Performance of smoke reservoirs protected by mechanical ... · PDF fileNo guidance is given in the BCA with respect to the analysis method or ... Figure 1ypical system design incorporating

eColi BR i u M • AuguSt 20 0 9 30

F O R U M

1. intRoduCtion1.1 Mechanical Smoke exhaustThe Building Code of Australia (BCA) [1] generally requires large spaces in buildings to be protected by a mechanical smoke exhaust system designed in accordance with the requirements set out in Specification E2.2b of the BCA. Essentially, it requires the spaces to be divided at the ceiling level into smoke reservoirs to contain the smoke layer; and exhaust fan(s) installed to remove the smoke from each reservoir (see Figure 1).

Even though this is a Deemed-to-Satisfy solution, the BCA sets out both prescriptive and performance specifications that the system must be designed to satisfy.

1.2 Prescriptive SpecificationsFor the reservoir size, the BCA prescribes the following limits:

• maximumhorizontalarea, Amax = 2000 m2

• minimumreservoirdepth, dmin = 0.5m

For the design of the mechanical smoke exhaust, it prescribes:

• heatreleaserate,Q&, for a given building classification and sprinkler protection

• minimumsmokelayerheight, hmin = 2m

• smokeexhaustrate,V&, for a given h and Q& (see Figure 3)

1.3 Performance SpecificationsIn addition to the prescriptive specifications, the BCA also specifies the performance the smoke exhaust system must be designed to achieve. The key one is that the smoke layer must be contained within the smoke reservoir. Other performance specified relate to plug-holing, make-up air and other aspects of the system.

1.4 design ProcessNo guidance is given in the BCA with respect to the analysis method or design process to achieve the specified requirements. Given that V&, h and d

are interdependent, an iterative design process may be required.

In a typical case where the ceiling height H is known or fixed, the designer must therefore provide sufficient d or V& to ensure the smoke layer can be contained within the reservoir (i.e. h ≥ (H – d)). This may involve an iterative process as shown in Appendix A of this paper.

It is unclear what process has been adopted in the design of many of the existing buildings. Anecdotal experience from Computational Fluid Dynamics (CFD) analyses revealed that the performance of many existing smoke exhaust systems falls significantly short of that specified by the BCA. In these instances, it appears that only the prescribed limits on reservoir area and depth (i.e. Amax and dmin) have been incorporated in their designs. The specified performance for smoke layer containment has either been ignored or not understood by their designers.

This raises the following fundamental questions with respect to the limits set by the BCA regarding the smoke reservoirs in the above situations:

Performance of smoke reservoirs protected by mechanical exhaust

dr Weng Poh, Associate director, Principal fire engineer, umow lai Pty ltd

Maisam Mirbagheri, fire Safety Consultant, umow lai Pty ltd

ABStRACtThe Building Code of Australia (BCA) requires smoke exhaust systems to be designed such that the smoke layer is contained within the smoke reservoirs.

Parametric studies using Computational Fluid Dynamics (CFD) analyses reveal that, with the BCA prescribed smoke exhaust rates, the smoke layer cannot be contained within reservoirs having a BCA prescribed minimum depth of 0.5m. For smoke containment, the necessary depths are significantly larger. Further studies also show that, to contain the smoke layer within a 0.5 m deep reservoir, the smoke exhaust rate will need to be significantly higher than that prescribed by the BCA and is unlikely to be practical.

An iterative procedure will likely be required to determine the appropriate combination of smoke reservoir depth and smoke exhaust rate to achieve the BCA specified performance. The results presented in this paper will provide useful guidance for the choice of initial trial values in the design process.

Control humidity with MuntersSupplying air at a 7°C dewpoint dries more than just the air, it dries the

building. To cool your building, you supply air colder than the desired

space condition. Likewise, to control humidity you need to supply air

drier than the desired humidity level. Munters’ equipment is designed

to condition outside air to a 7°C dewpoint without overcooling it. This

provides the perfect environment for your facility.

