BK /S(

Building Research Establishment Report

Design pr¡nciples for smoke ventilation

in enclosed shopping centres

HP MorgâÍr, BSc, CPhys, Mtnstp, AtFireE, MSFSE and JP Gardner*, BSc

-Colt lnternational Ltd

@@lãr
ld.ûtbot L¡ñi{d
Co¡r
N* L¡ft Hô¡d
Hmp3 ¡'G POg2LY
lckphæ H¡d (tð5) ¡jill¡
Tclc¡8óllt ¿ Eó:19
Fr¡ Hüd ((IÐ5t Jt¡!?0

Presented to

on behalf of
Colt International Limited
co-authors of this document

Acorrc.ñp€oñBôy Jerry Gardner BSc

eOgf
zñ

Fire Research Station

Buílding Research Establishment
Borehamwood, Herts
WD6 2BL

&iaq(/ ¿Á
Price lisis for ålt availêbte
BRE publications can be
obtaioed from:
ERE Bookshop
Building Research Esablishment
Carston. Watford. WD2 ?JR
Telephone:0923 6$aa4a

-''ì$tî'ìffi::::$,.ììåi:"

ER 186
{EgN 0 85125 462 4 --.- ã
@ Crown copyright 1990 ..
Fir$r Publi$hed 1990

Applications to reproduce extracts

frorn the lerr of this Publicarion
should be made to the Publications
Officer at the Fire Rese¡rch S¡ation

,Ar'gât,
Contents

Pas,e

Foreword

Introduction I

Chapter I General principles 3

Chapter 2 Design procedure for the mall smoke control system 6

Smoke reservoirs 6

Design fire size 6

Single-storey malls 1

Two-storey malls with srnall voids I

Two-storey malls with large voids 10

1A
Multi-storey malls l¿t

Calculating smoke temperaturc t(

ñlall sprinklers l7

-, Flowing layer depth 18

Local deepening 18

Inlet air 19

Minimum number of extract points 20

Natural ventilation - area required per reservoir 2l

Powered ventilation 22

Chapter 3 Large shop opening onto a mall 23

Chapter 4 Some practical design considerations 25

Factors influencing the design fire 25

The effect of wind on the efficiency of a smoke ventilation system 25

False ceilings in the mall 25

The use of a plenum chamber above a false ceiling in a shop 26

Stores with internal voids 4 ZT

Sloping malls 27

Assessment of effective layer depth 27

(continued)

tll
 Srrltikc l'ltlrv irt ltrrv lrarrtlrv tlalls 1i

Basemcnt scrvice levels 21

Encloscd car Parks 29

Smoke transfer ducts 29

Entrances within the smoke laYer 30

Other situations 30

Chapter 5 Some operational factors 31

References

.'ït;Lïïï|iffi;ïl::I'

lv
Foreword
Building Research
This Report has been prepared by the Fire Research Station of the
scientific knowledge and practical experience in
Establisirment (BRE) anå results from the latest
of 'Smoke control methods in enclosed shopping
smoke movement and control. It is an update
summary' published in 1979' and is based upon
complexes of one or more storeys: a design
pretiminary work carried out for BRE by Colt International Limited.

consisting of industrial and

The Report takes into account the comments of a Liaison committee
grateful to the Society of Fire Safety
government representatives. The Fire Research station is also
of the Society of Fire Protection
Ëngineers, the Institute of Fire Engineers and the UK Chapter
comments which have as far as possible been included'
eniin""., for providing detailed
The Fire Research station acknowledges the assistance
given by colt International Limited in the
preparation of the final RePort-

-{F
Introduction
This Report is intended to assist designers ol smoke comnronly try to escape by the same route they had
ventilation systems in enclosed shopping complexes. used to enter the prenriscs (sce for example, Canterl).
lvfost of the methods advocated are the outcome of It follows tllat cscapc via the ¡rralls must be assumed,
research into smoke control by smoke ventilation at even where other exits are providedl. In this situation a
the Fire Research Station, but also take into account long evacuation period can be expected. A detection
the recommendationsl of the Working Party on fire and alarm system is required to give early warning of a
precautions in town centre redevelopment, as well as fire, and sprinklers are needed to control its spread.
experience gained and ideas developed whilst the Statistics of fire deaths show that the majority of
authors and their colleagues have discussed many fatalities are due to the effects of smoke.
proposed schemes with interested parties. The
primary purpose of this Report is to summarise the The ideal option would be to prevent any smoke from
design advice available from the Fire Research Station a shop fire entering the mall at all. In the majority of
at the time of its preparation, in a readily usable form. cases, it would either be very difficult or extremely
As such, the Report is neither a detailed engineering expensive to fit a separate smoke extraction system to
manual nor is it a scientific review article. Perhaps each and every shop, however smalll. Note that large
most important of all, it is not a summary of the shops are, however, an exception to this rule and the
totality of approaches possible. New methods such as proyision of smoke ventilation systems for such shops
those based upon computational fluid dynamics, will is discussed in Chapter 3. Occasionally circumstances
be developed as time passes and there will always be dictate that smoke ventilation fitted to each shop unit,
special cases where existing alternative methods can be even small units, is the most viable option for
adopted. protecting part or all of a mall. (This can apply, for
example, when an old conrplex is being redeveloped.
At peak times a shopping centre can be occupied by but the mall is too low and/or too narrow to allow the
thousands of people, and some larger centres by more installation of a viable smoke ventilation system in the
than a hundred thousand. A typical centre may mall itself). There have been several examples of this.
comprise many individual shop units opening onto a Nevertheless it remains generally true that this option
common mall. Although the individualunits may be is rarely found to be appropriate for most malls.
separated from each other by a dividing wall of f ire
resisting construction, usually the shop is either open Occasionally, one still hears the suggestion that the
fronted or only separated from the mall by a glass shop mall should be pressurised to prevent smoke moving
front. This means that the public areas of the entire from a shop into the mall. This is not usually a viable
centre can be effectively undivided. option by itself where the opening between shop and
mall is large (eg an open-fronted shop, or a shop whose
Means of escape from within each shop unit will, in glazing has fallen away in whole or part). This is
general, be specified (eg BS 5588 Part 22) in the same because the airspeed needed from mall into shop in
way as for shops which are not part of an enclosed order to prevent the movement of smoky gases the
complex. This means that escape from within the shops other way through the same opening, can vary between
is specified as if the mall were as much a place of safety l/z and c. 2 ms'l depending on fire size, gas temperature
as the usual open-sky street. Unfortunately, the mall is etc. All of this âir must be continuously removed from
a street with a roof, and so cannot be regarded as being within the shop unit in order to maintain the flow. The
as inherently safe as an open-topped street. People quantities of air-handling plant required will exceed
escaping from a shop into a mall will still need to travel the size of smoke ventilation systems for most typicat
along the mall before exiting to a true place of safety. It shop-front openings.
follows that the mall constitutes an additianal stage to
the escape route, which needs to be protected from the Where smoke from a fire in a single shop unit could
effects of fire and smoke. Ideally, it should approach spread rapidly via the malls through the entire centre,
the same level of safety as a street for as long as people a smoke ventilation system in the malls is essential to
need to escape through it
- even if smoke enters the
mall from the shop on fire. Details of means of escape
ensure that escape is unhindered, by ensuring that any
large quantities of thermally buoyant smoky gases can
provisions for the malls can be found elsewhere¡. be kept separate from (ie above) people who may still
be using escape routes through the malls. Therefore,
A shoppingtomplex is a public building and the the role of a smoke ventilation s$¡lin¿¡F principally
occupants will be a cross section of the comrirunity one of life safety. It should also be remembered,
including the elderly, children and the disabled. They however, that firefighting becomes both difficult and
will not necessarily be familiar with the building, or dangerous in a smoke-logged mall. It follows that to
perhaps more importantlf with all the escape routes assist the fire services, the smoke ventilation system
that might be provided for them. In many types of should be capable of performing its design function for
building it is widely recognised that people will a period of time longer than that required for the
puhl¡c t() csc:.ìpc: so an imnlcdiutc iltrack <tn thc lire rcgu la t tlr.t' u rr I hori t ies, <Jcvelope rs and l'i rc scicn t is ts.
can be rùade aiter the arrival ol'thc lire brigade. ralhcr than by spccilic research. Such advicc has lrcen
inclucled i¡r this Report with the intention of giving the
Guidance on design prin.ipte. for smoke ventilarion fullest picturc possible. f t is therefore Iikely thar some
systems was summarised in a report by Morganl of this guidance rvill need to be modificd in rhe fururc,
published in 1979. This was based on the knos,ledge as the results o[ continued research beconrc availablc.
available at that time. Since then a great deal of
relevant research has been carried out, which for the A Code of practices for enclosed shopping centres is
most part has confirmed the guidance given in the currently being prepared by the Brirish Standards
original report, but has in some areas highlighred rhe Institution (BSl). The aim of this Report is to provide
need to modify rhat guidance. A great deal of practical
guidance only on design principles for smoke
experience has been gained in designing such systems.
ventilation and it is hoped to supporr the Code rarher
Also in the intervening years, shopping centres than pre-empt it. The Report cannot cover all the
themselves have become larger and more intricate with
infinite variations of shopping centre design. Iñstead it
gives general principles for the design of efficienr
many open levels, interconnecting voids, sloping floors
systems, with simplified design procedures for an ideal
and atrium features. These can result in a very complex
path of smoke flow from the shop in which the fire model of a shopping mall and then further guidance on
starts to the eventual point of extraction. frequently encountered practical problems. Because
the design procedures are of necessity simplified. the
Report also gives their limitations so that. when
The purpose of this Report is to update the guidance necessary a more detailed design by specialists can be
available in the earlier design summarya to reflect
carried out.
these changes, to assist designers of smoke control
.\,stems in enclosed shopping centres. As was the case A Code of practice for atrium buildings is also being
rvith the previous work, this Report only gives prepared by the BSI. An atrium can be defined as any
guidance in [ine with current knowledge and generally space penetrating more than one storev of a building
accepted practice. The guidance is based on results of where the space is fulty or partially covered. Mosr arria
research where possible, but also on the cumulative within shopping centres may be considered as part of
experience of design features. required for regulatory the shopping mall and treared accordingll,. Wherc atria
purposes. of many individual smoke ventilation have mixed occupancies (eg shops and offices) then
proposals. Many of these design features have been reference should be made to the above document.
evolved over a number of years bv consensus between when available, or specialist advice sought.

-:ì
Chapter 1
General principles

Smoke fronr a firc in a shop rises in a plurne to the quantities. but will not mix apprcciably into ir
ceiling. As the plumc ;iscs, air is êntrained into it, horizontally fìowing stream of hot smokv gases exccpr.
increasing the volume of smoke and reducing its under specia[ (and usually local) conditions.
temperature. The smoke spreads out underneath the
ceiling and forms a laver which deepens as the shop Smoke florving from the shop onto the mall rvill risc ro
begins to fill with smoke. As the layer deepens, there is the mall ceiling (Figures 1 and 2). Air will mix into rhe
less height for the plume of smoke to rise before it smoke as it rises. If no smoke control measures ar€
reaches the smoke layer. Less air is being entrained, present, the gases will then flow along the mall as a
with the result that the temperature is higher on ceiling layer (Figure 2a) at a speed typicaily between
reaching the smoke layer. As this continues. the 1 and 2 m/s8.
increasing smoke layer temperature at some point will
operate the sprinklers. The fire may shatter the shop This is faster than the probable escape speed of
front glazing (if present) unless that glazing is fire- pedestrians in a crowded malls. When the smoke
resistant. The smoke will then flow out of the shop into reaches the end of the closed mall it will dip dorvn to a
the mall (Figure 1). The change of direction as the lorv level and be drawn back towards the firel(r(Fieure
smoke florvs out of the shop front results in increased 2b).
mixing with the surrounding air.
If the end of the mall is open, even light winds blowins
There is so much mixing that, except close to the fire into or across the opening will cause severe'local
itself. the hot smokv gases can be regarded as disturbance and mixing, and once again smokc at lou,
consisting entirely of rvarmed air. rvhen calculating level rvill be drawn.back towards the firer0 (Figure 2c).
flow. The quantitv of smoke particles (and hence the Hence a single-storey mall could become smoke
optical density and visibility through the smoky gases) logged within minutes. An unsprinklered fire in a
produced in a fire, depends strongly on the nature of single-storey shopping centre in Wolverhamptonrr, is
the burning materialó. thought to have caused a lO0-metre-long mall to
become untenable within one minute. Similarly short
The quantity of hot gases carrying these particles is, times can be expected to apply to the upper floors of a
howeveç mainly dependent on the size and rate of multi-storey mall. Since it is not usually practical to
burning (related to the heat output per second) of the prevent smoke entering the mall, except for larger
fire. Increasing the visibility through these gases shops, a mall smoke control system is necessary to
requires their dilution with clean air, but improving control and remove heat and smoke.
visibility (and reducing toxicity) by dilution alone in a
mall to safe levels after smoke has entered, is not a The Fire Research Station has extended the ideas
practical proposition. One suggested safe level of developed during its earlier work on the fire venting of
visibilityT (about 8 metres
- rather a short distance for
a mall) would require the hot smoky.gases to be
factoriesll to apply to mallss in order to use the
buoyancy inherent in fire gases to keep those gases
diluted by a mass ratio possibly larger than 300, even safely above the heads of people in the malls (Figure
after the gases leave the shop of origin. It follows th¿n 3). There are three essential features of a smoke
that hot smoky -qases should always be kept spatiatly ventilation system, without any one of which it will not
separated from escaping people. function effectively:

One can state as a general principle that air will mix 1 There must be some means of forming a smoke
into a rising stream ot'hot smoky gases in large reservoir to prevent the iateral spread of smoke,

tìgure I Smoke spread and main entrainment sites

in single and two-storey malls
--F

fwo-storey mall Single - storey mall

 Figurc 2a Crcation of a nroving snl<lkr: lurcr l¡crrcath rhc ccilinc
of an unventilated mall, shrlrving rìtovcment of the
dìsplaced arr

