Fire and the spatial separation of buildings

Publisher’s version / Version de l'éditeur:

Vous avez des questions? Nous pouvons vous aider. Pour communiquer directement avec un auteur, consultez la première page de la revue dans laquelle son article a été publié afin de trouver ses coordonnées. Si vous n’arrivez pas à les repérer, communiquez avec nous à PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca.

Questions? Contact the NRC Publications Archive team at

PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca. If you wish to email the authors directly, please see the first page of the publication for their contact information.

https://publications-cnrc.canada.ca/fra/droits

L’accès à ce site Web et l’utilisation de son contenu sont assujettis aux conditions présentées dans le site

LISEZ CES CONDITIONS ATTENTIVEMENT AVANT D’UTILISER CE SITE WEB.

Technical Paper (National Research Council of Canada. Division of Building

Research); no. DBR-TP-212, 1966-02-01

READ THESE TERMS AND CONDITIONS CAREFULLY BEFORE USING THIS WEBSITE. https://nrc-publications.canada.ca/eng/copyright

NRC Publications Archive Record / Notice des Archives des publications du CNRC :

https://nrc-publications.canada.ca/eng/view/object/?id=24770389-bba7-4a3a-b1cc-4231af0b3614 https://publications-cnrc.canada.ca/fra/voir/objet/?id=24770389-bba7-4a3a-b1cc-4231af0b3614

NRC Publications Archive

Archives des publications du CNRC

This publication could be one of several versions: author’s original, accepted manuscript or the publisher’s version. / La version de cette publication peut être l’une des suivantes : la version prépublication de l’auteur, la version acceptée du manuscrit ou la version de l’éditeur.

For the publisher’s version, please access the DOI link below./ Pour consulter la version de l’éditeur, utilisez le lien DOI ci-dessous.

https://doi.org/10.4224/40001478

Access and use of this website and the material on it are subject to the Terms and Conditions set forth at

Fire and the spatial separation of buildings

1 1 3

- 7 7

CANADA

CONSEIL NATIONAL DE RECHERCHES

e9s?s

Fire and the Spatial

Separation of Buildings

by J. H. McGuire

A N A L Y Z E D

Reprinted from

Fire Technology, Vol. 1, No. 4 November 1965, pp. 278-287

Technical Paper No. 212 of the

Division of Building Research

OTTAWA February 1966

NRC 8901

/ 8 7cc?1 t7*il

LE FEU

SOMMAIRE

On sous-entend que les valeurs des distances de s6para-tion des bdtiments calcul6es sur la base d'un rayonnement maximal pr6viendront la propagation d'un incendie par rayonnement pendant une dur6b ind6finie. Les valeurs calcul6es des distances d6passent toutefois ce qui est r6alisable pratiquement. L'auteur pr6sente des tables de distances de s6paration suffisantes pour pr6venir la propagation des in-cendies par rayonnement pendant un laps de temps per-mettant aux sapeurs-pompiers de commencer leurs op6ra-tions. L'auteur explique comment ces tables ont 6te calcul6es

et 6tur'' emplr {e leur -:d\ :C\ : + --F1-iC 6-o =-o o-9 Q-A -a :@ -T -CD

REPRIIVTED FROM

FIRE TECHNOLOGY

V o l . I N o . 4 N O V . 1 9 6 5

FT.12

Norn: This paper is a contribution of the Division of Building Research, National Research Counciil,-Canada, and is published with the approval oT the Director of the Division. Acknowledgement is due-to Mr. G. Williams-Leir for programming the com-pr.rter to solve the configuration factor equations and to Mr. P. Huot for carr5ring out the computations.

