthdepthheight)
2 windows: 12 m (widthheight) each
Fire load:
30 Mcal/m2 (125 MJ/m2)
Distance:
3m
2. A h / Atot = (221 2 )/(6.35.52+6.332+5.532) = 0.040 m1/2
4. Atot,p = 140.1 m2
mp = 1
p = 1/2 = 0.5
Pmax,p = 10.5 (140.1/85) Pmax,t = 0.824 Pmax,t
distances.
-106-
60,00
50,00
40,00

30,00
(kW /m 2)

P rojec ted flam e
20,00
10,00
0,00
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
T im e (h )
Figure A4.5
Case 5
4 windows: 1.451.45 m (widthheight) each
Fire load:
50 Mcal/m2 (125 MJ/m2)
Distance:
3m
2. A h / Atot = (41.451.45 1.45 )/(1082+102.52+82.52) = 0.040 m1/2
4. Atot,p = 250 m2
mp = 1
p = 1/4 = 0.25
distances.
-107-
70,00
60,00
50,00
40,00

(kW /m ^2)

30,00
P rojec ted flam e
20,00
10,00
0,00
0,00
0,20
0,40
0,60
0,80
1,00
1,20
T im e (h )
Figure A4.6
Case 6
1 window:
Fire load:
30 Mcal/m2 (125 MJ/m2)
Distance:
3, 4, 5, 6 and 8 m
2. A h / Atot = (1.521.52 1.52 )/(442+42.52+42.52) = 0.040 m1/2
4. Atot,p = 72 m2
mp = 1
p = 1
Pmax,p = 11(72/85)Pmax,t = 0.847 Pmax,t
distances.
-108-
14
12
10
8
P rojec ted flam e

(kW /m ^2)
6
0
3
R ad iatin g distan ce (m )
Figure A4.7
-109-

External Fire Spread

Uploaded by

Copyright:

Available Formats

External Fire Spread

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

External Fire Spread

Uploaded by

Copyright:

Available Formats

What is the document about?

What is the document about?

What parameters affect external fire spread between buildings?

What parameters affect external fire spread between buildings?

EXTERNAL FIRE SPREAD TO

EXTERNAL FIRE SPREAD TO ADJOINING BUILDINGS

Building Code, enclosing rectangle, external fire spread, flame projection,

Copyright: Brandteknik, Lunds tekniska hgskola, Lunds universitet, Lund 1999.

Department of Fire Safety Engineering

Telefon: 046 - 222 73 60

Telephone: +46 46 222 73 60

Robert Jnsson, my supervisor, for always finding time to help me when

Per-ke Olsson, my second supervisor and a Fire Engineer at Arup Fire

Peter Johnson, Arup Fire in Melbourne, for giving me the opportunity to

Chris Marrion, Arup Fire in New York, for providing me with

Greg North, my dear friend from New Zealand, for proof-reading my

17th December, 1999

Since the radiation intensity is dependent on the temperature to the power

The prescriptive Building Regulations studied in this report use values of

The Building Regulations in Australia, England and Wales, New Zealand

The calculation methods studied in this report predicted approximately

The calculations performed with a method by Fredlund et al. (1976) as

PARAMETERS AFFECTING FIRE SPREAD BETWEEN BUILDINGS............................ 3

BUILDING CODES ................................................................................................................... 17

COMPARISON OF BOUNDARY REQUIREMENTS .......................................................... 51

ENCLOSING RECTANGLES ......................................................................................................... 33

IMPORTANCE OF RADIATION FROM PROJECTED FLAMES.................................... 57

DISCUSSION AND CONCLUSIONS...................................................................................... 77

FURTHER RESEARCH ........................................................................................................... 81

APPENDIX 1 COMPARISON OF BOUNDARY REQUIREMENTS FOR WAREHOUSE.... 87

At the present, there is no internationally accepted method to design against external

A combination of an extensive literature study and assessments of different

Parameters affecting horizontal fire spread

2 PARAMETERS AFFECTING FIRE SPREAD BETWEEN BUILDINGS

Fire spread between buildings

Parameters affecting horizontal fire spread

emitted radiation [W/m2]

Parameters affecting horizontal fire spread

Drysdale (1985) reports on a method to predict fire temperatures in compartments

Parameters affecting horizontal fire spread

temperature of ambient air [C]

Parameters affecting horizontal fire spread

Parameters affecting horizontal fire spread

window openings were assumed to be the only radiators radiating with an

Net calorific value (MJ/kg)

Heat release rate (MW/m2)

Parameters affecting horizontal fire spread

flame length above top of window [m]

Parameters affecting horizontal fire spread

width of window [m]

area of opening [m2]

flammable linings in the fire compartment,

Parameters affecting horizontal fire spread

x = 0.312 H 1.54W 0.54

No wall above window:

temperature at distance l along flame axis [K]

Parameters affecting horizontal fire spread

temperature at the window [K]

Absorption coefficient for flames from different materials.

An emissivity of 1 is often used for window radiation from compartments, (Barnett