It's not the heat, it's the humidity.

www.munters.com.auSydney • (02) 8843 1588Melbourne • (03) 8795 4544Brisbane • (07) 3826 0461Perth • 1800 008 379Auckland • (09) 524 0543

Improve IAQ

Help Prevent Mould Growth

Reduce Operating Costs

Affordable Comfort

CO2 Emission Reduction

2

5

8

11

15

17

20

10 15 20 25 30 35 40

Dry Bulb Temperature °C

0 5

Hum

idity

Rat

io g

rain

s of

Moi

stur

e pe

r g/k

g of

Dry

AirOUTSIDE AIR

34°C, 14.5 g/kg

MUNTERS PATH

TRADITIONAL REHEAT PATH

SUPPLY AIR23.5°C, 6.2 g/kg

HOSPITALS SCHOOLS NATATORIUMS RESTAURANTS

eColi BR i u M • AuguSt 20 0 9 32

F O R U M

• GivenaBCAprescribedsmokeexhaustrate V&, can the smoke layer be contained within the smoke reservoir of the limiting size of Amax and dmin?

• Ifnot,whatistheleastdvaluerequired for containing the smoke layer within the reservoir?

• Oralternatively,whatistheleastsmoke exhaust rate necessary for containing the smoke layer within the smoke reservoir of the limiting size of Amax and dmin?

In order to provide some answers to the above questions, three series of parametric studies were carried out using CFD analyses to examine the smoke filling of smoke reservoirs protected by smoke exhaust systems. These are described in subsequent sections of this paper.

2. PARAMetRiC StudieS2.1 fdS and SmokeviewThe parametric studies were carried by simulating the smoke filling of smoke reservoirs using a CFD program, Fire Dynamic Simulator (FDS) [2] and Smokeview [3].

FDS is a computer model of fire-driven fluid flow and specifically developed for high-level fire engineering analyses. The software solves numerically a form of the Navier-Stokes equations appropriate for low-speed, thermally-driven flow with an emphasis on smoke and heat transport from fires.

Smokeview is a scientific visualisation program that was developed to display the results of an FDS model computation. It allows the viewing of FDS results in three dimensional snapshots or animations.

Since their development, FDS and Smokeview have increasingly being used in fire engineering design and research. Many publications can be found in the literatures on validation of various aspects of the softwares. However, there appears to be a paucity of studies on the comparison of the analysis results with test measurements involving smoke exhaust in large building spaces. The only notable one appears to be a study reported in [4], where smoke tests were carried out in an atrium space protected by a smoke exhaust system. In this study, the smoke layer interface height and air temperatures inside the atrium were simulated using FDS and compared with the data measured during the tests.

The study shows that FDS output compares well with the experimental results and demonstrates that FDS is capable of closely simulating smoke behaviour involving smoke exhaust in large building spaces.

2.2 ideal Smoke ReservoirFor the purpose of the parametric studies, an ideal situation of a single smoke reservoir with an area of Amax was examined using FDS. A BCA prescribed heat release rate (Q&) of 5 MW was arbitrarily chosen as the fire size for the purpose of the studies. The fire was assumed to be located centrally on the floor within the smoke reservoir (see Figure 2).

For the sake of simplicity, the effects of plug holing were eliminated by modelling the ideal situation where the smoke was extracted uniformly over the entire surface of the smoke reservoir at the ceiling level. Any adverse effects of make-

Figure 2 A typical FDS model – Smokeview Visualisation

Height to undersideof smoke layer, h

Smoke ReserviorDepth, d

Prescribed SmokeExhaust Rate, V

Smoke Layer

Prescribed heatrelease rate, Q

Enclosureheight, H

Imperforate ceiling or roof

Figure 1 Typical system design incorporating mechanical smoke exhaust system

Reservoir area = Amax Smoke reservoir depth = dSmoke exhaust rate = V&

Q & = 5 MW

33Augu St 20 0 9 • eColi B R i u M

F O R U M

up air were also eliminated by placing the smoke reservoir within an infinite space without any bounding walls.

It is important to note that the situation examined represents the best case scenario where all aspects were assumed to be the most advantageous for extraction of smoke from the reservoir.