Fire and the Spatial

Separation of Buildings

Copyrisht 1965 NATIONAt tlRE FROTTCTION ASSOCTATION 60 EATTERYMARCH ST., BOSTON, MASS, O21IO J. H. McGUIRE, SFPE

Diuision of Building Research National Research Council (Canada)

It has been implied that spatial separations based on peak radia-tion levels will prevent igniradia-tion by radiaradia-tion, indefinitely. The specified distances, however, exceed practical limits. Separations calculated to prevent ignition by radiation long enough for fire ex-tinguishing operations to be initiated have been tabulated. The author explains how the tables were derived and discusses problems that may be encountered in their use.

,TtHE _{spread of fire from one building to another separated from the first} I by a vacant space may result from one or more of the following mecha-nisrrs:

o Flying brands.

o Convective heat transfer. o Radiative heat transfer.

Flying brands may initiate secondary fires at substantial distances from the primary fire, e.g., at least a quarter of a mile. It is not, therefore, practical to consider the spatial separation of buildings as a means of com-bating this hazard. Regulation of the choice of exterior cladding materials, particularly on roofs, minimizes such ignitions, and their extinguishment is usually easy, provided they are detected at an early stage.

Convective heat transfer will cause ignition only if ttre temperature of the gas stream is several hundred degrees Celsius. Such high gas temper-atures are only to be found in or very near the flames emanating from the windows of burning buildings.

Ignition by radiation from a burning building can occur at distances substantially greater than those to which flames generally extend. It is this mechanism, therefore, that will be the factor governing the specifica-tion of the spatial separaspecifica-tion of buildings from the fire point of view.

Separation of Buildings 279 The remainder of this paper is devoted to the formulation of a tech-nique for prescribing separation distances between buildings with a view to reducing the likelihood of spread of fire by radiative heat transfer.

T O L E R A B L E R A D I A T I O N L E V E L

When discussing the possibility that combustible materials will be ig-nited by radiant heat, the lowest level of intensity that proves to be sig-nificant is 0.3 cal/sq crnfsec; below this, most materials cannot be ignited in the presence of a pilot flame. Unfinished, untreated fiberboard does not obey this generalization and will ignite in the presence of a pilot flame at even lower intensities. In the present context, however, this feature will not be considered on the grounds that untreated, unfinished fiberboard is very unlikely to be a material exposed to radiant heat from fire in an adjacent building.

The mechanism discussed above involves the presence of a pilot flame, which constitutes a local high-temperature source. When a building is on fire and is exposing another to radiation, sparks and flying brands con-stitute the local high-temperature sources. In many cases, where a radia-tion level greater than 0.3 cal/sq cm/sec is incident on a building, a spark or flying brand will pass through the evolved streams of combustible volatiles.

It will therefore be assumed, that the spatial separation of buildings should be such that a fire in one building should not subject the facade of another to levels of radiation higher than 0.3 cal/sq cm/sec.

R A D I A T I O N L E V E L S F R O M B U R N I N G B U I L D I N G S ExppnrupNrar- FrNnrNcs

The radiation levels to be expected from burning buildings were investi-gated in the course of a program of full-scale burns known as the St. Lawrence Burns, carried out by the Division of Building Research, Na-tional Research Council, during the winter of 1958.1 The following were the principal findings:

. The nature of exterior cladding - brick or clapboard - did not noticeably influence radiation levels.

. Peak radiation levels at some distance from the buildings coincided with those that would result if window openings, at an appropriate hypo-thetical temperature, were taken to be the only sources of radiation.

. Peak radiation levels from buildings with highly flammable linings were twice those from buildings with noncombustible linings.

o Radiation levels were affected by wind direction, those on the lee-ward side of a building being, in general, much greater than those on the windward side.