Figurè 2b Rccirculation of smokc in an unvcnrilared and closed mall

Figure 2c Mixing of snioke into rhe air beinq drarvn inro an

open-cnded mall. causcd by u.incl

tt
Smoke ex ha ust

Figure 3 Principles of system needed !o conrain smoke rn a

well-defined layer (section along mall)

which would result in excessive loss of buovancv. buoyancy driven ventilation or mechanical
This reservoir must be designed to contain the extraction, depending on the circumstances. The
smoke layer base wéll above head height. Cenerally rate of the exhaust must equal the rate at which the
speaking, malls with high ceilings or rooflighrs will smoke enters the reservoir from belorv.
allow a deeper smoke reservoir and hence a more
effective ñokë ventilarion system than in low. 3 Since the gases being extracted conSÍÇ añost
narrow malls. entirely of air that has mixed with the orisinal fire
gases, fresh air must enter the mall to takã its
There must be extraction of smoke within the olace.
It must enter at a rate equal to the rate of extràction
reservoir. to prevent the smoke layer building down of smoke and at a low enough height nor ro mi.r
below the design depth. This can be natural, prematurely with the smoke.
ln a iire nlost ol' thc inrportanI lactors are time Il re srrc.
<iepcndcnt: the [irnc lbr a l'ire to grow lronr ignition ttl
the design tire size. tinle lor the mall to smokelog an<1 There are other essential iactors that should be bornc
time needed for evacuation. Currently there are no in mind when designing a srnoke control system. When
reliable data on fire growth rates in retail premises. thc mall has morc than one lcvcl. ir is also necessaq/ to
Smoke filling times are known to bc rapid l¡ut are ¡ror rcstrict or channel snroke otì tlìc ktrvcr lcvcl in orclcr to
always easily quantified. The time needed for restrict. enrrainmenr into the plume rising through the
evacuation is unknown but it could be considerably upper level. Once the sntokc has enrered a ceiling
longer than the 2.5 minutes escape time used to size reservoir it will flow towards the cxtraction points.
exit widths to cope with the peak flow rates during this Since this horizontal florv is driven by thc g¿ìses'own
evacuation period. There is considerable anecdotal buoyanc¡ there will be a nrinimum depth for any
evidence within the UK fire services supporting rhis particular combination of smoke temperature and
view. mass flow rate. This is discussed in greater detail in
Chapter 2, but this minimum depth will sèt a limit to
These uncertainties lead to many problems in the design for the smoke reservoiç which cannot be
designing a time dependent smoke ventilation sysrem. shallower.
The design principles given in this Report are based
instead on steady srate conditions. A design fire is used The capacity of the extraction system depends mainly
which has a low probability of being exceeded, and the on the height the gases have risen to the lâyer base,
smoke ventilation system is designed to remove the and not at all on the area of the floor or the volume of
mass flow rate of smoke necessary to maintain clear the space. Therefore a simple pcrcentagc rule for
conditions in the mall for escape. Sprinklers in shops natural venting, such as 3o/o of the floor area, or an air
are an essential part of the smoke ventilation design in change rate, such as six air changes per hour lo¡
order to prevent the fire growing beyond the design powered ventilation, are nrisleading.

-<ã
Chapter 2
Design procedures fbr the mall snroke control svstem
'l'he [ollorvirrg sccl.iorrs rlutli¡rc a ge ncr¿¡l proccdulc 'l'hc
snloke rcse lvoir c¿rn oltcn be lìlrnle d lrv thc
whiclt can bc followed wtren dcsignirrg the mall snrokc dorvnstand fascia ol' the shops. conlbined rvith scrccns
control svstcm. First plarr thc positions ol the sr¡rokc at intervals along the mall (Figurc 4). Thc fascias ¿rre
rcservoirs. calculate thc mass l'low ratc of snrokc thcn nccessary, not to prevcnt thc smoke florving into
enl.ering thc reservoir. and citlrer the vcntilatit)rr ¿rrr:ir thc nlall l'rorn a lirc in thc shop. but to prcvcnt snloke
Ol'a natural ventilation svstcrn or the extract ral.c 0l a containcd in thc nlall rcscrvoir llorving into thc shops
powered system to remove the same amount o[ smokc. adjoining the mall. If there arc raised root"lights abovc
These procedures [rave been developed from 'idealised the malls, these can oiten be utiliscd as smoke
models' of shopping malls studied at the Fire Research rese rvoi rs.
Station. Commonly encountered variations are
discussed furtlrer in Chaoter 4. Although thc he ight through rvhich thc smoke nlav
havc to risc to a vcntilat()r installed ât the top of a high
Smoke reservoirs rooflight may be greater than for lower malls. provicled
that the smoke is allorvecl to build down to the samc
No smoke layer has a pertèctly defined interface rvitlr lcvel. the mass florv r¿rte of smoke entering the
the colder. clearer air bclorv: there is alrvays a snr¡rll rcscrvoir is no greater. but thc deeper rescrvclir rlt
amount ol cross mixing. This cross mixirrg has no snroke rvill produce a grcater huclvancv pressurc at the
significant effect on the quarrtity of smoke in a ceilins.
orizontal ceiling Ia-ver but can c¿ìuse prouressivc loss
of visibilitv in the air beneath. This loss of visibilin The scrccns u'hich l'ornl the smokc reservoir shouki bc
()ccurs ntorc rapidlv rvhcrl:
arr¡ngcd stl Ihat s¡nokc l'ronr a lirc i¡¡ arrv shop carr
()¡ll\' tlo$,irìto orìe resc-rvQir.
' a ceiling layer is cool (indeecl verv ct'rol gases rvill
not persist as a layer) 'Evcn rvith a snrokc rescrvoir linlited to l(X)0 m:.
c\cL'ssi\/c cooling and/or d<lrvnrvard mixing can occur
o the hot smoke and the air beneath havc a
irl stagnant regions of thc smoke laver. To prevent this
high relátive velocitv thc crtraction outlcLs, be they naturaI or mechanical.
should he distributed over the reservoir so as to
. turbulerrce disturbs the illterface prelent stagnant regions being formed. Potentiallv
staqnant regions of the smoke reservoir can sometimes
Heat is lost from the smoke layer by radiation he avoided b1' using smoke transfer ducts (see Chapter4).
downwards and by radiation and conduction to thc
surfaces of the reservoir. To prevent excessive heat loss
the size of an individual snloke reservoir within the Design fire size
mall is usually taken to be limited to l(X)0 ¡1: (o¡ l-i00 Beiore any smoke ventilation system can be designed
m: if mechanical extraction is used as its etiiciencr is it is cssential to determine a suitable size of fire. for
lcss susceptible to hcat loss). This reconlmcndation clcsisn purposes. This fire size then forms the basis of a
,,olved during the early l!170s and is pcrhaps hrest snrokc ventilation systenr clesign.
understcrod brv notine that in rlany 'industrial' sinr¡rlc
untlividcd conrpartmcnt applicirt ions oI srnokc
vcntilation. it has long bccn thc practicc lt¡ lirlrit
reservoir areas to bctrveen 2(XX) m: ancl -3(XX) nl:. Ex haust

In a mall. sntokc frorrr shops ()l'up t() l(X)0 nt: ¿rrc¿r (il'
thc mall is naturally ventilatccl) or up to l-ì(X) nl: (il'thc ¡- Smoke reservorr
mall has extract Ians) carr cntcr thc nrall rcscrvoir.
giving a total area affected by smoke similar to thc
lcln*9,-sta ndi ng practicc- st a ted above. Thc m ax i rn u rlr
distance between scrcens lorminq thc b<lundarv of the
rcservoir slrould bc 60 m. This distancc
recomme nçþtios fol lorvs earl ier co nside rat it'lns I
(confirnred in this ftrrm bv the Liaiso¡r Pancl consultccl
in preparation of the prescrìt documcnt). and clcrivcs
l'rtlnl concer¡r over thc dist¿rncc pcoplc shoultl lrc
c.rpected to walk below a snroke laycr while cscaping.
Fttr complex georrretrics ol smokc rcservoirs spccialist Figurc 4 Sntrkc rcscrvoir in nrall with a cüilinq hcieht
advicc shtlultJ bc sclurlht. -grcatcr thln thlt in shtrps (scctirrn across nrall)

6
Irlc;rllr tlre desiqrr l'ile slloulrl rhorr tlre ¡'rltvsicitl sizc Il nl:¡r'lrc nccess¿trv l() riuse tlle .rrrtrkc lrrvcr h¿rsc
tll lhc lirc incre asi¡lg ivith time.
arìcl lìci.ìt ()utput l¡lrovc this lirr practicll re ¿ls()rìs. l'his rr rll rcsult in an
lllowrng tlre grorvin! threaL ttt occup¿lnts to be irrcre lrscd ¡rl¿tss l'lorv t'¿ttc ol sntokc tluc to tlle
calculated as time incrcascs. Un[ortunately there is n<r rrclditio¡lal cntntitrmctlt. but ¿rs Illc srttokc llvcr basc is
¿rvailablc rcsearch. at thc tinlc oi prcparing this hiqhcr. ¿r grc¿rte r dcglcc ol's¿tlct-r'ri'ill lurvc lrccrr
Rcport, which allorvs asscssntcnt oi thc probability ¿rchicr'ccl. Shop I'ascias should e rtcrtd bclorv Ihis hcis,ht
<Jistritrution describing the variation of iire growth t'¡therrvise the !a;":r Lrase rvould bc lorv enough to cnter
curvcs for retail areas. Clearly, one does not want an unalfccte d shops. lndecd. l'ascias scrvc no othcr usc[ul
'íìve ¡¡9"' fire lor safety design. since typically hall of smokc controlling function. cxccpt as part of a
all iires would grow taster. lt is rnuch simpler to assess 5cparatc s¡noke extracti()n svstcrìr rtithin ir sho1r. or to
Lhe nraximum size a fire can reasonably be expecte d to cont¿ritì smoke from a verv s¡nall lire .
reach during the escape period, and to design the
system to cope with that. Note that even here. the Having established the clear layer heieht in the mall,
statistical evidence is not strong (see for example the nrass florv rate of smokc can thcn be calculated.
lvlorgan and Chandlerrr) fôr shopping malls. Further Rece¡rt work by Hinkleyls has conl'irnrcd the rate of
research is currently in hand to improve this statistical entrainment of air into a plume o[ snloke risinq abovc
basis. a [ire as:

It follo*,s from the foregoing that there is a strongly M = 0.l9PY'r/: (t)

subjective clcmen¡ in assessing rvhat fire size is
acceptably infrequent for safetl' design purposes. t,herc nlass llorv rate of smtlke entcring the
A l2 m perimeter (3 nr x 3 m) -i l"{W sprinkler
lvf =
smoke laver within thc shop (kds)
controlled fire has become the accepted basis in the P= (hc ¡rcrinreter of the fire (nl)
UK for a smoke ventilation system in a srrrinklered Y= the he ight t'rom the base ol the fire to
sholrpi ne centre l.a.si.
the snroke lavcr' (nr )

St¡nre factors alfecting the desisn fire size are discussecl

The rc is no inlormation available to sho*,horv
in Chapter 4. The desi_qn principles outlined in this
Equation I (or anl'current altcrnatives) shoutcl be
Report are based upon a 12 m perimeter 5 lvf W fire.
modilicd to allorv for the effects of sprinkler-spray
Should a different design fire be considered for
whatcver reasons. the equations. figures etc given in intcractions. Consequcntlv it is uscd here unmodified.
this Report may no longer applv and advice should be
sought from experts. Other fire sizes have occasionally Experimcnts have shorvnl¡ that the s¡noke florving
been specified by designers, for both sprinklered and from the shop onto the mall becomes turbulent with
unsprinklered shops. The problem of unsprinklered increasing mixing of air. The mass llow rate of smoke
shops has been discussed in more detail by Gardnerl{, enterin_e the reservoir is approximatelv double the
who has shorvn the importance of considering amount given by Equation 1, where Y is now the
'flashover' in such units and the consequent need to height from the fire to the base of the mall smoke layer
consider potentially very large fires. (Figure 5). Figure 6 shows the mass flow rate of smoky
q,ases entering the layer at different laver heights in a

Single-storey malls sinsle-storev mall. This can be calculated from:

The minirnum height of the sn'roke layer base must be
2..5 m from the mall floor to ensure safety. and lvl=0.38PYrt: (2)
prcferably at least 3 rn in a single-storey mall (Figure -5).
Thc Fire Rcscarch Station is currentlv studving
entrainment into smoke flow trom compartments. The
purpose of this u,ork is to determine more accurately
thc irrlluencc ol'such factors as compartment openine
geometry. the presence of a do\À'nstand fascia and
balcony/dorvnstand combinations. It fc¡llows that
Equatiorr 2 may be superseded in due course.

ll the clear layer in the mall is much higher lhan the

shopfront opcning. thcn the plume of snroke rising to
the smoke layer rvill be similar¡q¡A Ui{,o-storey
shopping centre (Figure 7). Where this height
differencc (hr) ['retr*'een the shopfront opening and the
smoke layer base is more than 2 nt (Figure 7) the mass
Fire perimeter p flow rate should be obtained using the procedure for
¡.wo-storey shopping centres given later in this chapter
Figure 5 Smokc ventilation in a singlc-storcv nrall on prge l().
 Trvo-strlrcv malls r.r,ith srnall voicls
Whc¡r a lirc occurs ()tì [hc lorvcr lcvcl ol ¿r trvo-store v
nrall. a plunre ol'srn<lkc h¿rs furthe r to risc lrclorc
entering thc slnoke rcservoir. This results i¡r grcatcr
e rttraintrìcnt ol'¿r¡r. hcttcc ¿t lilrgcr qu¿ltìtitv ol'snì()kc.

