Application of the Zone Model FIRST on the Development of Smoke Layer and Evaluation of Smoke Extraction Design for Atria in Hong Kong W. K. CHOW AND W. K. WONG Department of Building Services Engineering Hong Kong Polytechnic Hong Kong (Received August, 1992) (Revised February, 1993)

ABSTRACT: Whether

a smoke extraction system has to be installed for an atrium building in Hong Kong is determined by its volume. This article reports an evaluation of this regulation using the zone model FIRST developed at the Building and Fire Research Laboratory, NIST, U.S.A. A survey of the geometrical configurations of the local atria is made. The general shapes of the atria are classified into three types: 1, 2 and 3, with types 2 and 3 further divided into three sub-types: A, B and C. Smoke filling in those atria is simulated by the 3 to model FIRST with the volumes of the atrium space varied from 2500 m . It is illustrated that specifying only the volume of the atrium space 3 35,000 m is not good enough to determine whether a smoke extraction system has to be installed. The geometrical configuration is recommended to be included as well. A time constant is defined for the atrium with a certain design fire.

1. INTRODUCTION

ATRIUM BUILDINGS ARE recently very popular in Hong Kong. Although an atrium is not a new architectural feature, the fire safety in such space, particularly when it is located in a shopping mall, is still a matter of great concern to architects, engineers and fire officers. Since there is not much research work [1] reported on atrium fires, design guides 329

330 are set up without strong scientific background. Works reported include the full-scale atria experiments on smoke extraction (e.g., Tanaka and Yamana [2], Yamana and Tanaka [3]), scaled models for studying the smoke movement and smoke control (e.g., Tsujimoto [4]), semiempirical studies (e.g., Thomas [5], Morgan and Hansell [6], Hinkley [7], Hansell and Morgan [8,9], Milke [10,11]) and finally, the numerical simulations on the fire environment using both zone and field modelling techniques (e.g., Tanaka and Nakamura [12], Chow [13,14], Waters [15], Morita et al. [16]). Results from those studies are not good enough for formulating regulations [17,18] and codes for designing safe atrium buildings. Further investigational work is urgently required. However, it is quite clear that spread of smoke in an atrium building is a serious problem, and hence being able to design an appropriate &dquo;smoke control&dquo; system is extremely important. Earlier studies by Chow and Wong [19] using a zone modelling technique supported this point already. Therefore, a smoke extraction system is an important active fire protection system in atrium buildings. The heat released by burning materials is unable to heat the large volume of air enclosed (e.g., 28,000 m3 ) to a flashover fire temperature (e.g., 600 ° C) within a short time interval. For a fire with thermal power of 5 MW, the time taken for flashover will be at least an hour (e.g., Chow [20]), if air is assumed to be stagnant. Obviously, the actual time required will be much longer since about 30% of the heat release is lost by radiation. In Hong Kong, whether a smoke extraction system has to be installed is determined by the volume of the atrium space [21-23]. The critical volumes are 7000 m3 for a basement, 7000 m3 for an atrium with fire local high density and 28,000 m3 for a normal atrium space. Evaluation of this guide with full-scale burning tests is good [3] but too expensive. An alternate way is to simulate the fire growth and smoke filling process using fire models. With input data on the building geometry and a suitable design fire, the hot gas temperature and smoke layer thickness in the atrium can be predicted. Well-validated zone models [24-26] are available in the literature and several of them can even be executed on a personal computer. Zone models are good for providing design data for fire engineering systems quickly since the computing time required is rather short. Although some training in using the models is required, they are suitable for use by the local building services industry since most of the firms are equipped with personal computers. Applying zone models to study the development of the smoke layer and to evaluate the design of a smoke extraction system in an atrium based on the space volume is the objective of this article, and the model FIRST (Mitler [27], Mitler and Rockett [28]) is chosen to be the fire simulation tool.