Other model parameters and assumptions made in the analyses are summarised below:

• A=40x50m(i.e.=Amax of 2000 m2)

• hrangeexamined=2to15m

• Totalsimulationtimet = 1800 s

• = 5 MW and remains constant over the simulation time period

Figure 3 BCA Spec E2.2b, Figure 2.1 Smoke Exhaust Rate [1]

350340330320310300290

280

270

260

250

240

230

220

210

200

190

180

170

160

150

140

130

120

110

100

90

80

70

60

50

40

30

20

10

02 4 6 8 10 12 15 20 25 30

Height to underside of smoke layer (metres)

Smok

e ex

hau

st fo

r eac

h s

mok

e re

serv

oir (

m3 /s

)

Classification Unsprinklered Sprinklered

Class 2,3 or 5 5 MW 1.5 MW

Class 6 10 MW 5 MW

Class 7 or 8 15 MW 5 MW

Class 9

generally 5 MW 1.5 MW

exhibition halls 10 MW 5 MW

theatres, stages

and public halls

complying with

Part H (note 1) 10 MW 5 MW

Note 1: If the smoke reservoir above the stage is smoke separated from

the audience area, the fire load specified applies to the stage area

only and fire load for audience area is as per Class 9 generally.

1.5 MW

5 MW

10 MW

15 MW

SMOKE EXHAUST RATE

Figure 4 Assumed h = 12m, d = 0.5m, t = 1800s

Figure 5 Assumed h = 12m, d = 0.5m, t = 1800s

• Combustionreaction=timber(Douglas Fir)

• Effectsofsprinklersonthesmokebehaviour are ignored

• Smokeexhaustfanstartsupinstantaneously and operates at its full capacity V& over the entire simulation time period

• Modelgridsize=0.5m

For the purpose of determining the level of containment of the smoke layer, devices for measuring the smoke layer height were incorporated in the model. The smoke layer was considered to be contained within the smoke reservoir if h ≥ (H – d) over the total simulation time. This was further verified by visually examining the analysis results using Smokeview.

2.3 SeRieS 1: testing BCA prescribed minimum reservoir depth, dmin

The first series of analyses was carried out by setting d = 0.5 m for each assumed h values with the corresponding smoke exhaust rates as prescribed by the BCA (see Figure 3).

This was aimed at examining whether a smoke reservoir with the minimum dmin = 0.5 m could fulfil the requirement of smoke layer containment.

The results of the analyses in this series show that, in each case, the smoke layer falls below the smoke reservoir and the smoke spreads out of the reservoir.

Figures 4 and 5 show the Smokeview visualisations of two examples with a relatively small and a relatively large assumed h values, respectively.

dmin

dmin

Amax

Amax

Smoke spreads outside reservoir

Smoke spreads outside reservoir

H = 2.5m

H = 12.5

eColi BR i u M • AuguSt 20 0 9 34

F O R U M

In all the cases examined, the reservoirs could not contain the smoke layer despite the fact that the smoke reservoirs and smoke exhaust were arranged in the most advantageous situation as discussed earlier. The performance would likely to be worse in less ideal situations. This leads to a firm conclusion that, with the BCA prescribed smoke exhaust rates, the prescribed minimum reservoir depth of 0.5 m is insufficient to contain the smoke layer within the reservoir for a constant 5 MW fire and for the h values considered.

2.5 SeRieS 2: increasing reservoir depth, d

Having completed Series 1, a second series of analyses was carried out to determine the least smoke reservoir depths for smoke containment for various h values and the corresponding BCA prescribed smoke exhaust rates.

This was achieved by increasing d in steps of 0.5m for each h value until the smoke layer was contained within the smoke reservoir (i.e. h ≥ (H – d)).

9

8

7

6

5

4

3

2

1

0

0 2 4 6 8 10 12 14 16 18 20 22

Height to underside of smoke layer, h (ma)

Smok

e re

serv

oir d

epth

, d (m

a)

Depth necessary forsmoke containment

(with BCA prescribed V&)

BCA prescribed

Figure 6 Least Smoke Reservoir Depth Necessary for Containment of Smoke Layer.

Figure 7 Assumed h = 2m, d = 3.5m, t = 1800s

d increased to 3.5m (smoke contained)

H = 5.5m

Figure 8 Assumed h = 12m, d = 5.5m, t = 1800s

d increased to 5.5m (smoke contained)

H = 17.5m

Figure 9 Assumed h = 4m, d = 0.5m, V& = m3/s, t = 1800s

H = 4.5m

dmin V& = 60m3/s

Smoke spreads outside reservoir

35Augu St 20 0 9 • eColi B R i u M

F O R U M

The results of the analysis of this series show that the d values required to contain the smoke layer within the reservoirs are significantly larger then the BCA prescribed minimum of 0.5m. The findings are summarised in Figure 6.