I'hc smoke l¿rvcr on thc uppe r lcvcl rvill Ilrcn lrc corllc¡.
and less wcll dcl'incd. In this casc ths smokc laycr
should be not lcss tha¡l 3 m above the upper floor lcvcl
and prefcrablv nrorc than 3.5 m. Again. shop fascitrs
should extend be lorv this hcighr othcrrvise the laycr
base would be lorv enough to entcr unaficcted storcs.

80 In a shopping cenrre which has small voids co.nnecrirìg

¿
levels, smoke from a fire in a lorver level shop rvill fiorv
ø
O out of the shop and spread in a complicared lrorizonr¿rl
.'n
circulation pattern bcneath the ceiling (ie beneath rlrc
upper deck) (Figure tl). Where smoke reaches the ecigc
of a void linking the two levels. some rvill florv over thc
60 edge producing an extensive plume above each void.
rising throu,qh thc uppcr lcvcl.
L i_l
e 40
iii
r ,r- .l
Air mi.res into tlìcsc plunrcs. re sulting in e xtr.cnlclv
large quantities ol very cool sas collecting in the uppcr
level ceiling reservoir. This in turn reduces thc
efficiency of buorancv-drivcrl venting as well as
increasing downrvard mixing lronr the ccitinr¡ laycr. 'lìr
minimise this mixing of air into thc plume. smoke
screens of at least l.-5 m depthli (actuated bv smokc
detectors or as permanent features) shoutd be hung
below the lower level ceiling (ie belorv the upper de ck)
in order to restrict the lateral spread of smoke an<j
ensure that all the smoke from a fire passes through
only one void (Figures 9 and 10). The approach
outlined in this section will only apply rvhen rhe void is
small enough in relation to the mall. and the screerìs
are arranged so that smoke can flow freely into the
45678 void around its perimeter in a 'swirling' pattern
Heighr of smoke base - m
(Figure 9).
Figure 6 Rltc trl producrion Of hot smokv gascs in
a sinqlc-storcv nralllronr a -5MW fire The perimeter of the rising plume. rvhich strongly'
affects the rate of the mixing of air into the plumc.
then depends onll'on the size of the void. Experiments
with scale modelslT suggest that the largest void
consistent with this type of smokè control svstem
rvould have a perimeter of betrvce n -3-5 and 4-5m.

There are no adequate theories to relate the quanrirics

ol smoke entering the reservoirs. to (he heieht from
the upper tloor to the smoke base. Figure I I shows the
values ol the mass flow rate (M) entering the rescrvoir
for different heights of ceiling reservoir smoke base
ahove the upper floor. and for three differenr. sizes of
void (indicated by their respective perimeters), based
on the voids used in the scale-model experiments.
Other void perimeters must be inferred by
interpolation. --F

Figure I I applies to a 5 MW fire occurring in a shop orr

the loler levels. and is derived from experiments on a
Rgurc 7 Atl<Jitional cntrainrncnt rvith lhc smokc basc onc-tenth scale modellT representing a mall of 5 m
front opcn¡nq
rYqll ¡lr1lvg tlrc shop floor-to-floor height. Note that extrapolating too fâr

8
 illUrne -Snroke plume
I
I

Shoo on f ire

Figurc ll Plan vie s'ol s¡nr¡kc circulatit¡n pattcrn bcl()rv u¡r¡:cr dcck.
Without smokc-rcstraininq scrccns

Mal

Shop on fire

Figurc 9 Plan vicrv of smokc circulation pattern bclow upper <Jcck

With smoku-rcstruininq scrccns

Ceilrng reservorr
Exhaust

.- R eser vor.
screen

i::F:.'i.
:::j:::::lÌ
Ris ing
prume
4- <-
i-fuV.d _¡\\ <-¡
Smoke screen Shop on f ire

Figurc l0 Smoke and air movcmcnts ¡n two-storcy mall rvith small

conncctine voids

9
 130

120 .l
J=64m
l
v)
.. J =32m i
s¿.

' ttu
I
-- ¡
¡

. J=22ñ
Ø
I
I

o 1ôn
;,

(ovv
orì

ì 'l
;
I
I
I

6 Ral I

6--
lrl
Where J = perimeter of void
I pass¡ng smoke I

60
25 3 3.5 4 4.5
Height of rise through upper level (dç) - m

Figure ll ñ{ass ot smokv uas cntering cciling rcscrvoir pcr scc()no

lrom a 5!lW firc small voi<js
-

will risk an increasing (but unknown) error. sincc thcsc L'ntrainnlent of air that takes place mav result in verv
curves âre purely empirical. large quantities of cool smoke.

The fire on the upper level can be rreated for design Screens installed beneath the balcony to channel rhe
purposes as a fire in a single-storey mall. If the snroke smoke from the shopfront to the edge of the balcony
ventilation system will cope with the smoke from a fire will restrict lateral spread of smoke (Figure l3).
on the lower level, it should usually be able to copc thereb.v producing a more compact line plume. This
with fires on the upper level. Note, horvever. that results in less entrainment of air and therefore a more
upper level fires can result in higher remperature manageable quantity of hotter smoke. These screens
smoke in the reserv<;ir. can be permanent structures or. alternatively. screens
activated automatically on smoke detection.
'rwo-storey malls Ìvith large ì,oids Recent. as yet unpublished. researchlr. suggests that
channelling screens may be unnecessary if the balcony
vVhen a shopping centrc has large voids throuqh thc
projects no moie than 1.5 m beyond the shop front.
mall connecting lcvels. the snroke flou' is sttrncrvlrât Thc nlinimum depth required for a pair of rhese
different from that gir.en earlier for small voids. An screens to channel all the snrr:ke is dependent on thcir
idealised model of a two-storey mall with large voids is separaricn at the void edge (L). Some values for a 5
shorvn in Figure 12. rvhr..re the upper pedcstriatì
ñlW fire in a typical mall are siven in Table la f<-rr a
walkwavs take the lbrnr of a balcony on cither sidc ol'a
balcon¡'rvith a downstand at the void cdge. rvhich is
large void.
deep in relation to the approaching laver. and Table I tr
for a balcony without a downstand at the void edge
Again. the 'worst case' is rvhcn a fire occurs in a shop (based on Morganre). If sidervays spillage occurs in
on the lorver level. Snloke rvill llow from the shoplront quäntity, the visibility on the upper level can be
to the underside of the kralcony. f r rvill then t'low significantly worse.
forward to the balcony edge. as well as sidcrvays
beneath theåalconv (Figure l2). Because of this Thc mass flow rate entering the reseryqif ç¡$. in
sideways flow beneath the balcony, lhcre nray bc an principle. be related to the height from the balcony to
extensive line plume flowing upwards frorn the halcony the reservoir layer base and to the plume length (ie to
edge like an 'inverted rvaterfall'. lhe screen separation) by the theory of Morgan and
Marshalllt¡. Unfortunately. this theory applies to a
Unless the length of this line ¡rlume florving over rhc plume rising through free space. where the air outsíde
edge of the balcony is limitccl. thc cxtcnsivc the plume is uniformly at anlbicnt temperature.

10
 r Smoke plume . Snroke plume

Void Mall

Balcony Ba lcony

Shop on f ire

(a ) Plan

Exhaust

Ceiling
re s et voir
Smoke plume

Upper level

Lower level

F'igurc 12 Sm<lkc sprcad bcncath a balcony producing

(b) Section a long linc plumc

The smoke layer in a mall ceiling reservoir does not Results from such calculations for a number of valucs
have a well defined base (especially in a two-storev ol' channelling screen separation. are shown graphicalll'
mall rvhere there might be a deep layer of cool smoke). in Figurc 1.5. These results include an allowance for
Even below the nominál layer base (ie d¡ below the cnt.rainment of air into thc cnds ol'the plurne. This
ceiling. Figure 13, which corresponds closely to the corrcction rnethod follorvs that t-¡l Morgan and
visible. layer baselT). there is a temperature excess Marshall:t. which supersedes the ir earlier approach:t'.
relative to the temperature of the incomin-q air which
increases with height (Figure l4). Oncc the height of the nonrinal laver base (h - cl¡) lras
bccn chosen on safety grounds. and the channelling
To apply the theory2tt to a calculation of the mass flow screens separation L (and hence also channelling
rates of smoke entering the reservoir. one must screen depth using Tables la or I b) has been chosen on
introduce a correction factor for the smoke layer dcpth practicability grounds, (eg such that the screens contact
in a reservoir. Experiments with flat roofed models:l the walls separating the shop) then Figure l5 can lre
have shown that for calculating plume entrainment. the used to find the mass flow rate of smoke cntering the
effeõfive [ayer depth (d2) is 1.26 times the nominal and reservotr.
visible layer depth (d,) which has been chosen for
reasons of visibility and safety. ln practice many It should be noted that once the height of the layer
shopping malls have roofs which are not flat. and it is base (h-d') has been selected, d¡ can be used instead ol'
necessary to assess the result of this on the effective d2 tbr greater simplicity in interpreting Figure l5 ancl
layer depth. This is examined more closely in in designing the consequent smoke extraction. This
Chapter 4. rcsults in an 'overdesign' of thc cxtract capacity. whiclt

ll
 +l+
l-l - Srrrokrz p¡r,."
I

VO ,ö/
aI
rá.t
''.t,€,,.1ä,a J
i,r3t
í e,i,li
8a rcony
ttt 8a rcon y

ÏIÏ
, ræ,.
Chan nel I rng screen
*s
Shop on fire
Channel I rng sc reen

(a) Plan

Exhaust

r ---ì------
,¡
I Layer
Effsç¡¡ys dep¡h
layer deÞth d1 Smoke plume
Ba lcon y

Channelling r.r."n/
(b) Sectron Figure 13 The usc ofsmoke-channelling screens to producc a
compact rising plumc

thus errs on the side of safetv. In experiments values ol. should bc measured from the bottom of the
( ,) as low as a full-scale equivalent 0.75 m rvere downstand. Figure' 15 can a-eain be used to estlntate the
obtained. For a ve rv deep laver. one sometimes fincls untrainnlc¡tt into thc plume, but a nrore Drectse
that (h-d:) can sometimcs be negative. This c¿rlculation lor this case is feasiblc::.
corresponds to a plume nroving do$,nwards rvhich is
impossible in this conrexr and shows that the merhod S.irnilarlv. resulrs given in Figure l5 are for a line plume
breaks dorvn undcr these conditions. It lollows that risins Ihrou_qh an upper level whcrc air is entrainecl on
rvlìc'revcr (h-dr) is less than 0.75 m, it rvould be salc rcr botll sidcs. Il the plunte is rising againsr a verrical
use (h-d,) instead for estimaring entrainment. surlace (such as a wall or shopfront on the level
ahrovc). then air rr'ill only be enrrained into one side.
The resulrs given in Figure l-5 are representative of Reu^ent research ry6¡lç21.?.1 has enabled a more detailed
typical shop/balcony geometries. In practice the analysis of the fire compartment conditions and
shopfront geometry. presence or absence of a deep subsequcnt plunte entrainment to be carrietl out rakin_g
tlorvnstand fascia and a balcony rvill affect the ntass i¡rl.tl account. these factors. This fire engineering
tlow rate of sn¡cke. For example. many malls rvill havc rupproach is oI necessity more complicatedsnáneeds
the upper walkway ser back above the shop units on individual co¡rsideration.
(he storey below with no balcony projecting
beyond
the lower shop front. ln such designs the shop rvalls A¡l alternative method of calculating the entrainment
themselves act as channelling screens. Where such a rrìto thc linc plunre is due to Thomas:r. This treats the
shop has a downstand fascia. the plume.s rise (h-<J,) plunrc in a 'far plume' approximation apparently rising
t2
1ìrhle la 1'li¡rinrrrr¡l dcptlr of'chunnclling scrccns
cluwrrstand at v<licl e dgc
-
Scrcen scpurulion ' ñlinin¡um scrccn
ät edge L dcpth
(nr¡ (m) Fascia Downstand

4.l ! 2.'l
6j2.4
8i2.2
l0 l.I
r.1 I Lr,r

Note: Thc minimum <jepth is thar bclow thc Downstand at void edge
lowest transverse obstaclc

Tablc I b Minimum depth of channelling screens

no downstand at void edge
-
Scteen separation ùlini¡num screen
at edge L depth
(m) ; (m)

4 , 1.6
6 t-4
I 1.3
r0 I t.l
l¿1 i l.l
Note: The minimum deprh is rhat below thc No downstand at void edge
lowest transverse obstacle

lrom a line source of zero thickness some dist{nce iVlorgan's re-analysisl5 of Lau,\ earlier paperlo,
below the void edge. The relevant formula is: concluding that the effective depth d, of the reservoir
layer should be used *'hen the plume rises within a
qrl closed space such as a mall. sllrluld appty equally to
vr = o.sso[e-ç!a
'[ocnrìl (z + a) (3) Thomas s rvork. This meâns that z shoulcl be taken to
be (h-d2). as earlier in this section. Again from
l-. 0.zl(z + zL\1%
X rlr
Morgan:s. one can take A = 0.3 times the height of the
LLI compartment (ie shop unit) opening, although
comparisons between Equation 3 and the modified
form2z of Morgan and Marshalli methodzo su_egest that
where lvl = mass flow of smoky gases passing height the value of A may be sensitive to the parameters of
z (kg s't) the horizontal flow approaching the void. ln this
tr- density of warm gases at height z (kg m'r) context it is noteworthl' that Law:ó derived a value of A
rl heat flux in gases (kW) = 0.67 times the height of the compartment based on
t- length of void edge past which gases spill different experimental data. -+ F
(m)
(. specific heat of air (kJ kg-t ç't¡ It should be realised thât the derivation of Equation 3
'rrl absolute ambient temperature (K) limits its application to scenarios where smoky gases
empirical height of virtual source be[ow issue directly from the compartment on fire, with a
void edge (m) balcony projecting beyond. For two-storey malls
height above void edge (m) Equation 3 and Figure l5 should give broadly similar
IJ
 _r--
lr