331

Section 2 of the article describes the types and configurations of atrium buildings in Hong Kong. From a survey, the atrium buildings are classified into three types, 1, 2 and 3, with types 2 and 3 further divided into three sub-types A, B and C. Fire simulations and smoke filling processes in those three types of atria using the model FIRST are carried out in Section 3. There, the hot gas temperature, smoke layer thickness and time required to reach the maximum smoke layer thickness are predicted. Section 4 is on the effect of smoke extraction and Section 5 is on a recommendation made to specify the design of atrium for smoke control purposes; then, a time constant proposed by Chow [29] is discussed and the results on the atrium spaces simulated by the model FIRST is compared. Section 6 is a conclusion. 2. ATRIUM BUILDINGS IN HONG KONG

In Hong Kong, there are many large atria, and they could be of quite different design in their shapes. For example, the one in the China Bank Building (Ross et al. [30] ) is large in its height, being more than 60 m; in the Hong Kong Bank Building the atrium [31] is large in its volume, being 11,000 m3; and in Shatin New Thwn Plaza the atrium [32] is large in its length, being 88 m. A survey on the geometrical shapes of the atrium spaces in Hong Kong had been carried out, and it was shown that they can be classified into three main types (e.g., Chow and Wong [33,34], Chow [35]). Such a classification is adopted in this article. 2.1 Atrium

Type 1: Cubic

The atrium space is of cubic shape and the design is commonly found Hong Kong. About 60% of the atrium spaces can be classified as this type. They are smaller in scale (i.e., usually of length less than 20 m) and most of them are integrated into the shopping centre. The dimension (length x width x height) can be approximated by L x L x L, where L is the length as shown in Figure 1. in

2.2 Atrium

Type 2: Flat

The atrium has a large transverse dimension compared with the height. This kind of atrium is constructed in large multi-level shopping malls and 25% can be classified as the &dquo;flat&dquo; type. It is classified into three sub-types as 2A, 2B and 2C with dimensions (length x width x height) specified as 2L x L x L, 3L x L x L, 4L x L x L respectively, as shown in Figure 1.

332

333 2.3 Atrium

Type 3: High

This is the atrium with a height to width (or length) ratio of more than two. They are usually found in prestigious office buildings such as the Hong Kong Bank Building [31] and luxurious hotels like the Royal Garden Hotel [36]. About 15% of local atria belong to this type. It is subdivided into types 3A, 3B and 3C with dimension (length x width x height) approximated by L x L x 2L, L x L x 3L and L x L x 4L, and is shown also in Figure 1. The local fire authority [21,22] is interested in the volume of the atrium space. The volumes to be considered are 7000 m3 in the basement ; 7000 m3 for an atrium with high fire load density, and 28,000 m3 otherwise. Whether this is reasonable and sufficient can be evaluated using the zone model FIRST. Simulations are performed on the above three types of atria with volumes varying up to 35,000 m3. Their fire environments and the effects of smoke extraction are also studied.

3. SIMULATION ON THE ATRIUM FIRE ENVIRONMENT

The fire environment due to a fire occurring at the atrium floor is simulated using the zone model FIRST. A 3 m diameter growing fire with heat release rate plotted in Figure 2 is located at the centre. Simulations are carried out in studying the fire environment for the three types of atria with dimensions shown in Table 1. Volumes of the three

Table 1. Dimensions of atria simulated.

334

335

Figure 2. Heat release

rate of the fire.

types of atrium space vary from 2500 m3 to 35,000 m3. A smoke curtain is assumed to be operated so that all higher level spaces open to the atrium are covered except for leaving four openings of 3 m height at the bottom as in Figure 1. Results for the maximum smoke temperature and smoke layer thickness are shown in Figures 3(a) and 3(b). The maximum temperature of the smoke layer is less than 400 K (123 ° C) for a type 1 atrium of volume 2500 m3. For atrium volumes of 28,000 m3, the temperature will not be hotter than 320 K (47 ° C) for all three types. Therefore, the smoke will be quite &dquo;cool&dquo; and the oxygen concentration will be high. The combustion process will be completed on the burning regions and the concentrations of carbon monoxide will be quite low. However, in view of Figure 3(b), smoke will cause a problem. For the same type of atria under this design fire, the smoke layer thickness increased as the volume of the atrium space increased. But the time required to fill the atrium to the maximum smoke layer thickness was longer. Compared to the others, the type 3 atria will have a very thick smoke layer filled. The smoke layer thickness is more than 50% of the atrium height even for a type 2C atrium with the smallest volume of 2500 m3. Smoke will cause problems, but from the above simulations, the atrium might not be wholly filled with smoke. The time required to fill the atrium to the maximum smoke layer thickness is important and is plotted in Figure 3(c) for the three types of atria at different volumes.