Figures 7 and 8 show the Smokeview visualisations of two examples with a relatively small and a relatively large h values, respectively, when the smoke layer is contained within the reservoirs.

It is noted that in this analysis series, by keeping h constant, H increases with increasing d. This differs from the design process discussed earlier where H is fixed.

Again, it is important to note that the situation analysed is an ideal case

where the reservoir, fire and smoke exhaust are arranged in their most advantageous conditions. Hence, the resulting d values in Figure 6 apply only to these best case scenarios. The reservoir depths for smoke containment could be larger in less ideal situations.

2.6 SeRieS 3: increasing smoke exhaust rate,

A third series of analysis was carried out as a limited study to determine the smoke exhaust rate V& necessary to contain the smoke layer for the BCA prescribed dmin of 0.5 m. For the purpose of this study, the case where h = 4 m was arbitrarily chosen. The BCA prescribed smoke

exhaust rate in this case is approximately 30m3/s (see Figure 3).

The study was carried by increasing in steps of 30m3/s until the smoke layer was contained within the reservoir. Further analyses were then carried out to refine the limiting value to an accuracy of 5m3/s.

The results of the analyses show that the smoke exhaust rate required for containing the smoke layer is approximately 155m3/s.

Figures 9 and 10 show the Smokeview visualisations of two examples with different V& values in the series.

It can be seen from this case study that, with a reservoir depth of 0.5m, the smoke exhaust rate required to contain the smoke within the reservoir is significantly higher than the corresponding value prescribed by the BCA. The resulting high exhaust rate may also introduce significant issues with make-up air. The practicality of increasing the smoke exhaust rate alone in containing the smoke layer is therefore questionable.

www.bigfootsystems.com.au

the complete solution

Pro Pipe Supplies Pty Ltd, 13-15 Main Street, Beverly, SA5009t: +61 8 8268 8633 [email protected]

Figure 10 h = 4m, d = 0.5m, V& = 155 m3/s, t = 1800s

H = 4.5m

dmin V& = 155m3/s Smoke spreads

within reservoir

eColi BR i u M • AuguSt 20 0 9 36

F O R U M

3. CoMMentS on BCA PReSCRiBed SolutionS

Review of literature indicates that the BCA prescribed exhaust rates may find its root to a simple smoke production model devised empirically by Thomas in 1963 [5]. The smoke exhaust rates were derived simply by equating the exhaust rates to the respective smoke production rates. This gives smoke exhaust rates in the same format as that prescribed by the BCA (see Figure 3), i.e. V& = k1Q& + k2h1.5, where and are constants. An account of the derivation can be found in [6, 7].

If the Thomas model could accurately predict the smoke production rates, then the prescribed smoke exhaust would maintain the smoke layer for each of its corresponding h value, and the smoke layer would be contained within the smoke reservoir. From this viewpoint, the choice of a minimum reservoir depth is somewhat arbitrary, since the smoke layer would be contained within a reservoir with any assumed depth.

The results of the parametric study using FDS, however, show different results. Not only the smoke layer could not be contained within smoke reservoirs with the prescribed minimum depth, the prescribed smoke exhaust rates also could not contain the smoke layer in deeper reservoirs. This calls into question the validity of the BCA prescribed exhaust rates, which are based on a simple model devised almost half a century ago.

FDS, on the other hand, is a modern CFD program that is scientifically based and numerically solves the fluid motion Navier-Stokes equations. Although, currently, there is a paucity of data in the published literatures on the comparison of FDS output with the experimental results, the limited comparison in [4] shows that it is capable of closely simulating smoke behaviour involving smoke exhaust in large building spaces. With its increasing use throughout the world in the fire engineering field, it is to be expected that there will be more testings and comparisons with FDS that could further validate the accuracy of the software.

The findings of this study point to the fact that the BCA prescribed solutions may need to be revised in light of the more advanced methods of analysis the engineering fraternity now has at its disposal. A new set of prescribed solutions that satisfy the desired performance can be derived. However, this will require further study, development and agreement by the appropriate authorities.