,-
I

I
I

/
I
7
t)
I
/
o q

Nom¡nal layer base

4
I

¿345679
Temperature above ambient - oC

Figure t4 .Ä. rvpical ,.rnp"rutur. profilc for a rescrr.oir lavcr

results, bur undcr some circumstances significanr down from the void edgcs. thc region below the upper
discrepancies can occur because of the apparent deck can be turned into a ceilin_g reservoir similar to
variability of Ne\ertheless. Equation 3 can be a very that of a single-storey nrall. albeit a more complicarecl
useful wa), of ^.
estimating entrainment for geometries geometry.
departing significanrlv from Figure 15.
This reservoir can then be provided *'ith its orvn
Multi-storey malls smoke extraction system. Other screens can be
From thc previous section it can be seen that rvhen a positioned across the rnall to limit the size of this
fire occurs in a ground floor shop and rises through an reservoir, as for a single-storev mall.
rper level. a verv large quantity of smoke is
If this is estended to a three-storey nrall the The screens around thc voids rvill. in qeneral. be fairlv
"roduced.
result is an impracticably tdrge. quanritv of verv cool close to potential fire compartments (ie shops). Beinu
smoke. Therelbre it is not usually possible to design a close. smoke issuing [ro¡:r such a compartmcnt sill
practical smoke \enrilation sysrem which allorvs smoke deepen locallv ort nìccting u transverse barrier. Thc
from more than rhe top two levels of.a multi-srorev depth of these screens should take into accounr local
mall to rise up through the mall and mainrain a clear deepenine2T see paee lli. Srnoke remor.ed from
layer for escape on rhe upper level. The limiting factor -
these lower level reservoirs should usual[y be ducted to
here is not thc heicht to the top of the smoke rcservoir. outside the building bur can be ducted inro the ceiling
but the height of rise to the smoke layer base. reservoir of the top l'loor ( Fisure 1tì). The mass tlow
rate of smoke exhaust for the top floor can be
A multi-storev mall can instead be treated as a stack of calculated as if it were a sinele-storev mall.
single-storev malls. rvith each level having a separare
smoke ventilation svsrem. Clearly this technique can There rvill often be some small smoke spillages under
also be usedín a i=r*b-storev mail if so desired. Figures the void screens. rvhich rvill contributffi aaÌogging'oi
l6 to l8 illusrrate in schematic form a mall whose the upper levels. This can be controlled bv permitring
upper floor (two lcrels only are shown in the figures) is some ventilation ol thesc uppcr levcls to operate. to
penetrated by voids rvhich leave a considerable area f'lush out tlris stray smoke. To minimise such spillages
for pedestrians. On the lower level there is a large area and limit the amount of smoke belorv the ceiling layer
situated belorv this upper deck. If screens (activated bv on thc affected level. ir is desirable to simplifv the
smoke dctectors ()r as permânent features) are hunq qc()mctry ot'the ceiling reservoir rvhere possible.The
l4
 L=7m
tt
L=5m
v
l1 90
x

E1
õ
I
ìl 50 I

; I

Ø I
Ø I

ãl 30
---l
I

I
I
I

screen

34567
Effective height of rise (n-02) through upper level _ m

Figure 15 Mass of smokc entcring cciling reservoir pcr second from

a 5ñfW fire larec voids
-

void screens tend to split thc smokc flttrv into scparatc

streams rvithin the reservoir (Figure l6).
Calculating smoke temperature
'lhc tcmpcrature rise of the smokc
layer. above
These streams can meet further along thc rescrvoir. ¿ìs a¡¡rbicnt. is given in Table 2. lor a 5 MW lire (ignoring
shown. Such opposed smoke flows producÈ turbulence anv further cooling) and can be calculated from:
and downward mixing olsmoke into the air belou,. It
o (4)
follows that it is an advantage to keep a simple
MCn
reservoir geometrv. It is also important to provide the
full 'flushing'clean air inflow below rhe ceiline where Q = heat flow rate (kW).
reservoir at. the affected level. 0 = temperature rise of the smoke layer
above ambient ffi).a
A lesser amount of 'l'lushing air florv' is desiratrle on cn = specific hcat capacity of the gases
the top level when lower. reservoirs are vented lo a (kJ/kgK).
common top level reservoir (Figurc I8). This can easilv M = mass tlow rate of smoky gases (kg/s)
be provided by increasing the smoke extraction from
tlìe top reservoir by l0% above tlrat needed t() remove Hieh snloke layer temperature rvill result in intense
all the smoke arriving lrom bckrw. hcat radiation causing dil'licultics t-or people cscaping

l5
 Reservo¡r
sc reen .Vord-edge screens ,, f Reservorr
I screen
?

Local lurbulence Local turbulence

and mixirrg and rnrxing
Shop on fire

-+>- Drrection of smoke f low

tr Smoke extraction f rom ceiling reservoir formed by screens

F-igure 16 Schematic plan of multi-storey mall with a smokc rcscrvoir

on cach lcvel

Smoke t x haust Smoke e xhaust

Smoke extract duct

Upper level
Lower level reservoir{
Void -eoge screen
Lower level

Figure 17 Schcmatic secrion of a trvo-storcv mall rvith a smoke rcscrvoir

on the lowcr level

Smoke exhaust

Upper leve
Lower t"uut ,"s".uoirJ
-F

Lower level

Figure 18 Schcnratic scction of a two-storcv rnall *.ith lt:wcr smokc

' rescrvoir vcntinq into thc upper rcscrvoir

l6
 Iì¡hlc 2 vrlrlr¡¡c llr¡*'r:¡rc .,nd tcmpcrature r¡f'guscs l'ron¡ u 5 iVl!v lire
. (igrtoring cooling)

i\las.r lhw r:rtc Tcnrpera(urc of gasrs Volunrc f.J(c of crtrtcli()tr

(,11:¡ss ¡:rtc of cttritrtion) ahoyc amllicnt (:rt rrr:¡xi¡rrr¡m tcntpcntlurc )
klr'. "C mVs

a
556 2l .5
t2 4t7 at
I5 333 26
t{J 278 29
20r,r -t+
:ì() to/ 39
3ó 139 43
.s0 t00 55
65 71 h/
It0 OJ 79
95 )J 9t.5
It0 45 t0-3.5
l:10 Jð r20
l-50 JJ ¡ -JO

beneath the smoke laver in the mall. To reduce the Sprinkler cooling can be uscd in rhc malls ro recluce
intensity of lìcat r¿rdiation rhc srnoke layer the snroke laycl tcr.n¡rcraturc t() bck¡rv 200oC. abovc
temperature in rhe rnalls should be less than 200oC. In rvhich heat radiation f'rom rhc. laver is likcly to impecJe
general a higher clear laycr beneath the smoke layer
escape beneath.
will lead to more air being enrrained into the risine
smoke plume and therefore lower smoke A natural ventilation svstem relies r¡n the buovancv ol
temperarures. If rhe height of the mall is restricted. thc smoke for exrraction. rherc.forc if sprinklei .ooiing
then it may not lre practical to increase the clear laver is underestimated, the use of unrealisrically high ,*oÈ.
height in order to reduce the smoke layer temperaiure. temperatures could lead to the svstem beins
in which case considerarion ma-v be given to installing un<Jerdesigned. Converselv a porvered extrãct system.
sprinklers inro the malls specifically to reduce rhe to a reasonable approximation. removes a fixed
smoke layer temperature by sprinkler cooling. This is volume of smoke irrespective ol' remperature.
discussed further in the following section.
Therefore if the extent of sprinkler cooline is
overestimated the system could be underd-esigned. The
Mall sprinklers heat lost from smoky gases to sprinklers in the mall is
Sprinklers in shops are an essenrial part ol the smoke currentlv the subject of research although data suitable
ventilation design in order to prevent a fire growing lor design application is nor ver availabie.
bey,ond the design fire size. Sprinkler operaiion in the
malls will lead to increased heat loss reducing the An approximate calculation ap¡troach can be used. as
buoyancy of snroke, rvhich i¡r turn can contribute to a Iollorvs:
progressive loss of visibility under the smoky layer.
If the smoke passinu a sprinklcr is hotrer than the
However. gases sufficienrly hor enough to sér off
sprinkle r tlpcrating tenlperilturc tlrat sprinklcr w,ill
sprinklers will remain initialty as a lhermallv buovanr
evcrìtu¿rllv be set off. The sprinkler spray rvill then cool
layer under the ceiling.
thc smtlkc. lf thc smoke is still hor enOush. thc ncxt
sprinklcr rvill operate, coolins rhe snroke l'urther. A
When the fire occurs in a shop. operation of sprinklers
stagc rvill bc reachcd when the snroke tcmperature is
in the rnall will nor assist in controlling it. lf too manv
insufl'icient ro set off furrhcr sprinklers. The smoke
sprinklers operated in the malls sprinklers in rhe shops
lavcr temperaf urc can tlrereafrer símply be assumed to
could become less effective as thc available rvarer
bc approximatelv equal to the sprinkler operating
supply approaches its limits.
temperature if natural vents are used.
Mallssirouhd be sprinklered if they conrain sufficient Thc cooling eftèct of sprinklersìr1!lhFlnalls can be
combusribles to support a fire lars,er than the dcsi_qn ignorecl in determining thc volu¡nc cxtract ratc
fire size of 5 MW l2 m perimeter. durine therr required for fans. This rvill err on thc side of siìfetv.
operational lifetime. Note however thatìorinklers Alternativclv this cooling and the consequcnt
installed at high level in a multi-storev rn"ll ur. c<lntraction of the snroky gases can be appro.rimately
unlikely to operate unless the fire size reached is much cst¡rìrated on thc basis of an average value betrveen the
larger than this. sprirrklcr operatinq tempcraturc and thc calculated

TI
 ini(ilrl srrloke tc!ììllcnlturc. sincc ir l'irc nllrr ,rcc[¡r closc strr.¡ctur¿tl l'lc¿rrlrs tlr cluclrvork) r;rthcr tll¿rrr tlle t r.rre
Lr) ltt lc¿tst rrrlc 0l tllc cxtract Opcrrings, rvhcleils tllc ceilirrg. Wlrcrc such structu¡.cs c.rist ¡nd irre ¿r¡r
othcr o[)cninqs n]ay bc yell outsidc any probablc zone apprcciablc t'raction ol' thc ovcrall l¿rvcr cle pth. thc
ol opcratinu sprinklers. The number oi potcnrial ,lrot, dcpth be low thc obstacle should bc l'ou¡ld usirrq'[ìrhle
and 'cool' intakes ¡nust []e assessed separatclv lor cach ib rarhcr than 3a.
sho¡l¡ling contplcx whcn calculating ttrc ,,u"r,,g"
temperature of extracted gases. It is alsr¡ neccssarv in s<lrrre malls to dete rmirrc thc
florving layer depth bctwccn clrannelling scrcens
If the sprinkler operating temperature is set high beneath balconies. This l'lowing laycr depth rvill bc
enough and is abt¡ve the calculated smokc laycr altered by {.he prescncc or abscrrcc of a clownsia¡lcl
temperature. then sprinkler cooling in the rnalls can be which is deep in cornparison ro the approaching lavcr.
ignored. If there is a deep downstand at the balcony eqgc usc
Table lå-
Note that the effect of sprinkler cooling is to reduce
the heat flux Q without significantly changing the mass Local deepening
t'lux. It follorvs that once a new vqlue of 0 has been
When a buoyant layer of hot smoke flows along
es(imated. the new heat flux can be found usine
beneath the ceiling and meets a transverse barrier it
Equation 4.
deepens locally against that barrier2z; the kinetic
energy o[ the approaching layer is converted to
Note that .spritrklerr- mu:it trot be installcrl clr¡se tt¡
v e tt t i lato r e -yI r(tc I
buoyant potential energy against the barricr as utc
¡t <t s i t io tts.
gases are brought to a halt.
¡-'rwing layer depth
Smoke enterins the ceiling reservoir will flow from thc,
point of entrr torvards the vcnt or fans. -l'his flou,is Table 3a Minimum reservr¡ir depth (m)
driven by the buoyancy of the smoke. Even if therc is a
very large ventilation area downstream (eg if the M¡ss flow r¿te I *¡dth of rescrvoir (m)*
downstream roof rvere [o be removed) this flos,ing entering reservoir I

.f t5
layer would srill have a depth relared to the rvidth of
the mall. the temperature of the smoke and the mass t0 tt 0.8 a.7 0.6 05 0*s
tlow rate of sm<-¡ke. Work by Morganre has shorvn that 15 IA
t.l 0.9 0.8 0:Ì 0.6
this deprh can be calculated for r,lnidirectional flow 20 t.t 13 Ll 0.9 0.8 0.7
under a flat ceiling. as follows:
75 2.0 15 t2 1.1 1.0 0.{l
JtJ ?3 t.7 1.4 12 l.t 0.9
40 ¿8 u 1.8 15 1.4 12
. I Mr
-' -Lt
d -t
I

!
l lrl
t-''
I
(-5)
50
70
J.+
45
L6
3.4
LL
28
1.8
z4
t.6
22
t.¡r
t.9
e r1: w -l
90 5.6 43 35 3.1 z't u
lr0 o-t 5.1 42 3.ó 33 ¿8
rvhere dr = 130 t.ó 60 4.9 42 3.8 32
florving smoke layer depth (m) 150 9.0 6.8 5.6 4.9 43 31
T. = absolute temperature of the smoke
laver (K) ' For bi.directional flow of smoky gases rhis should be twice thc actual
rcservoir width
0 - rcnìpcrature rise of the snroke laycr
abovc ambienr (oC)
W= clrannel rvidth (rn)
.( - do$'nstand factor 3ó if deep downstand Table 3 b Minimum reservoir depth deep downstand
is prescnt ar right angles to the t-lorv. across the flow
-
dorvnstand factor 78 if no downstand is
M:¡ss flow rate I width of reservoir (m)*
prcsent at right angles to the tlorv enlering rescrvoir j