336

337

For a certain type of atrium, the larger the space volume, the longer is the time required to fill it up with smoke. From the simulations, the maximum smoke layer thickness will be different for different atria. A reference value is taken as the smoke layer thickness being equal to 80% of the atrium height [10,11]. The times tr required to fill 80% of the atrium with smoke for those cases are plotted in Figure 3(d) as well.

4. EFFECT OF SMOKE EXTRACTION

Without a zone model, design of the smoke extraction system is possible only when the smoke production rate is known. The rate of combustion products emitted (e.g., in kg s-1 ) from the burning materials is not too large but the upward moving fire plume will entrain air. The air entrainment rate will determine the actual smoke production rate. An empirical equation has been used [37,38] for estimating the rate of smoke production Mp (in kg s-1) in large fires through the clear height y (in m) and the perimeter p (in m) of the fire:

Taking only mass transfer into account, the smoke extraction inert (in kg s-1) rate for the fan can be estimated from the air density Q, the floor area Aand the height H of the atrium as:

However, the above expressions do not include the enclosure and layering effects and are good only for quick engineering estimation. The application of this plume equation is reported separately by Chow [29] and will not be repeated here. With the zone model FIRST, almost all the physical effects concerned in a building fire are included. Since the ventilation rate is specified to be 6 air changes per hour in the local regulation [23,24], simulations with ventilation rates of up to 10 air changes are performed. Results on the types 1, 2A, 2B, 2C, 3A, 3B and 3C atria are shown in Figures 4 to 10 respectively. For a type 1 atrium, the time required to fill up 80% of the atrium with smoke is much longer than would be the case for natural filling. When the extraction rate is increased to higher than 8 air changes per hour, smoke cannot fill the bigger atrium with 80% smoke. Similar results are found for the types 2A, 2B and 2C atria.

338

339

340

341

342

343

344

345 For a type 3A atria, an extraction rate of 6 air changes per hour would not change the values of t, significantly. But the value of this time is much longer when the atrium is bigger than 10,000 m3. Effect of changing the smoke extraction rates on reducing t~ is not obvious for either the type 3B or 3C atria. 5. THE TIME CONSTANT OF THE ATRIUM A time constant

T

has been defined for

an

atrium with

a

fire

by Chow

[29]:

~ is the geometrical aspect factor given in terms of the atrium floor area

Aand height

H:

For the fire simulations carried out in this article, the perimeter of the fire p is 9.425 m. Taking ~ to be 1.22 kg m-3, the time constant is

given by:

The time constant describes how fast the atrium will be filled up with smoke. The values of T for the atria simulated in this article are also shown in Table 1. Values of t, are plotted against T in Figure 11. Using Equations (1) and (2) with mass transfer only, a crude model has been proposed on relating tr with the time constant by Chow [29]:

This is plotted in Figure 11 and fits the type 3 atria well, but not so well for the type 2 ones. Since the model FIRST is well validated and can be used as a fire simulator, the data on t, and T are fitted by linear

regression:

346

However the plume entrainment in an atrium is different from plume entrainment in the open. It is fair to say that Equations (5) and (6) are in semi-quantitative agreement and perhaps the following equation can be used to relate the time required to fill 80% of the atrium with smoke and the time constants of the atria:

where

If a reference on t, is used and longer than 150 s, the time constant of the atrium is recommended to be lying between 121 s and 179 s as from Equation (7). Whether a smoke extraction system has to be installed will then be determined by this value of the time constant, rather than by the volume only. Also the full-scale experiment on smoke filling reported by Yamana and Tanaka [3] at the Building Research Institute is considered. The atrium is of length 30 m, width 24 m, height 26.3 m and burnt with a methanol fire of thermal power 1.3 MW. This can be classified as a type 1 atrium with ~ equal to 0.98. The time constant T is then 4.25 minutes and the experimental value of t, is 5.40 minutes. This is plotted in Figure 11 and quite a good agreement is obtained.

6. CONCLUSIONS

The fire environment in the three different types of atria are simulated using the zone model FIRST (Mitler [27], Mitler and Rockett [28] ). With it, the development of temperature and smoke layer thickness are predicted. The predicted hot gas temperature profile indicated that flashover is unlikely to occur if 600 ° C is taken to be the criterion for flashover [39], but smoke will develop quite fast to a thick layer during an atrium fire and installing a smoke control system seems necessary. The effect of the smoke extraction system on the development of the smoke layer is also simulated with ventilation rates up to 10 air changes per hour. It is recommended that a time constant is to be used for specifying the development of smoke in the atrium.