Using the second result described above, it was found that peak hy-pothetical radiation levels at window openings on the leeward sides of

280 Fire Technology the buildings came to nearly 40 and 20 cal/sq cmfsec, respectively, for buildings with highly flammable and noncombustible linings. These values are, in fact, much greater than the maximum level to be expected at win-dow openings - about 4 cal/sq cm/sec - because it is assumed that radiation from the flames above windows is actually emanating from the window openings themselves. Spatial separation calculations using these results gave a range of values that were inordinately large and virtually impractical. An attempt, therefore, was made to justify basing spatial separation calculations on lower levels of radiation from burning buildings. In re-examining the results of the St. Lawrence Burns, it was noticed that, although the fires had been arranged to develop very rapidly, radia-tion levels did not exceed about one-flfth of the peak values listed, i.e., 40 and 20 cal/sq cmf sec until at least 16 min had elapsed. As fire fighting is in progress at this stage for the great majority of fires, it is possiblethat spatial separation would perform adequately if it merely gave protection against the spread of fire during this period. In many cases, spatial separa-tion calculated on this basis would, in fact, protect a building indefinitely, for the radiation levels previously discussed are maxima and would not always prevail.

Frnr.o Sr:unrns

To throw more light on the possible hazard. of adopting the less stringent approach just referred to, it is worth examining the results of the very Iimited number of field investigations carried out to date.

It is preferable to discuss this question in terms of a quantity called the configuration factor rather than in terms of radiation levels, which obvi-ously are not recorded during a fire. A configuration factor is defined as the ratio of the radiant intensity at the receiving swface to that at the (one or more) radiating surfaces. Assuming that these are at, a uniform black body temperature, a configuration factor is calculated solely from the relative geometry of the radiating and receiving surfaces. If it is assumed that radiation may be represented as emanating solely from win-dows and other openings, then this latter calculation is usually feasible following a fire.

The configuration factors that would be specified on this basis, to offer protection against the peak levels of radiation measured at the St. Lawrence Burns would be 0.3/40:0.0075 (hazardous cases) and 0.3/20:0.015

(normal cases). To guard against radiation levels of about one-fifth the peak value would call for configuration factors of 0.035 (hazardous cases) and 0.07 (normal cases).

The first record of configuration factor calculations made on this basis during field investigations may be found in a British technical paper pub-lished in 1950.' The results relate to two fires. For the first, the results refer to the condition of a number of window frames in the exposed buitd-ing and are given in Table 1. The exposbuitd-ing buildbuitd-ing was a multistory clothing store.

Separation of Buildings

Tanr-n L. Damage Related to Configuration Factor Configwation foctor Condition of window frame

0.093 o.L12

Paint blistered

Paint blistered, little charring Surface charring

Burned Burned

The second fire gave only one result - a timber billboard with a con-figuration factor of 0.092 ignit€d.

A fire that occurred in Winnipeg in 1956 also offers interesting informa-tion on this subject. At one stage, an exposd building had a configurainforma-tion factor of 0.05 but did not becorne involved in the fire. Shortly afterwards, ,another building ignited, raising the configuration factor of the exposd building to 0.1. Many of the window frames of the exposed building then ignit€d.

The choice of a corifiguration factor of 0.07, based on the St. Lawrence Burns rreeults, appears to be eompatible with the above field obeervations and appropriate for normal use. The high intensities recorded during some of the St. Lawrence Burns are so disturbing, however, that it is suggested that a configuration factor of 0.035 should form the basis of separation calculations involving buildings that can be expected to burn extra vigor-ously.

Since the above suggestion was adopted in the 1960 edition of the Na-tional Building Code of Canada, a field fire investigation involving two dwellings has further justified it. The two dwellings were separated by a dietanc€ of 17 ft, which is 2 ft gxeater than the 15 ft given by a configuration factor of 0.035 together with a constant addition of 7 ft, i.e., as for Table 2. Despite this substantial separation, ignition still occurred, suggesting that the distances prescribed are not excessive. The fire was started with gasoline, which pertly explairrs the very rapid development and the attain-ment of poak radiation levels before the arrival of the fire departattain-ment.

T H E T A B L E S DnnlettoN

Tables 2 and 3 are samples of calculations based on configuration factors of 0.035 and 0.07, respectively, for particularly hazardous and normal con-ditions. In other words, the specified separations theoretically reduce the radiant intensity at an exposed building to 0.035 or 0.07 times the equiva-lent intensity at the window openings of the exposing building.