(ke/tl I 1 6El0t2t5
fhe resulting value s ol l¿vcr depth tbr differenr
i-eservoir rvidths and mass l'low rates of snroke are t0 l.tì 1.4 t2 1.0 0.9 O.ri

;hown in Table 3a. This ig¡¡ores the effects of cooling in l5 23 1.8 t5 t3 l.l t.0
20 2.8 u 1.8 t5 t.4 l:
:he layer. theretbre iIsprinklers are installed in the ?5 33 25 Z1 t.8 1.6 1.4
nall Equation 5 should he used atier estimating the 30 3.8 L9 L4 ¿1 r.8 1.6
:lfects of such cooling (scc previous section). Each 40 4.7 3.6 3.0 L6 23 L0
Jepth shown irffhis table is the minimum possible
50 5.7 4_3 3.6 3.t 2.7 ?.4
'egardless ol the smc¡ke e xtract.ion
70 . s.8 4.8 4l- 3l 3.1
method enrptoyed 90 9.4 7 2 ó-0 5.t 4J 1.9
jownstreâm: consequentlv it represents il0 It t 8.ó 7.t 6.1 5.1
the minimum 4.1
Jepth for that reservoir. 130 r3.l r0.0 82 7.t 63 5.4
150 r5 ll5 95 82 72 62
fhe depth musr be measured below the lowest * For hi-dirccrional flow of smoky gases this should
be twice thc âctuäl
ransvcrse dounstand obstacle to the flow (ce rescrvoir width

r8
 Whcn dcsigltilr.i lr s¡nokc vcrttilitlion sr.le ru lor- l() (lclcnnlrìc itl')l)t()\ilììille vlrlucs lol. l(ìc:rl tlc(J[re rtin,:
(ho¡tlling ce¡ìtrcs ()l ¡ìror'c
tll¿rrl onc lcrle l. it is olte n
ncccssarv to contr()l thc paLh ol s¡noke I'lorv usirlrl l(t - lntj,,,)
A d.u = (6J
dorvnstand s¡noke curtains. These are tvpically I l-lw
installcd around tlrc cclgc ol'voids Io prcve r]t srnokc
florving up Ilìrough rhc voicls. lf thc void cdgc is ckrse
r," ltc rc A d,* = ¿¡dclition¿rl dcc¡rcnirrc hclorv thc
cstablishctJ srn<lkc laver at thc
to the shop this local deepening could cause snroke to
rransvcrsc barrier (nr)
under-spill the smoke curtain and flow up through thc
void. possibly affecting escape from other storevs. dn, = csrablished layer depth (m)
Clearly, the void edge screens must be deep enoueh to = distancc betrvcen the shop front ancl
the transvcrse barrie r (m)
contain nor only the established layer. but also rhc
additional local deepening outside rhe shop unir on
Note that extrapolating too far will risk an increasing
fire.
(but unknown) error, since equation 6 is a purely
empirical fit to the data and has no rheoretical
The extent of local deepening can be found from
dcrivation. Note also that Equation 6 rtnly applics
Figure 19. The depth of the established laver (d,,, in
strictly to a 5 m high ceiling. Figure 19. on the other
Figure 19) in the matl immediately downsiream of rhc
hand, can readily be scaled.
local deepening must.first be found using the design
procedure given in the preceding sections. Usually this
means in the channel tbrmed bctrveen the voict cttgc Inlet air
screen and the shop front. The additional depth A d,.. There must be adequate replacement air for thc
can then be found by inspection of Fieure 19. allorvins e lficient operation of a snrokc ventilarion svstc¡1. J'hc
the necessary minimum overall depth (d. + ô d.,¡ ol' ratio of ínlet area to extract area is used in ihc
the void edge screen to be found. Alternativcly ií has calculation of natural ventilatioÍt area required to
been shown:x that the lormula given belorv c¿rn bc uscd ircc()unt for the elttct ()f inlct rcstrictions on thc

W=7.5m II

W=1Om

{ fO- wide channel

f 7.5m wide channel
I
-.1'- 5m wide channel
¿

NB, Flow away from barrier is

bidirectional in all cases
N B, Error bars are one standard
deviation

F-igurc l9 Local dccpcning at a transvcrsc barricr

l9
e llicicllcr'()l lhc svstcrn (scc plrgc ll). lf rl,)()r\ iu.r- it can distu¡lr tllc srlrrrke i¿ril irltcrf'ucc ¿rnd clrusc
rcqtl¡rccl It'r ()¡tcn l()t rcplaccn]cllt ¿tir. iln ¿rpl.)r()l)t.[ttc cxccsslvc litss ol visibrlrtr'. ll thc l¿rvc¡.bitsc is tjcsrgrrecl
r:ntrv coclt'icient should be used. Naturul r.e ¡rrs r¡r at lc¿rst 2 rn atrovc Llìe t()p ol'the doors (or air inlcrs)
unalfccte d smoke zones can often be ortcne d thcre is no nceci [o sct ['rack the reservoir scrcerì17.
iìut()nlatic{ìllv. atld sirnultancouslv rr ith tlle nr¿rirl snt<lkc
'u,c¡rtilation svst(] ln extracting smoke. contributc t() tlrr: Anv stagnant rcgiorr i¡r the ct¡ld clcare r air lrcncath thc
total air inlet requircd. If a powered smoke extract snroke laver rvrluld sul'ler l'rom a slelc!r.accumulati<¡n
systern is used, smaller areas for inlet air nlav bc of smc¡ke. Thc l'ire and other major entrainme¡lt sitcs
sufficient. will act as air pumps and cause air to florv from thc
inlets towards thcrnsclvcs. The air inle ts shoulcj
If the area available lor inlet beconles too resrricted. thcrefore be chosen to ensure that thesc florvs ol cold
incoming air t'low through escape doors mav be at too air will flush through all arcas of the malls below tlie
high a velocity for easy escape. Studies ar r.he Building ceiling reservoir. Any smoke wisps that enter tlte lower
Research Establishmentze have shown that rvinds clearer air will thus bc swept back into the main bocty
above 5 m/s can cause discomfort to pedestrians. Such of the hot smoke.
air inflorvs through doors would hinder escapees and
could be dangerous since they mighr alreadl'be A fan-driven inlet air supply can give problems when
predisposed to panic. It rvould perhaps be rvise ro mechanical extracrion is used (the building will usually
design a smoke venrilation system such that air be fairly well sealed in such circumstances). This is
velocities throuch doors are less than 3 m/s. bccause the warmed air taken out rvill have a
-greatcr
volume than the inlet air. As the fire grows and
A high relative vclrJcit-v betrveen the snlokc lavcr ancl declines. the ¡nismatch i¡r volume bctween the inlut ¿rir
oming air occurs rvhen air is drau'n in through an and the extracted fire-rvarmed air will also change.
,..ret of linrited area and the resulting air streanr (or This can result in significant pressure differences
jet) passes below the smoke laver immccliarclr'. S¡nrtkc appcaring across any doors on the escâpe routes. For
will be drarvn dorvn into such a jet b¡.. thc venr.uri this reason simplc 'push-pull' s),'srems should bc
effectrT, causing a sisnificant loss of visibilit¡- in thc avoided.
lower cold air regions: this can occur rvhcn doors are
used for inlet air (Figure 20). It can be minimised by Minimum number of extract points
placing screens defining the end of the reservoir at
least 3 m back from the air inlet. givin_e the intlow an Thc beginning of this chapter outlined rhe imponance
r-¡f distributing extraction points about a smoke
increased cross-secrion and a drop in velocirv. This
measure also permits turbulence in the entering air rescrvoir to prcvent the formation of stagnant cooling
stream (caused bv external winds) to damp our bcfore regions. The number of extraction poinrs within the
reservoir is also important since. for any specified layer
depth, there is a maximum rate at which smoky gases
can enter any individual extract vent. (be it natural,
chimney, or mechanical). Any furrher attempr to
increase the extraction through that vent merely serves
to draw air into the orifice from below the smoke layer.
This is sometimes known as'plug-holing'. Ir follows
that. for efficient extraction. the number oI extract
points must be chosen to ensure that no air is drawn up
in this s,av. Table 4. which is based on experimental
work.r{) subsequently modified bv Heselden.ì1, lists thc
minimum numbers needed for different reservoir
conditions and for a variety of mass flow rates being
Figurc 20a Smokc [r<¡nt a huoyanr cciling lar.cr mi\ins ir¡ro ir
cxtracted from the vcnts in the rcsc'rvoir. Table 4
high vclocitr. air inflorv strictly applies to vents rvhich are small compared to
the laver depth below rhe venrs.

If calculation is preferred to using Table 4, the

[ollo,*'ing apply at the critical point rvhere air is about
R eser voi r to be drawn into the openings.ll:
screen

At an opening. -.ã
eã7
m = q (gd5 Tr 0/T.2)t/: (7)

rvhere m = critical extract rate for efficient venting

Figure 2{}b Mixing is rcduccd hy allorving thc inconlint ¿ir lkrrv at one vent (kg s-t )
to slow bclorc c(lntactinq thc snrokc lavcr cr = l.3tbraventncarawall (kgm.'ì¡
20
lìlrlc { ilvlinilrrttnr nu¡nl)cr of cxtructio¡l pcints nce(lcd ¡tì u
s¡t¡()kc rcscrl'(lir

T()tul rnass l)cpth of lavcr llckrrr cxtructi¡rrr ¡roint (nr)

rrte ()f crtracti0n '---
¡ tsi
!

( ks/s) -í 7 l(,

9 l-l t-t tt

l2 t-l l-l
li tl l-l t-l
IS t-l l-t l-l
2() l-l t-l t-l
2i ,t< tt t-l
-30 44 t-l l-l l-l
40 ñ J+t¡-¿il-t l-l l-t
50 ð-t I 4-5 l2-21 l-r tl t-l l-l
70 ffi12-3lt-2 tt l-l ¡-t
90 34 I 2-2 t-t t-l
ll0 4-s ¡ ?-1 It t-t t-l
130 t-l
r50 3-4
:.-
!-J t-l

Notc: In reading thc above tahle. the first number is for extractio¡r points well
awav from the walls. the second is for extraction ooints closc to thc walls

cr = 1.8 for wall (kg nr'l)

a vent distant fronr a ratc ol extraction is largclv dL'pcnden( upon the depth
-l'hc
g = acceleration due to gravity (ms'l) arrd tem¡reraturc <¡l s¡rlokc. advantage of a natural
d = depth of smoke layer below thc vent (m) vcntilation system is that it is very simple and reliable.
Tr = ambient absolute temperature (K) and can cope with a s'ide range clf l'ire conditions.
0 = excess temperature of smoke layer ("c) Should for any reason the firc grorv larger than the
T. = Tt+O' dcsign fire sizc. a greater dcpth and tt: mperature of
s¡noke leads to an increased Èxtraction rate. so to an
The required number of extract vents (N) is then given extent a natural ventilation svstem has a self-
by' compensating mechan ism.
M
N> :-: (8) Care must be taken to ensure that natural ventilators
ttl
are not sited in a position subject to positive wind
where M is the total extract rate required from the pressures (see Chapter 4). If smoke ventilation is
reservoir required in such positions porvered extraction must be
used instead of natural vents. Note that natural vents
Where verv larqe or physically extensive vents are used and powered extracts should never be used together in
(eg a long intake grill in the side of a horizontal duct) the same reservoir.
an alternative method is possible. For this case. Table
3a can be used rvith the 'rvidth of reservoir' being takerr Table 5 gives thc rninimum acrotlvnanlic free.aiea of
as the total horizontal accessible perirneter of all thc vcntilatiorr require d. i-l¡nori¡rg thc efl'cct ilf anv inlet
vents u,ithin the reservoir (eg the total len-sth of intake restriction. To allorv l'or thc' cffcct ol lirnited lresh air
grilles in the example above) and the 'minimum inlets the follorving Suidc can bc uscd:
reservoir depth' corresponds to the depth of the snltlke
Ìa_ver beneath the rop cdge of the intake orifice. In If thc inlct area for thc u'liolc ¡rrall is thc same as the
practice for a given mass florv rate and layer dcpth one vcnt ¿ìrea lor thc rcscrvoir givcn hv Table 5. tltis
can use.Table 3a or Equation 5 to find the minimum indicated vent area should be irrcreased by
value ol' accessible pcrinteter. approxinìalr: ly 359á.