ACKNOWLEDGEMENT The authors wish to thank the

and Fire Research Laboraof the zone model FIRST.

Building

tory, NIST, U.S.A. for allowing the

use

347

REFERENCES

Proceedings of the seminar on "Atrium Engineering," Inst. of Mechanical Engineering, London, UK. 2. Tanaka, T. and T. Yamana. 1985. "Smoke Control in Large Scale Spaces-Part 1," Fire Science & Technology, 5:31-40. 3. Yamana, T. and T. Tanaka. 1985. "Smoke Control in Large Scale Spaces-Part 2," Fire Science & Technology, 5:41-54. 4. Tsujimoto M., T. Nagaoka and S. Uehara. 1990. "A Scaling Law of Smoke Movement in Atrium," private communication. 5. Thomas, P. H. 1987. "On the Upward Movement of Smoke and Related Shopping Mall Problems," Fire Safety Journal, 12:191-203. 6. Morgan, H. P. and G. O. Hansell. 1987. "Atrium Buildings: Calculating Smoke Flow in Atria for Smoke Control Design," Fire Safety Journal, 12:9-35. 7. Hinkley, P. L. 1988. "Comparison of an Established Method of Calculation of Smoke Filling of Large Scale Spaces with Recent Experiments," Fire Science & Technology, 8:1-8. 8. Hansell, G. O. and H. P. Morgan. 1990. "Smoke Control in an Atrium Building Using Depressurization-Part 1," Fire Science & Technology, 10:11-26. 1. 1990.

9. 10.

Hansell, G. O. and H. P. Morgan. 1990. "Smoke Control in an Atrium Building Using Depressurization-Part 2," Fire Science & Technology, 10:27-41. Milke, J. A. 1988. "Fire Hazard Assessment in Atriums," Roundtable on

Safety in Atriums-Are the Codes Meeting the Challenge? Washington, DC, December 15. 11. Milke, J. A. 1990. "Smoke Management for Covered Models and Atria," Fire Fire

Technology-NFPA (August):223-243. Tanaka, T. and K. Nakamura. 1989. "A Model for Predicting Smoke Transport in Buildings," Report No. 123 of the Building Research Institute, Ministry of Construction, Japan. 13. Chow, W. K. 1989. "Smoke Movement and Design of Smoke Control in Atrium Buildings," International Journal of Housing Science and Its Ap-

12.

plications, 14.

15.

16.

17. 18.

13:307-322. W. K. 1990. "Smoke Movement and Design of Smoke Control in Atrium Buildings-Part 2: Diagrams," International Journal of Housing Science and Its Applications, 14:147-159. Waters, R. A. 1989. "Stansted Terminal Building and Early Atrium Studies," J. Fire Protection Engineering, 1:63-76. Morita, M., Y. Yamauchi and A. Manmoto. 1991. "Numerical Simulation of Fire in the Temperature Stratified Atrium with a Mathematical Field Model," poster paper presented at the Third International Symposium on Fire Safety Science, Edinburgh, U.K. 1990. Building Code of Australia, Australian Uniform Building Regulations Co-ordinating Council. 1991. NFPA 92B. "Guide for Smoke Management Systems in Malls, Atria and Large Areas," Quincy, MA: National Fire Protection Association.

Chow,

348 19.

Chow, W K. and W. K. Wong. 1993. "On the Simulation of Atrium Fire Environment in Hong Kong Using Zone Models," Journal of Fire Sciences, 11(1):3-51.

20.