Meantime, engineers will still need to design smoke exhaust systems to satisfy the current prescribed requirements of the BCA, perhaps using FDS or other modern analysis methods. An iterative procedure will likely be required to determine the appropriate combination of smoke reservoir depth and smoke exhaust rate. In this instance, the results presented in this paper will provide useful guidance for the choice of initial trial values in the design process.

Another alternative to the above approach is for the designers to seek a performance-based fire engineering solution whereby the fire safety of the building is assessed holistically rather than examining the smoke exhaust system in isolation. This may mean that containment of

smoke layer in each smoke reservoir may not necessary form the governing criterion for the design, if it can be demonstrated that a high level of fire safety of the building occupants can be achieved using alternative design criteria.

4. ConCluSionSParametric studies using CFD analyses have been carried out to examine the smoke filling of smoke reservoirs. The studies reveal that with smoke exhaust rates as prescribed by the BCA, smoke reservoirs having a depth of 0.5 m are insufficient to contain the smoke layer within the reservoirs. For smoke layer containment, the necessary reservoir depths are significantly larger than the BCA prescribed minimum of 0.5m. These were investigated in the parametric studies and the necessary depths for smoke containment in the ideal, best case situations are given in the paper.

Further studies were also carried out to investigate the smoke exhaust for a smoke reservoir with a depth of 0.5 m. The results show that the smoke exhaust rate required for containment of smoke within the reservoir is significantly higher than that prescribed by the BCA. The high exhaust rate is unlikely to be practical and may also introduce significant issues with make-up air, hence the viability of adopting the minimum reservoir depth and increasing the smoke exhaust rate for containing the smoke layer is questionable.

In order to achieve the performance specified by the BCA for smoke containment, it is necessary to use a reservoir depth greater than the BCA prescribed minimum. An iterative procedure will likely be required for determining the necessary reservoir depth d and/or smoke exhaust rate to satisfy the BCA specified performance. The results presented in this paper will provide useful guidance for the choice of initial trial values in the design process. They may also be useful for use in gauging the performance of smoke reservoirs in general.

The findings of this study also indicate that the BCA prescribed solutions may need to be revised in light of the availability of more advanced methods of analysis. A new set of prescribed solutions that satisfy the desired performance can be derived. This will require further study, development and agreement by the appropriate authorities.

5. RefeRenCeS[1] “Building Code of Australia. Volume 1 – Class 2 to Class 9

Buildings”, Australian Building Codes Board, 2008.

[2] McGrattan, K, B., Forney, G. P., Floyd, J.F, Hostikka, D., Prasad, K. “Fire Dynamic Simulator (Version 5) – User’s Guide”, National Institute of Standards and Technology, NIST 6784, 2008

[3] Fourney, G.P, “User’s Guide for Smokeview Version 5 – A Tool For Visualizing Fire Dynamics Simulator Data,” National Institute of Standards and Technology. NIST Special Publication 1017. Gaithersburg, MD, August 2007.

[4] Chow, W. K. et al, “Numerical Studies on Atrium Smoke Movement and Control with Validation by Field Tests,” Building and Environment, Volume 44, Issue 6, June 2009, pp 1150-1155.

37Augu St 20 0 9 • eColi B R i u M

F O R U M

[5] Thomas, P. H, at al, “Investigations into the Flow of Hot Gases in Roof Venting”, Fire Research Technical Paper No. 7, HMSO, London, 1963.

[6] Wild, J.A., “Fans for Fire Smoke Venting”, Woods of Colchester Limited, November 1990.

[7] Courtier, G.A.C, and Wild, J.A., “The Development of Axial Flow Fans for the Venting of Hot Fire Smoke”, Fire Technology, August 1991, pp 250 – 265.