Intermediate size intakes (ie where the vent size is It the total inlet area is trvice thc rcservtlir vclìt arcll.
comparable to the layer depth) cannot be treated so the indicated vent arcn should be increased by l0'7ô'
simply and it is recommended that Table 4 be usecl
sincgit errs on the side of safetv" The precise relationship betrvg tþ mass flow rate
extracted, t.he vent area. the inlõt a-rea and the smoke
layer isll:
Natural ventilation area required
per reservoir -
A natural ventilation system uses the buoyancy of the A*c" = lìr + (A* C. r.l ''
/Aici)rT, ''l (e)
smoke to provide the driving.force lor extracticln. The #[ 29 dn 0. l, j
2l
\\lìcte clr¡rlrlrlc rrl rr ithst;rntling thc lrrrtrci¡rlrted rnlrrke
\,C, = ,rcrodvnilrììic trce ilre¿.ì ol nlrturill tc rìr l)c r¿l t u t'cs.
vcntilation (nrl) 'fhc contrttls irrrd rviring shoultj
ol coursc bc ¡rr-orccrcu.
= rncasured throat arca of ventilators fitr thc to nlairrtain the electrical supplv to thc l¿rns durins
rcservoir bcing considcreci (nr:¡ ¿r

[i rc.
¡ \| = tr¡tal arca of all inlets (nr:)
= coefl'icierrt of disclrarge (usua llv.betrvccn Thc nrass l'lorv ¡'ate ol'smoke deterntined lronr rhc
0.5 and 0.7)
previous section c¿r¡r lre corrvertccl to the corrcs-
Li - entrv coefficients for inlets (tvpicallv about
ponding volurnc fltlrv ratc irnd tcrnpe rature. usinr
0.6)
Tablc 2 (clr thc [olkru,ins equarion) lor sclection ol'rhc
lvl = mass [ou, rate of smoke to be extracred (kg/s) appropriate l'ans:
Po = ambienr air density (kg/m:¡
o
= acceleration due to gravity (m/s:)
)vb = depth of smoke beneath ventilator (m) v=M .i- ilo)
A
= tenlperature rise of smoke la¡.'er above P,,T,,
ambient ('C)
Tc = absolur.e temperature of smoke layer (K) rvhere V = volumc l'lorv rate of gases to be extractecl
Tu = absolure tenlperarure of ambient air ( K) from the srnoke-layer (m:/s)

Natural ventilarion can sometimes be enhanced by

using chimneys to increase the buoyanr head of hor
smoke. Hr)rvevcr. a svst.c¡n of chimnev vents can bc Table 5 The miniu¡unr total vent are¡ (mI needed in r¡ne
'"icult
to design. The florv resistance must be taken ceiling reservoir (from Equation 5 with C" = 0.ó)
r..-o account. T1'pical values may be found in thc
Mass rale i Smoke dcpth hencath lents (ml
CIBSE Cuide¡:. hut it should be noted that ofexlr¿ction i-
unpublished experirnents at FRS have shown rhe cnrrv (kt/s) l_i
coefficient to take on a value of approximately 0.T (as
compared rvith the more usual value of 0.5 used in
4.E .1.{ 1.9 l.o ' !.¡i
t2 6.1
.;- 5..'ì -1.7 :.¡i 1l
HEVAC calculatio¡ls) for a sharp-ed-eed opening in t-5 / -_l o.) 5.i 4.ó 1.9
drarving gases fronr a relativelv shallow buoyant laver. l¡ì 9.0 7.ri 6.J ,i.5 -1.9 ' J.l -i. )
?J t?.2 I0..5 $.6 7._s Ò./ '' 5.6
30 r5.ó I -1.:l II 0 9.5 8.-5
?a 6.()
A powered extract system should be used where 3ó 19.? I O.O r 3.6 I t.S 10.5 ; a.z 1t
positive wind pressures are likely to be a problem. or 50 l.i 2t) 17._s l)./ 13.2 ll.t
r.,''here it is necessary to extract smoke via an extensive ó0 25 21.0 t9.7 lo./ ' IJ.Y
t) -ì.1 l9 7) . 18.6
ductrvork system. 90 34 ¿ó
ll0 57 19 JI
Powered ventilation 130 7t 62 55 :39
r50 87 75 o/ iqS
A powered smoke extract svstem consists of fans and
associated ductwork designed to remove the mass fiow Notc: Add on l0o/" if thc rora¡ inlct arca in rhe mall is tu'ice rne venr àrca.
Add on -15o/o if the toral air inlcr area in rhe mall is eoual ro the
rate of smoke entering the smoke reservoir. and to be vent area

22
Chapter 3
Large shop opening onto a mall

As outlincd irr Chirptcr 2. thc prcfcrrcc.l opti<lrr tìrr rhc tcnl[)craturc rcqu¡rcr]ìctìts Irtr I¿rns artcl drrcts¡
nrajority ol'sh<l¡rs is to provide a comnlon smokc
ventilation svstcm in the nlalls unless the shops are An alternative, pitrticul¿trlv appropriittc wlìcrc thc
grcater than l(XX) nr: (il the mall is naturally vcnte<J) or total width ol o¡renines bctwce n the storc ancl thc rl¿rll
l-j00 rn: (if tlrc rnall lras porvcrecl extraction). Shops arc restrictcd. is to providc l¿rrgc ca¡racity 'slir'
larger than these sizes necd adtlitional measurcs to cxtraction-ì:l in the cciling ovcr rlrc wholc width <ll suclt
protect the nlall. openings. including <Joors, but not inclu<ling fixed glass
windows (Figure 22). Such a system is likely ro work
A smoke layer rvitlrin a large shop will lose heat as best with further extractiorì <Jistributed witlrin thc
smoke spreads throughout rhe store. This heat loss is store. which may possihlv tre provided by the normal
caused by' thc cooling action of the sprinkler spray ancl ventilation extraction systcn't. rvith thc nornlal
heat loss bv conductìorr and radiation to the building ventilation input and rccirculation of air bcine stoppcd.
structure. Thus the cooling is related to the size of thc Whilst this system is designe<J to prevent cool smokc
shop. If there is too nluch heat loss, the effect in the cntering the mall. it will not rrcccssarily nraint¿rirr a
mall will be as if thc mall reservoir itself were too clear layer within the store itsclf. The exrraction should
large. sincc there rr.ill asain be excessive downwarcl be provided very close t<l thc opcning l'ronr a
nrixing and loss oI r,isihilitv in the mall. To ensurc sal'e continuous slit rvhich nlirv bc situatccl i¡r thc irlallc ol
contlitions in thc ¡nall thcrefore. smoke from lar_l,e
stores nlust be prcvcntcd from florving onto it. To
achievc this thc store nlust he eithe r isolated from thc
mall or har c its orvn snlokc ventila{.ion systeln. Firc
shuttcrs should onlr' ['rc used to isolate the store frorn

- and if
the mall if thcrc is no other practicable choice
thð im¡llicatirlns tìrr n.ìcarìs ol escape and the possible
psvcholcleical c-ffect on escaper:s have beerr fullv takcn
into account.

This excessive cooling o{'snroke is thought only to be q

serious in stores over 1000 ml in area (or 1300 m: if
porvered extraction is used in the rnall). If they are not I

isolated from the mall. such large stores should have (¡)
ø
ceiling reservoirs formed by similar methods to those
for malls. Since no smoke is allowed to enter the mall.
x
each resen'oir can be linlired to 2000 m2 (or 2600 m: if
E
porvercd extraction is used) to prevent excessive
cooling. ic to the same m¿rrinrum area as the combined
nlarima for the mall plus ¡r small shop unit. Note that
a¡rv ¿rrea less than ItX)O m:'lcft over'from this internal
sub-dir isitrn. antl adjaccnr ro thc nrall. can (hen be ':
ventilated lil the nrall in thc samc,way that a small
shilJr rr.ould Lre. Thc' qu¡lnti[v of snroke entering the
ccilinq rescn'oir rvithin the shop is given by Equation
l. ancl rcsults for a 5 l\.lW t'irc are shown graphicallv in
o
Figurc 21. Havinq dctcrminetl the mass flow rate of
snroke. the tJesign proccdurc given for the mall
vcntilation s\.stcnì in Chaptcr 2 can be used ttl
dctcr¡nine the e.\tract capacitv required within the
storc. lt shoultl bc ncltcd that. provicJed smoke cannor
enter thc mall. the hciglrt of the srnoke base in the
shop need not bc the sanre as it rvoul<J be in thc mall.
Whe n*alc+rlating the overall smoke layer temperature
9
thc procetjure for mall sprinkle rs given on page l7 of smoke base-m
shoukl be used (but note rhat any individual intake or
duct rnight bc located inrnrediately above any fire antl Figurt 2l Ratr: of pro<Juclion ol hot snroky gascs in lt
this should be taken into account when specifying thc storc l'rom a 5MW Urc

23
tite llrlsr' ccilirlg.'l'he ùxtrircti()rl ri¡tL' ( \' l cirn lrc lrrurltJ
I'rorn:

'lrfl 0.74V.]
V = : 'j'l l24 - - ,-' 1W,,, rìr ' s-r (l l)
L Hw-'tI
wlrere W = total width of stores g¡rctring (rn)
Wu, = rvidth of storcs openinq onto the mall
(m)
H = he,ight of store opening (m)
Vs = volume extract rate from store
ventilation system (mrls)

Plan

Mall

ll ooo.
Section
Figure 22 Slir exrracrion

24
Chapter 4
Some ¡rractical design considerations

Factors influencing the design fïre The efïect of wind on the efliciency of
Unsprinklered shops a smoke ventilation system
If sprinklers are not insralled in the shop. it is likcly
When natural ve ntilators arc uscd t'c¡r smokc
that the fire will grort l¿¡c(:¡ than the.5 MW. l2 m
c\traction. it is inrportant tllal. thcy arc positi<lnctl
perimeter. design firc;r. An analvsis of lires irr public
rvhere they will not bc a<Jverscly af'fecrcd by extcrnal
areas of retail prenriseslì has suegestcd that fervcr tlr¿rn
rvind conditions. A pr_lsitive rvind pressure can be much
4lo o( fires exceed the design fire area. (Note that thc
greater than the pressure head developed by a smokc
'sprinklered' curve in Figure 23 is drawn to conform to
layer. Should this occur the ventilator may act as ar.t
Morgan and Chandler'slr interpretation of rhe
inlet rather than as an extract. However. if sited in an
statistical data and to havc approxintately the sanrc
area of negative wind pressure. the resultant suction
form as the 'unsprinklered' curve: it can be seen that
force on a natural ventilator rvould assist smoke
the latter is far more reliable). The same statistical
extractions.
probability of being exceeded rvhen sprinklers are nol
present gives a fire size in excess of 100 m: (Figure 23)
Tall buildings or taller areas of rhe same building (such
potentiallv l0 times larser.
- as roof top plant rooms etc) can crcatc a positivc wincl
pressure on lorver nearbv ro<lt.s. Stce plv pitchc<.i r<lol's.
This could leacl to 'tl¿rsitovcr'. s hcn rhe lirc rvill
ic those rvith over 30" pitclr. nl¿rv also havc a positivc
develop very rapidlv ro involve the entire sho¡r.
rvind pressure on the windrvard slope.
Evidence from actual lircs sugqe sts Ihat t]ashovcr cln
happen qu¡ckl\'ìr.
r\ sue._gestion solìtetinlcs aclvanced lor ol'f sctting wintJ
()\'cr-pressures is to incre asc thc f otal area of natural
A fully involved shop tire in a 'rvpical'slrop unir nrav ventilation per reservoir. Since the over-pressure is. bv
not have a large enoueh openine onto thc rnall for
dcf inition. force per unit area. this rvill usually not
sulficient oxygen to be available ro support complere
rvork and indeed could exacerbate the problem by
combustion rvithin thc shop. l'his rvill result in unLrurnr
allorving even greater quantities of air to be driven
gases flowing from thc shop. nririnc with air and
through the vent to mix into the smoke. A powered
further combustion raking place in rhe mall itse lt.
smoke extraction systenl should be designed to
resulting in considerablc downivard heat raciiati<¡n.
overcome the anticipated wind pressures.
Unless the màlt roof is high it is unlike [y to be practical
to design a smoke venrilation svsrem that will ensure In some cases it may be pos.sible to retain natural
safe escape, for people in nornlal clothing. below the ventilation openings in a vertical plane by arranging
mall ceiling smoke reservolr. them to face inrvards to either a region sheltered from
rvind action, or rvhere the wind rvill ahvays produce a
Clearly if sprinklers are nor installcd in the shops a suction. In other cases the erecrion of suitably designed
5 MW design fire size is inappropriate . and rhe scree ns or rvind baffles (outside the vertical wall or
fornlula. equations and nluch oi the guiclancc given in rvindow holding the vents) can overcome wind
this Report are invalid. l-his is cliscusscd in more dct¿ril i¡rterference and mav even be able to convert an over-
in a paper by Gardnc.r:j. prr'ssure into a suction. Therc is also the.possibilitt, of
Multi-occupancv nalls selectively opening vents in rcsponse t<'r signals from a
As shoppinl centres arc trcconling increasinqlv nlulti- rvind directir)n sensor. Expert advicc sh<¡uld be sousht
use developments. it is intportant trl consider the for such designs.
potential fire risk in occupancics othcr rhan sho¡rs. it'
thev adjoin the rnull u irlrou¡. ¿nr lirc sc¡raration. Due to the complexitv of rvind induced air florv ovcr a
lnformation on fire sizes i¡l sonlc othcr classes of shopping centre and surrounding buildings. it may
occupancies has bcen clcr e lcl¡letl tJuring research intt¡ sornctinres be desirable to carn. out brtundarv layer
smoke control irr atriunl buildinss.ìs..ìô. u,ind tunnel studies to establish the rvind pressure ovcr
the building\ envelope. Once areas of over-pressure
It is importanr to assess the likelv fire size in anv rtl and suction have been identified for all possible wind
these occupancics, arrcl a chcck slrt¡uld be nrade to tesr directic¡ns, design of vents or fans can ¡rroceed as
whether¿ sys¡em desicned o¡t the'basis of a l2 m before. -_ F
perimeter 5 MW fire rrill be atJequirte to de al rvith a
fire from any alternativL' ()ccupa¡tcy opcning onto that False ceilings in the mall
section of the mal[. The nr¿rll snroke ve¡rtilation systenl Where there is an unbroken false ceiling in the mall it
should be designed to cope rvirh thc ntost severe nrust be treated as lhe top of the smoke-layer. If the
consequences of any of thc sclected fire sizes in any of false ceiling is porous to smoke. ie if it has an
thesc occupancies. appreciable free area. anv smoke scrccns forming the
 d
I

Unsprinklered
public areas

t¡

Sprinklered
public areas

l'-igrrre 23 Rctail premises horizonral firc

-
30 40 50
damaged area

Sample At
-'/.