Chow, W K. 1991. "On Designing a Smoke-Extraction System for an Atrium Building," Building Journal, Hong Kong and China (May):

103-104. 21. 1988. "Smoke Extraction System," Fire Services Department Circular Letter No. 13/88, Hong Kong: Fire Services Department. 22. 1990. "Code of Practice for Minimum Fire Services

Installations," Fire Services Department, Hong Kong. 23. Chow, W. K. 1989. "FSD Circular Letter No. 13/88: A Comment," The Hong

Kong Engineer (Nov):19. Jones, W. W. 1983. "A Review of Compartment Fire Models," NBSIR 83-2684, Centre for Fire Research, National Bureau of Standards, MD. 25. Cooper, L. Y., J. A. Rockett, H. E. Mitler and D. W. Stroup. 1987. "A Pro24.

gram for the

Development of a Benchmark Compartment Fire Model Com-

puter Code," in Mathematical Modelling of Fires, J. R. Mehaffey, ed., ASTM STP 983, pp. 116-127. 26. Mitler, H. E. 1991. "Mathematical Modeling of Enclosure Fires," chapter 23 in Numerical Approaches to Combustion Modeling, E. S. Oran and J. P. Boris, eds., Vol. 135 in Progress in Astronautics and Aeronautics, Washington, DC, published by AIAA (American Institute of Aeronautics and

Astronautics).

Mitler, H. E. 1978. "The Physical Basis for the Harvard Computer Fire Code," Home Fire Project Technical Report No. 34, October 1978, Division of Applied Science, Harvard University. 28. Mitler, H. E. and J. A. Rockett. 1987. "User’s Guide to FIRST: A Comprehensive Single-Room Fire Model," NBSIR 87-3595, Gaithersburg, MD: Center for Fire Research, National Bureau of Standards. 29. Chow, W. K. "Smoke Development and Engineering Aspects of Smoke Extraction Systems for Atria in Hong Kong," Fire and Materials, accepted to 27.

publish. 30.

31.

Ross, D. E., W. S. Lewis, A. LoPinto, S. P. W. Wong, A. Wong and H. C. Mak. 1989. "Case Study: Bank of China Building," The Journal of Hong Kong Institution of Engineers (May):35-42. 1985. "Hong Kong Bank Headquarter Building-An Eye-Catching ’SpaceScraper’ Explores the Hong Kong Skyline," Building Journal Hong Kong

(August):62-67. 32. 1984. "New Town Plaza-Texas-Size in 33.

34.

Shatin," Building Journal Hong Kong (December):78-82. Chow, W. K. and W. K. Wong. 1990. "Fire Aspects of Atrium Buildings in Hong Kong," Proceed. 5th International Fire Conference: Inteflam’ 90, University of Kent, Canterbury, UK, Sept. 1990, pp. 361-365. Chow, W. K. and W K. Wong. 1991. "A Study of the Fire Aspect of Atria in Hong Kong," Proceed. 3rd International Symposium on Fire Safety Science, University of Edinburgh, Scotland, UK, July 1991, pp. 335-344.

349

Chow, W. K. "Simulation of the Atrium Fire Environment in Hong Kong Using FIRST," ASHRAE Transactions-accepted to publish. 36. 1981. "The Royal Garden: Magnificent Landscaped Atrium," Building Journal Hong Kong, (December):55-65. 37. Thomas, P. H., P. L. Hinkley, C. R. Theobald and D. L. Simms. 1963. "Investigation into the Flow of Hot Gases in Roof Venting," Fire Research Technical Paper No. 7, HMSO, London. 38. Hinkley, P. L. 1981. "Rates of Production of Hot Gases in Roof Venting Experiments," Fire Safety Journal, 10:57-65. 39. Quintiere, J. G. 1977. "Growth of Fire in Building Compartment," in Fire Safety and Standard, A. F. Robertson, ed., ASTM STP 614, Philadelphia, PA: American Society for Testing and Materials, pp. 131-167.

35.

BIOGRAPHIES W. K. Chow Dr. Wan Ki Chow is a Principal Lecturer at the Department of Building Services Engineering, Hong Kong Polytechnic. He taught many courses in building services engineering up to Master degree level at the Polytechnic. Interested subjects include computational fluid dynamics, fire services engineering, building energy, building materials science and indoor environment. In the past ten years, he worked on many research projects related to the above topics and at the moment, there are over ten Ph.D. research students under his supervision. He is a Chartered Engineer; a Member of the Chartered Institution of Building Services Engineers, U.K.; the Hong Kong Institution of Engineers and the American Society of Mechanical Engineers. W. K. Wong Mr. Wing Kwong Wong is a research student at the Department of Building Services Engineering, Hong Kong Polytechnic. Before joining the Polytechnic, he worked as a building services engineer in the industry for many years. His research project is on modelling the fire environment of atrium buildings and evaluating the associated fire services design.