Further distances of 7 ft (particdlarly hazardous) and 5 ft (normal) have been added, following the basic calculations, to account for the fact that flamee have a horizontal projection and that the equivalent radiating

0.067 0.067 0.081

?42 Fire Technology Tnnr,n 2. Building &parations (hazardous conditions)

Width of Per cent

compartment of window (ft) opening Height of comportment (ft) 12.5 25 37.5 50 75 100 20 63 D D 45 32 40 35.5 30 22.5 53.5 47 39.5 29 64 56 46 33 86 74 60 4l 100 t o 50 25 100 t o 50 25 100 75 50 25 100 ' t c 50 25 100 l o 50 25 30 60 100 200 100 I D 50 25 53.5 47 39 28 73 64 53 38 88 76.5 63 45 t20.5 104.5 85 59.5 L52 131 105.5 7I 204.5 173.5 136 85 87.5 76.5 62.5 44.5 106 92.5 ' / D . b O.t.A 147 L27.5 LO4 72.5 186.5 161 130 89 255 218 173.5 1t2.5 7 l 61.5 50 34.5 100 86.5 7L 49.5 t2t.5 105.5 86 60.5 169 146.5 1 1 9 . 5 83 2 L 5 . 5 186 151 103.5 297 254.5 203.5 135 84 93.5 72 79.5 57.5 62.5 37.5 39.5 119.5 135.5 103 116.5 83.5 93.5 57 62 146 166.5 126.5 L43.5 L02.5 116 71 78.5 205.5 235.5 L77.5 203.5 L44.5 165 100.5 114 263 303 227.5 26t.5 184.5 212.5 L27.5 146.5 366 423.5 315 365 253.5 294.5 17L 200.5 106 90.5 7 L . 5 16 136.5 r 1 3 . 5 85.5 50

surface is thus in front of an actual building facade. The two dimensions, 7 ft and 5 ft, were results given by the St. Lawrence Burns during peak levels of radiation. As the separations ane not intended to offer protection in these circumstances, it might well be that these dimensions are somewhat excessive and should be reduced by some 2 or 3 ft.

To cater for the almost infinite variety of window shapes and distribu-tions that exist in building facades, a variable "percentage window open-ing" has been introduced. Where windows are uniformly distributed and are close together in comparison with the spatial separation distance, this action will not introduce noticeable

error-Calculation of Tables 2 and,3 was made on a computer suitably pro-grammed by a colleague in the Fire Section of NRC.*

-._*Amplified ve{Fions of thee tableq, together with others involving, for example,

different story heighte, and different increments in percentage window opening, are available on r€queot from the Division of Building Research, National Research Council.

Separation of Buildings 283 Pnnctrcer- Appr,tcattot r

The first feature to be considered in applying tables of this nature is whether or not the adoption of an average value of percentage window open-ing will give a valid distance of separation. If, say, windows occupy a greater proportion of exterior wall area at one end of a building facade than at the other, then a greater separation is called for in that region. To be on the safe side, it would be desirable to require a separation in that par-ticular region based on the adoption of a higher value of percentage window opening while retaining the true values of the height and width of the building.

If an individual window or other opehing proved to be very large, i.e., to have dimensions comparable with the separation dimensions, further modification would be necessary. It would be essential to provide for greater separation that that given on the assumption that a particular window was the only radiator. Without delving more deeply into the evaluation of configuration factors,3 it is not practical to offer

recommenda-Tasr,p 3. Building Separations (normal conditions)