smoke reservoir must be continued above the ceiling. ceiling. In the event oI a [ire a fan of suitably larger
If the proportion of free area is large enough the capacitr.starts up a¡rd draws smoky gases into the
reservoir and its screens may be totally above the false chamber in a similar *'av. This System can ol coursc
'iling. The permeable ceiling ought not to inrerfere also be use<J for malls.
ppreciably w,ith the flor¡, of smoke from the fire to the
smoke ventilation openings above the false ceiling. A potentially r,irluablc honus of such a systenr is that
the sprinklcrs rvlriclr arc normallv required in the spacc
It has been shorvn experinrentafly:r ,¡ut' a minimum above the t'alsc ceiling rvill cool the smoky gasL.s bclbrc
iree area of 25"/" can be used as a .rule ol thumb' value thelr reach thc f¿rn.
for allowing salè cscape. Cool srnoke can sontetimes be
expected to aflect nearby shops but would not The plenum chamber should nol be larger in area than
si-enificantly hinder safe escape. Free areas of less rhan its associated reservoir. Larger chambcrs should be
25Y" are possible in some circumstances; expert advice subdivided by smoke screcns thar are the full height of
should be sought where this possibility is felt desirable. the chamber and rvhich cxtend downwards to form a
complete smoke reservoir below the false ceiline. Thc
The use of a plenum chamber above a mininrum nunrhcr ol opcnings throueh the false cciling
false ceiling in a shop required rvithin a single subdivision can be found front
Table 4. Thc total arJa of such openin$Frescrvoir
Some designs have been seen in which the space above should be decitJcd by consideration of the design
a mainly solid false ceiling in a large store is used for pressure differences between chamber and smoke
the extraction of air tbr normal ventilation purposes. A layer, and of the flow impcdance of the openings
fan extracting air from this space (effectivety a plenum concerned. A system of reasonably wide (perhaps onc
chamber) reduces its pressure and so draws air from or two metres) slots surrounding a region crf false
thc store through a number of openings in the false ceiling could pcrhaps be used instead of screens belou,
26
tltc l'irlse ccilin{ lìr' ;utv givctl re sctvrlir'. ( ;¡rttl tlì()rc ¡'e li¿thlc rnlìrrrrr-
irtirlrt trccol¡tcs ar'¿ttlrrllle . Il scu¡ììs rclsoltul'llu t() ¿lssurììù
It c¿tn bc clc:irahlc ttr lcavc thc Ialsc cciling bckru'the tlr¿rt Lhc cquivalcrrt Il¿tt rool is orrc rrl.¡crc thc lavcr birsc
extraction points 'solid' (ic not ¿rble to pass smoke ) t() is at thc s¿rrnc hcighl. arld tltc lat,er h¿rs thc samc cr<lss-
l')rcvcnt ¿rir []cing tJr¿rrvtr ull through tlrc snl<lkc lrrvcr. A sccti()nal arca. Thc cl'l'cctivc lave r dcpth can thcn be
sul't'icicrrtly cxtcnsive arc¿r t¡l 's<llicl'lalsc cciling rvill tlctcrnlinctl with cl ¡. norv tltc dcpth ol'snroke bcncltll
cnsure that thc smoke passes through at least onc thc cquivalent flat rool' p<lsition.
sprinkler spray en route to the extract.
Snroke flow in low narrow malls
The principlcs rJcscrihed in this section could bc usccl Thc problcms ol s¡ntlke l'ltlrv i¡r lr'¡rv n¿rrrorv malls arc
rvith extraction througll a shaft vcnt (or chimney) frorn rnore often encountcrcd whcn de aling with the
the space above the false ceiling provided that refurbishment of older existins, shopping centres.
sprinklers were not installed in this space, since they Table 3 gives the depth of flowins layer kr malls of
would rob the gases of buoyancy. diffcrent widths and varying mass flow rates. Narrorv
nlalls less than 5 metrcs flo<lr-l.o-flo<lr or rvith deep
Stores with internal voids beams across the mall bencath which the smoke must
When a store of more than one levcl, with internal llorv. could have insulficicnt clear height for escape and
voids. is open to a multi-level mall. smoke flow orrto correspondingly high smoke tempcratures. These high
the mall at more than one level simultaneously should smoke layer temperatures can bc reduced by installing
be avoiclcd. In si.¡rnc circunrstances this nray nreau sprinkle rs in thc malls. spccificirllv (o cool clown the
isolating the store lrom the mall <¡n one or more levels. smoke lavcr.
This can ['re achie ved i¡r ¿r nutnbcr ol'rvays. including
the use o[ fire shutters on dctection of smoke.to iscl[¿rtc 'l-hc snloke layer basc cannot bc raisecl hv incrcasinq
the storc. Whenever thcsc are use d a nunrhe r ol'l'actors the rate ol extraction unless thcrc is suf'ficient floor-to-
nccd ttr bc assesscd. suclr ls the irnplications lor nlcan\ ccilins hcieht to achicvc this. Fronr T¿rhlc --ì it can be
of escape a¡ltl the psrcht'rlclgical cffect on pcoplc t¡i f irc sccn t.lìat iI the sntokc lÏlrvs i¡l trvo dircctirlns in thc
shutters operating. smrlkc reservoir. the flowinq lavcr dcptlr is smallcr tharr
l'or a singlc direction llorv.
Sloping malls
l-his c¿rn be achievecl by ensurirrg. that 5()% ol thc
lf the mall floor slopes significantl¡'. a 3 m clear la.v-er at
onc end ot'the smoke reservoir may be consiclerabl-v- c\tract capacity is inst¿rlled at cither end of the
lcss at the other end. The smoke ventilation system resc'rvoir such that bi-directionrl flow rvill occur. Again
should be designed to maintain an acceptably clear t'rom Table 3 lt can be scen that if thc rnall is made
layer at the high end for a fire in a shop at the lorv end. wider the flowing layer depth is reduced. One method
resulting in a greater height of rise to the smoke layer ol'achieving this without nroving the shop fronts back.
base and mass flow tate of smoke.
u,hich has been used in practice. is shown in Figure 25.
The mall is widened at high level rvithout any change
There mav be a diftèrence in floor level between the at lorver levels by 'stepping back' the shop ironts. For
mall and the shops either side. If so. the height of risc aesthetic reasons this coulcl be above a permeable false
to Lhe smoke layer base should be measured from the cciling. allorving the tìorving la.vcr to exist s'holly above
lorvest shop level and the cle ar la1,er in the malls tllc [alse ceiling line.
should be measured from the his,hest mall level. ':
Basement service levels
Assessment of effective layer depth \\¡hcrc reg,ulatory authorities l'ccl that a smoke
As explained in Chapter 2 the elfective la¡'er de ptlt '"crrtilation system is requircd tilr thc hasement scrvicc
introduces a correction in the procedure for lcvel of a shopping centre . thcre has previouslv bcen
rro ¡rublished guitJancc.
cletermining entrainment. to account for the [ayer ol
rvarmed air beneath the visible s¡noke layer base. 1'ltc
experimental work rvas carried out on a scale nloclel A firc sizc of 7 MW. l5 m perimctcr corresponding to a
rvith a tl¿lt roof as shorvn in Figure 1-1. The depth ol' t'ullv burning tractor (of art articulatcd vehicle) or a
smoke beneath both the ventilator and the ceilins rvas ¡'rartlv burning trailer or a van, has become widelv used
[ound to be the same. as a maximum practical fire size tbr smoke ventilation
dcsign. Note horvever that larger fires are possible for
ln practice pitched roofs. pyramids. domes or llrgc deliverv vchiclcs or trailers. The clesign
rooftithtsare commonly used as smoke reservoirs. No proccdurc is similar to that foc¿rla¡g¡: store q,iven in
experimental data exist on these other rescrvoir lbrnrs Chapter 3.
and using the depth beneath the apex of a pitched rool'.
for example, will increase the elfective laye r deptlt. Firlurc 26 gives mass flow rates for varving clear laver
possibty resulting in an underestimate of the mass flow hciehts. for the basement service level design fire size.
rate of smoke (Fi-eure 24a and 24b). Therefore it is The temperatures and volume extract rates for fans
desirable to determine an equivalcnt llat roof positiort can thcn bc tjctermincd frc'lnr Tat'rk'6. lI natural

27
 Layer
depth Effect
d1 layer
depth
A
w2

Figure 24a Assessment o[cffcctive laycr depth from apex of

pitched roof

Equivalent flat roof

LA ye.r
de pth Effective
dr layer
depth d2

Figure ãb Assessmen( of effectivc layer depth from the equivalent

tla¡ roof oosition

lmpermeable Flowing smoke Mall widened

constructton I ayer. at high level

Shop front

Figure 25 Reducing thc flowing smokc laycr dcpth by widcning

thc mall at hirih level

28
 lrtrilding dcsign. it clrn lru tl()teLl tlìitt ir lirc in ()tìc cil.t\
tunlikclv to s¡tread to ir ncrglrbouling pirrkctl eAr. uve lì
onc rvitlr a"plastic'l'loclv shcll. rvlrcn sprinklcrs have
becn litted in the car pirrk. 'l'his comrncnt should not
bc taken as a gene ral rccorlrrlrcrrcl¿ìtion tlìaI sprinklcrs
are regardcd as cssc¡rtial in all enclttsed car parks.

The design procedure is similar to that for large storcs

given in Chapter 3 si¡tce the perimctcr is the same as
I
lor a shop fire. Figure 2l can be used to obrain rhe
-v, rnass flow rate for varying cle ar laver hei_ehts.
a

Horvever, since the heat flux is 2.5 MW the

x, temperature and volume extract rates must be
o
ú,
determined from Table 7. If natural venriia(ion is to be
used the aerodynamic free area can be determined
from Equation 9.

I Smoke transfer ducts

; Stagnant regions of a smoke reservoir will suffer from
tt continued heat loss resulting in do',r,nrvard mixing into
U'
the air below as discussed in Chaptcr 2. Cood
= clistribution of ventilatol' cxtr¡lct ¡rositions can prevcnt
this being significant. Where this solution is
impracticable smoke transfer ducts can bc installed to
nrc)r'c' smoke fro¡n thc stagnunt rctioll t() arlother part

Table 6 Volume flow rate and temperature of

smoke from a 7 MW 15 m perimeter
fire (ignoring cooling)
i Temperature of i V'olume extrao
o1234: lllass flow I smoke above ambient i rate
of smoke kgs "C : mVs
Height base - m
;

Figure 2ó Mass flow rate of smoke cntering rhe ceiling

9 77rì ; 2E

reservoir from a TN1W iirc. l-ím perimeter

t2 _s83 i j0
t) 467,-ìj
IE .ltig , .ì.i
29? J0
2:ìl i J.i
ventilation is used the aerodvnanlic free area required l{} 175 .i.'ì
can then be tbund lrom Equation 9. -i{) l-10 ol
70 l(X) t-()

Enclosed car parkS

Should regularory aurhorirics require a smoke
t'c'ntilation system for an e nclosed car park. there has
previouslv been no published guirlance. This section Tatlle 7 V<¡lume florv rate and tenrpcrature of
outlines an approach that has bee¡r used tbr some timc smokc lion¡ a 2i lVl!V l2 rn perirncter
hy the Fire Research Starion. firc (ignoring cooling)
A design fire size of 2.5 lvlW l2 m perimerer i Temperaturc of \'()lumc ertr¿ct
corresponding toa single burning car has become fk¡rv smoke abovc urr¡birnl r:¡le
* idely used as the design fire size basetl on a k/r : "C , ml/s
reca lculation of experinre ntal resultslt{. However. 9 27s 15
concen?harbeen expressed about the increasing use ol Ij ¿ils -+ +i3
plastics in car bodies. and of plastic perrol tanks ancl 15 167 l(l
the implications that these rvould have on the design lri . l.ì9 : ll
fire size. Further research is needed to evaluate the
l+ l0-r : l7
io ' ll.ì .11
practical significance of such concern. Whcre the {(} i c.l , .lt
approving aurhorities feel that such aspects might io:50Je
si,lnificantly inlluence their opinion ol any individual 70 : -j(r (rô

29
 +Potentially Stagnant
Region

Figure 27 Use of smoke transfer ducts in otherwise stagnant regions

of the smoke reservoir to rise with an existing florv

towards a vent or extract fan (Figure 27).

As a 'rule of thumb'. if the reservoir continues more

'han 3 times as deep beyond an extract openine as the
:servoir is rvide, then a smoke transfer duct may be
necessary. A currentlv recommended value of the
minimum extract rate is 4% or the smoke lavcr'r net
flow of I mi/s. whichever is the greater.

Entrances within the smoke layer

There have been shopping centres designed rvith
entrances within the smoke layer (eg from car parks
above the shopping levels). Any pedestrian area inside
these entrances and open to the mall needs to be small
enough to be evacuated quickly through smoke sealed
doors.

Other situations
The possibilities for special features are by no means
exhausted by the abor,e. Whenever another special
case occurs thar is not expticitly covered by this
Report. advice should be sought from experts.

-+F

30
Chapter 5
Some operational factors
All rcscrvoir scrccns, srnokc channelling screcns l¡rd ru high lcvcl. l'requentlv rvhcrc snr<lkc would bc l'lorving.
natural vents that âre not permanent features should This air would simply incrcase the quantity of snlokc
operate autontatically, actuated by smoke detectclrs. as that has to be dealt with. Hcncc the norntal ventilation
should mechanical extract systems. Other tvpes of inlet systcm should be autonratically shut down rvhen
detector are less uscful since, at best, they have longcr snroke is detected.
response times and many (eg fusible links) rnay not
operate at all. lt is much easier to prevent smoke f f the smoke to be removed lrom a reservoir is too hot

spreading to escape routes than it is to clear such for the ducting or for the fans, sprinklers or water jets
routes subsequently. If doors are to be relied on as air could be installed in a suirable part of rhe extract
inlets, devices to open them automatically are needecl. ducting to cool the smoke. A temperature rating
quoted for many (but not all) smoke extract fans is tr-r
When non-permanent automatic drop curtains are survive a temperature of 300'C for half an hour. Many
used as reservoir screens the gaps between curtains authorities regard this time as sufficient for escape
should be minimised. If a powered smoke extracr from most malls
system is installed. care should be taken to ensure that - but note that this should not be
regarded as being universally true. All such criteria
smoke leakage through gaps between curtains does nor slr<-ruld be agreed rvith the approving authoritics. All
inadvertentlv activate adjacent powered extract zoncs. e lectrical apparatus and porver supply cables used in
It can be an advantage to open all the natural vents thcse smoke ventilation systenrs should, ol course. be
installed in a shopping centre, since those in zones protected to ensure sustaincd operation in a fire.
unaffected by smoke can contribute to the freslr air Cold smoke tests are sometimcs used for the
inflorv. lVherever it is feared that rvind pressures on ¿rcccpt¿ìncc tesl"in_u t¡f smrtke venti[aticln systems.
vents to unaft'ected reservoirs may be much more Whilst this cold smclke can be used ro operare thc
negative than on vents in a smoke-filled resen'oir. it is smoke detection system and therefore activate all the'
recommended that wind-tunnel studies and/or a co¡nponents of the smoke ventilation system, it should
detailed pressure analysis be carried out to evaluate he noted that since the smoke is cold it would not have
the effect on venting efficiency. When mechanical the buoyancy that smoke in a true fire condition would
extraction is used for a number of separate smoke have. and cannot therefore adequately test the
reservoirs, it would be better to operate the smoke ventilation efficiency of the system.
controI system individually by zones. Facilities for
manual override should be installed to allow greater Smoke ventilation systems should be regularly tesred
tìexibility in firefighting, tesring and maintenance. and adequatelv maintained.