6. noMenClAtuReA = Horizontal area of smoke reservoir (m)

Amax = BCA maximum allowable area for smoke reservoir = 2000 m2

d = Depth of smoke reservoir (m)

dmin = BCA minimum smoke reservoir depth = 0.5m

h = Height to underside of smoke layer (m)

hmin = BCA minimum allowed smoke layer height = 2m

H = Height of enclosure containing the smoke reservoir (m)

Q& = Heat release rate of fire (MW)

V& = Smoke exhaust rate for the smoke reservoir (m3/s)

t = Elapse time from start of fire (s)

APPendiX A: determining d and v& for known h

h

A

QH

d

V

eColi BR i u M • AuguSt 20 0 9 38

F O R U M

APPendiX A: determining d and v& for known h

try new d (> dmin)h = (H - d )

ish > 2.0m ?

establishbuilding classification,

Q, H and A (<Amax )

determine V(BCA Spec 2.2b Fig 2.1)

determine hactual

(smoke flow analysis)

[optional] optimised = (H – h actual)

ish actual > h ?

design smoke exhaust system, including checks against

plug holing etc

design complete

wish to increased? increase V

start

are make up airrequirements

satisfied?

yes

no

yes

no

yes

no

yes

no

ABout tHe AutHoRSDr Weng Poh is an Associate Director and the head of the fire engineering group at umow Lai Pty Ltd. Throughout his career, Weng has completed numerous fire engineering projects. He has been closely involved in the development of methodology for fire safety designs of buildings in Australia and has published extensively in international journals, conferences and technical reports. Weng also gives lectures in postgraduate courses, aimed at training building practitioners in fire engineering design.

Maisam Mirbagheri is a Fire Safety Consultant at umow Lai Pty Ltd. In the past two years, Maisam has been involved in the fire engineering design of commercial, educational, healthcare and retail buildings. He has recently completed his graduate Diploma in Building Fire Safety and Risk Engineering and has been awarded the Best Student in the 2008 graduate Diploma Class (CESARE) Award. He also has a vast and practical experience in the design of fire/hydraulic/

mechanical pump and tank package systems for building services.

Short courses to degrees

You can elect to participate in a number of individual units of study

that form part of the graduate program in Building Services, or enrol in

a Graduate Certificate, Diploma or Master of Design Science (Building

Services). You may then want to combine the Building Design stream

with Sustainable Design to extend your knowledge even further.

Flexible delivery designed for professionals

Most units are taught in intensive blocks of full-day sessions, so you are

not having to make a weekly commitment for a whole semester. Some

are taught on weekdays and weekends, so the time you take off work is

kept to a minimum.

What’s more, you decide whether to just attend the sessions (or even

just the introductory session) or whether you want to enrol as a student

and complete assessment tasks. Completing the assessment tasks

successfully can see you upgrade into one of the Graduate Programs.

Programs taught by the experts

Our units are taught by some of the leading academics and

practitioners in the business, ensuring you learn both the key theoretical

foundations and the practical application of them in real-world settings.

Like all our graduate programs, it is guided by an industry advisory

panel ensuring it teaches you not only what you need for today but into

the future as well.

Study with professionals

Most of the participants in units are like you. Professionals keen to

cement their knowledge to secure their future career progression.

Although it is a graduate program you do not need a Bachelor’s

Degree in order to enrol in individual units or a Graduate Certificate,

with many students bringing knowledge into the class from extensive

trade backgrounds.

Choice of units to studyUnits to choose from include:

Air Conditioning Design• 5-day intensive, August 5, 6, 7, 12, 13 (Code: DESC9001)

Energy Management in Building• 4-day intensive, August 17, 18 & October 12, 13 (Code: DESC9111)

Building Energy Analysis• 5-day intensive, August 26 & 27, September 2, 3, 4 (Code: DESC9015)

Energy Code Compliance• 6-day intensive, September 16, 17, 26, 27 & October 10, 11 (Code: DESC9192)

Services Control Systems•

5-day intensive, October 7, 8, 14, 15, 16 (Code: DESC9170)

Professional Education options for refrigeration, air conditioning and heating professionals at The University of Sydney The Faculty of Architecture, Design & Planning at The University of Sydney offers a range of programs designed to meet the

needs of busy professionals in the heating, air conditioning and related building service and facilities managment industries.

Contact Us

For more information, contact Sue Lalor in the

Professional Education Office at Faculty of Architecture,

Design & Planning on 02 9114 0941 or at slalor@arch.

usyd.edu.au or visit www.arch.usyd.edu.au/cpd

The University of SydneyFaculty of Architecture, Design and Planning