Width of Per cent compartment of window (ft) opening Height of compartment (ft) 12.5 25 37.5 50 75 roa 43 D J 30 20 60.5 52.5 42.5 29 100 ' l D 50 25 100 75 50 25 100 I O 50 25 100 75 50 25 100 75 50 25 100 75 50 25 1 0 20 30 60 100 28 24.5 20.5 1 5 o , 32.5 27 19 44 38 3 1 2l 58 49.5 39 24.5 69.5 58 44 26 83.5 67 48 26.5 J ' 32 26 1 8 5 1 44 36 25 61 53 43 29.5 83.5 71.5 57 37.5 103.5 88 69 4 2 . 5 134 1 1 1 83 46.5 73.5 63.5 5 1 . 5 35 LO2 87.5 70.5 47 t28 109.5 8 7 ' 55.5 r 7 t . 5 t44 1 1 0 . 5 64.5 48 4 T 32.5 20.5 69 59 47.5 3 1 . 5 84 7 2 . 5 D t ' . D 39 tt7.5 101 8 1 54 t49 t27.5 1 0 1 . 5 66 201.5 1 7 0 . 5 r 3 2 . 5 80.5 55.5 60.5 46.5 50 36 37.5 2L.5 22 8 1 . 5 9 1 . 5 69.5 77.5 55 60 34.5 36.5 100.5 1L4 86.5 97 69 76.5 44.5 48.5 t42.5 163 122.5 140 98.5 118 65 73 782.5 2tO.5 156.5 180.5 125.5 L44.5 82.5 95 251.5 292.5 214.5 250 169 198.5 106.5 127 200

284 Fire Technology tions that are both simple and valid. Having derived separation distances based on a mean percentage window opening and on the window opening itse[ it must be left to the designer to assess an appropriate separation.

Radiation levels at a fixed distance from a building facade will decrease as the distance from the center of symmetry increases; therefore, it would be reasonable to relax spatial separation requirements near the corners of buildings. A number of sample calculations indicate that appropriate separations near the corners of buildings range between 65 and 95 per cent of those listed in Tables 2 and 3. It might be reasonable to suggest a re-laxation to 80 per cent of the value tabulated. The resulting separation requirements are illustrated in Figure 1.

3 5 0 r g s 9 8 0

ir

Figure 1 also gives the conditions required beyond the extreme corners of the building. In the case illustrated on the left of Figure 1, it might be considered some hardship that the boundary of the restricted area extends beyond the projection of the imperforate fire resistant side wall. This re-striction can be eliminated by ensuring that there are no window open-ings in the section CE of the adjoining wall.

The above measure has made use of two virtually self-apparent defini-tions. First, the equivalent building facade whose width and height will be looked up in either Table 2 or Table 3 will probably not coincide with the actual building facade. It is only necessary to include those openings that will be radiating freely during a fire. Thus, each story of the average build-ing will be separated from iLs neighbor by appropriately fire resistant con-struction and can be treated separately in the present context. Second, openings may be described as portions of the facade that might collapse and fall out during the course of a fire. Thus, any portion that does not meet

eoz.3rl ..

F I R E R E S I S T A N T W A L L N O O P E N I N G S

6 0 % D l

Separation of Buildings 285 the integrity requirements associated with fire resistance considerations will fall in this category. There is no call for it to meet any temperature requirements.

A complicating feature to be catered for is that the exterior wall of a building is often irregular in shape (Figure 2). In such cases, preliminary considerations should refer to a line joining the extremities of the exterior wall. Where the building is entirely contained behind this line no further steps are required, for so far as radiation levels are concerned the irregular external wall is closely represented by an imaginary one having the same percentage window openings and located on the line referred to. Where a portion of the building projects beyond the line, separation requirements will be largely fulfilled by a composite boundary line as illustrated in Fig-we 2. It is made up of a boundary line as calculated above, together with one referring solely to the projecting portions of the building.

B U I L D I N G

I B O U N D A R Y

L I M I T S

Figure 2. Boundary mnditions for irregularly shaped buildings.