The normal ventilation systems fitted to many. if not

most, malls and shops blow air into the mall or shop at

-.{ Ã

JI
 Ret'erences

Hon¡e Oftice:¡nd Scottish Honre and l{ealth 12 Thr¡nras P H ct al. f ¡lvcstigations in[o thc t]ow ol'
I)c¡rurlnrctrt. Firc precauti<tns in towtì ccntt.e Itot gascs in rool' vcnting. Fira Rt'.tcurclt -lccltttit'ul
redevclopme nt. Fire Pravenrion Guitle I . London. Pa¡rcr No 7. London. HMSO, 196-ì.
HlvlSO. 1972.
t3 Morgan H P and Clrandler S E. Firc sizes ancl
) British Stundards Institution. Fire precaurions i¡l sprinkler eflecriveness in shopping complexes ancl
thc dcsign and consrruc¡ion of buildings. Code ol retail premises. Fire Surveyor, 1981. l0 (5) 23-2,S.
practice for shops. British Standard BS 5588: parr
2: 1985. London, BSt, 1985. L4 Gardner J P. Unsprinklered shopping centres.
Design fire sizes for smokc ventilation. Fire -
Canter D. Studies of human behaviour in fire: Surveyor,1988. 17 (6) 41-47.
entpirical results and their inrplications for
education and design. Building Research l5 Hinkley P L. Rates of production of hot gases in
Establishmenr Report. Carsron, BRE, 1995. roof venting experiments. Fire Safety Journal.
1986. 10 57-65.
Morgan H Sntoke contol methods in enclosed
P.
shoppittg cctrnpletes of one or more storev-s: e t6 Heselden A J M, Wraight H G H and Watrs p R.
design sunrntar\,. Building Research Estahlishment Fire problents of pedestrian precincts. Part 2.
Report. London. HMSO, 1979. Large-scale experiments with a shaft vent. F¡rc
Research Stution Fire Rasearch Nota 954.
British Standards Institution. Fire precautions in Borehamwood. BRE. lc)71.
the design and consrruction of buildings. and part
10: Code of pracrice for enclosed shopping 17 Morgan H B lVlarshall N R and Goldstone B l!I.
complexes. British Standard BS 5588: parr l0: Smoke hazards in covered multi-level shopping
London. BSI. To be published. malls: some studies using a model two-storey mall.
Building Reseurch Estublishment C¿¿rrent Paper
Heselden A J ùf. Fire problems of pedestrian CP45/76, Borehamrvoôd. BRE. 1976.
precincts. Part l. The smoke production of various
materials. Fire Research Station Fire Research Nc¡te 18 Hansell G O, Marshall N R and Morgan H P.
856. Borehamwood, BRE. i971. Private comnrunication. Fire Research Station
1988.
McGuire J H, Tamura G T and Wilson A G.
Considerations in the design of smoke conrrol t9 Morgan H P. The horizontal flow of buoyant gases
systems in tall buildings. Proceedings of ASHRAE toward an opening. Fire Safety tournal,1986, 11,
Symposium on Fire Hazards in Buildings, r93-200.
San Francisco. Januarv 1970.
20 Morgan H P and Marshall N R. Smoke hazards in
Hinkley P L. Work by the Fire Research Starion covered, multi-level shopping malls: an
,n the control of smoke in covered shopping experirnentally-based theory for smoke
centres. Building Research Establishntettr ('urrcttr pro<.1 uction. B u i I d i tt
1i Ras ea rc lt Es ta h ! is h m en t
Paper CP83t75. Borehamwood, BRE. 197.s. Cttrrent Pupcr CP48l7-5, Borehamrvood. BRE.
I 975.
London Transport Board. Second Reporr ot'the
Operational Research Team on the capacitv of 2t lllorgan H P and Manhall N R. Smoke control
footways. Researclt Report 95. rncasures in coverecl two-storey shopping mall
having balconies as pedestrian rvalkways. Building
r0 Heselden A J M and Hinkley P L. Srnoke rravel in Ilaseurclt Est¿tltlishnrcrtr Current Paper CPll 179.
shopping malls. Experiments co-operation with Borehamwood. BRE. 1979:
Clasgow Fire Brieade.- Part l. Fire Researclt
Station Fire Research Note 832. Borehamrvood. )', ùlorgan H P and Hansell G O. Atrium buildings:
BRE. 1970. calculating smoke flows in atria for smoke control
< design. Firc Sa.fety Jt¡urnal,1987, U (ry-¿[.
u Silcock A and Hinklcv P L. Fire at Wulfrun
shopping cen tre. Wolverhampton. Fire R cs e u rc. I t a Hansell G O, Morgan H P and Marshall N. Smoke
Station Fire Research Note 878. Borehamwood. l'lorv experimcnts in a model atrium: Part 2 plume
BRE. r971. en.trainment in the atrium. To be published.

32
'tl 'l'horrr¡¡r l' I l. ( )n thc ttpwirrcl nt()vcnlclìl ()l \tìì()ñe Chirrf crerl lnstitrrtiorr ol' lluil<ling
Se rviccr
rrrrl lirted sh<l¡r¡rin{ nrall prolrlems. I:irc,Stt.fett,
re f:ngincers. CllJ.\l'- (ìuitlc. Vrlu¡ne C, tgl.iil
.lourttttl . l9ii7. l2 (3) l9l-203.
ùtarshull N R and Hcscldcn A J ùI- Smoke control
t< ,\Iorgan I{ Crlnrme nl.s on 'A notc ()n stnoke
P. in largc strlres opcrrirìg ()tìt() c¡lclosed shoppin_rl
plumes lrorn I'ires in multi-level shoppirrg ¡ntlls'. rnalls. Firc Survevor, l9¡i6. l5 ( I ) lS-22.
Fira Su.f?tv Jr¡untul, 1987, U (l) 83-.34.
Morgan H P and Savagc N P. A study of a larsc
Larv l\{. A notc on smoke plumes from fires in fire in a covered shoppine complex: St Johnls
nrulti-lcvel shopping malls. Fi¡e Sufary Jounrul, Centre 1977. Builditr.g Rcseurch Estahlishment
t9ri6, l0 197-202. Currettt Papar CPlÌlf.ì0, Borehamwood, BRE l9tj0.

Morgan H P and Marshall N R. The deprh of void- HansellG O and Morgan H P. Fire sizes in horel
edge screens in shopping malls. Fire Engineers
Journal, I 989, 48 (152\ 7-9.
bedrooms
- implicarions for smoke control
design. Fire Sufetv Journul,lgS-5.8 (3) 177-1S6.

Hansell G O. Fire Engineers Journal,l989, 4{t Morgan H P and Hansell G O. Fire sizes and
(ls2) 9. sprinkler effectiveness in offices implications for
smoke control desi_sn. Fire Safety-Journal,l985, g
Penwarden.4, D..Acceptable wind speeds in town. (3) r87-r98.
Building Research Estabtishmeil Current paDer
CPI/7+. Garsron. BRE. 1974. Nlarshall N R, Feng S Q und r\torgan H p. Thc
infìuence of a perforated.l'alse ceiling on the
Spratt D and Heselden A J M. Efficient exrracrion performance of snroke ventilation systems. Frre
of smoke from a thin laver under a ceiling. Fire Safe* Journul.l9S.5. I ß) 227-237.
Rcsearclt Smtir¡¡t Fire Reseurch Note ltòtJl.
Borehamçood. B RE. 1974. Butcher E G, Langdon-Thomas G J and
Bedford G K. Fire anti cár park buildin gs. Firc
3l Heselden .4. J M. Private communication. Fire Researclt Staùott Fire itlotc 10. Borehamwood_
Research Station. 1976, BRE. l9ó8.

-F

JJ
 Liaison Panel
The preparation of the final report was undertaken with rhe assistance of a Liaison Panel which
consisted of the following members, representing both industry and Covernment interesrs:

Department of the Environment, Construction Directorate

Home Office, Fire lnspecrorate
Scottish Development Department
Colt International Limired
Gradwood Limited
Nuaire Lirnited

--F

34 Frínråd ¡f, rhè UK tor HMgO. Ðd-8276047, lO¡9O, Cio. 38938

 Also available from BRE o
,^-'
Smoke control in shopping malls The interaction of sprinklers and roof venting in T

3RE Video UK 2888, t30.43 + €4.57 VAT industrial buildings: the current knowledge
lovered shopping complexes have become increasingly Ref 8R57 €4.0O ?RE Reoort 1984 I
:ommon. Smoke control systems are required to ensure An examination of the basic features of the operalion of
ceople can escape if a fire causes smoke and loxic gases sprinkler systems, vent¡ng systems and their combination. ='
-(J
to enter the malls. This video shows model studies dt the The Report describes recent work in detail, and gives o-
Fire Research Station and illustrates wavs in which smoke results of surveys by several national fire protection u1
control can be achieved. (7 minutes) associations and by UK fire brigades. o
u1
A simplified approach to smoke ventilation Selection of sprinklers for high rack storage ín ¿
calculations warehouses o
BRE lnformation Paper lP|9/85 t1.75 BRE lnformation Paper lP5/88 t1.75
;ç
Outlines the calculations for either fan capacities or vent
o
Demonstrates the likely benefits of a 'fast-response'
areas for single-storey, large undivided-volume buildings solder-link sorinkler head for use ¡n a warehouse. A 'fast- o
where the fire is directly below the smoke reservoir. lt is response' solder-link sprinkler has been shown to provide
not ¡ntended for use with more complicated building effective in-rack protection for goods stored in high racks =.
geometries, such as shopping malls. Of particular use to uo to about lOm. The information will be useful to
building control officers, fire officers and others who have manufacturers, those concerned with high rack storage, o
to check smoke ventilation proposals or give approvals. insurers and the fire authorities.

Smoke control in.buildings: design principles Guidelines for the construct¡on of fire-resisting o
?RE Digest 260 t2.00 structural elements g
lines the design principles for systems which will Ref 8R128 tlO.OO ?RE Report 1988 o
p...rvide safe escape routes f rom buildings. U'
This Report establishes guidelines for the f ire resistance o
of elements of structure. lt includes tables of notional
Fire spread in buildíngs oeriods of fire resistance based on current test data; it
BRE Information Paoer lP21/84 t1.75 also makes revisions to tables and text concerning o
This Paper is concerned with the spread of fire in concrete, masonry and timber included in the eailier 1982
occupied buildings and its implications for safety of life. lt edition. E
makes suggestions on how problems are avoided. (o
ASKFRS o
Smoke spread within buildings Part 1 The scientific content o
?RE lnformation Paper IP22/84 t1.75 Part 2 A programmer's guide
Complementary to 'F¡re spread in buildings', lP21l84, this BRE Computer package AP46 t'lOO.OO + f 15.00 VAT o
Paper is concerned with smoke spread to areas beyond An interactive computer program comprising a manual Ø
the source, its implications for life safety and for property and 5'l. " disk, to help fire engineers quantify various fire
damage. scenarios us¡ng common algebraic expressions and a
materials database. Part 1 provides ten types of
calculation - Part 2 describes ASKFRS. The program is
Fire safety in buildings written in FORTRAN-77 and can be run on various VAX
Ref 8R96 î1O.OO 7RE Report 1986 comouters and IBM PCs.
The author's view of how Building Regulations and fire
protection measures relate to actual fires. The Report
makes a number of suggestions for the improvement of Experimental programme to ¡nvest¡gate informative
. 'tect¡on measures and its leqislation. fire warning character¡stics for motivating fast
evacuation
Aspects of fire precautions in buildings Ref 8R172 f25.OO BRE Report 1990
Gives details of an experimental programme to evaluate
Ref 8R137 t9.00 ]RE Reporr 1988
This book introduces students and others to the oass¡ve the relative effectiveness of alarm bells, text messages,
computef generated voice and graphics displays to
aspects of fire precautions - the measures taken and
protection provided in a building, for example, to minimise mot¡vate escape. The quantitative values of each, in
promoting a decision to escape, are given togelher with
the risk to occupants, contenls and structure from an
the sub¡ects' message assimilation times.
outbreak of fire. lt describes the nature of fire. structural
fire protection techniques and means of escape.
Psychological aspects of informative fire warning
systems
Ref 8R127 t5.0O BRE Report 1988
A summary of research using simulations of informative
fire warning systems in a number of buildings. and how
Please send order to: the information they carried, and training procedures, can
_ most effectively promote rapid and safe evacuation in the
-'-^
BRE Eookshop event of fire.
Buìlding Research Establishment
Garston, Watford, WD2 7JR Studies of human behaviour in fire: empirical results
and their implications for education and design
Cheques payable to DOE/BRE
Ref BR61 î8.5O BRE Repon 1985
Minimum order value î3
A model of human behaviour in fire is presented. The
implications of the model and the findings of the various
BR 186 studies are d¡scussed in relation to education, training,
publicity, design and legislation. Essential read¡ng for all
@'
rsgN 0 85125 462 4 those involved in decision making in these fields. m