Building codes usually discuss the location of a building with relation to the lot line rather than to another building. It is difficult to see how this type of specification can be soundly framed. The only practical suggestion that has so far been conceived is that buildings should be separated from their lot lines by half the distances derived according to the principles here discussed. Where this rule is adopted for two adjacent buildings that are mirror images of each other, the separation between the two will, in fact, be appropriate. For dissimilar buildings, however, this will not be the case, and the separation may be more than adequate if the one building catches fire and less than adequate if the other ignites. It is doubtful whether this

286 Fire Technology incompatibility will ever be resolved. The only mitigating feature is that, in most cases, the building exposed to unnecessary hazard, will be the smaller of the two - the greater the difference in size, the greater the hazard.

C O M P A R I S O N W I T H O T H E R W O R K

The Joint Fire Research Organization in the United Kingdom has de-veloped recommendations concerning the spatial separation of buildings along the same lines as those described in this paper.a The choice of the level of radiation to be considered tolerable at an exposed building is the sanne, both being based on JFRO results.

In specifying the radiation levels to be expected from burning buildings, it is stated that the radiant contribution from flames issuing from windows may virtually be neglected. It is assumed that windows and other open-ings radiating at a temperature not exceeding 1,100'C will be the only sources of radiation. In terms of configuration factors, the recommenda-tion is that calcularecommenda-tions should normally be based on a value 0.075.

This value corresponds closely with the one used in this paper, although it is not claimed that separations based on the latter (together with the additional 5 ft always included) will prevent the spread of fire unless fire fighting is undertaken before the fire attains its peak. The British report implies that the separations wilI be adequate in their own right.

The British report also suggests that where fire loads are low, 5 lb/sq ft or less, much less stringent separations based on a configuration factor of 0.15 are acceptable. It is probable that this relaxation is appropriate for certain types of buildings now being constructed. Relaxation of the separa-tions suggested in the present paper could, in fact, conveniently be achieved without computing additional tables. By multiplying the percentage window openings by a factor of 2, Lhe "normal" tables, instead of being based on a configuration factor of 0.07, would be based on one of 0.14 (with the constant addition of 5 ft). The values thus obtained would correspond closely to those given in the British table, except for the 5-ft addition re-ferred to.

The lowest value of percentage window opening available in the table would then be 2 X20% :4070, and it might be considered some hard-ship not to have lower values available. However, the use of lower values might be somewhat dangerous. Separation values would be small and might well become comparable to the dimensions between windows. Such conditions would invalidate the use of the variable "percentage window opening," which assumes a continuous distribution of very small windows. A relaxation, such as the above, might be recommended where wall linings and the contents of a building have very low flammability ratings and constitute a low fire load, say less than 5 lb/sq ft.

In the British report, the absence of especially stringent requirements with regard to buildings that might burn extra vigorously would seem

Separation of Buildings

undersirable as far as Canadian conditions are concerned. The results of the St. Lawrence Burns and the field investigation of the dwelling house fire, here reported, emphasize the need for a stringent requirement here in Canada.

The British report states that "A wall clad with timber would be con-sidered as an opening, since the burning timber would act as a source of radiation . . ." The St. Lawrence Burns results suggest that the radiation level from clapboard cladding can be neglected, provided the wall remains intact and is fairly thick. A plane vertical sheet of thick timber will burn vigorously only if it receives supporting radiation or convection on' its front side, or alternatively, supporting conducted heat from the reverse side.

R E F E R E N C E S

r "The St. Lawrence Burns," G. W. Shorter, J. H. McGuire, N. B. Hutcheon, and R. F. Legget, NFPA Quarterly, Vol. 53, No.4 (April 1960), pp. 300-316.

, "Radiation from Building Fires," R. C. Bevan and C. T. Webster, Investigations on Building Fires - Part III, National Building Studies Technical Paper No. 5, 1950.

3 "Heat Transfer by Radiation," J. H. McGuire, Fire Research Special Report No. 2, 1953.

a "Heat Radiation from Fires and Building Separation," M. Law, Joint Fire Re-search Organization Technical Paper No. 5, 1963.