1、2008 ASHRAE 147ABSTRACTThe rapid smoke spread through an atrium in case of fireis a major concern. Although natural ventilation can be usedto keep the smoke layer at high levels, in some cases, such asystem may not be effective allowing smoke to reach low levelsin the atrium and it is not frequently
2、 used in North America,where mechanical ventilation is the preferred atrium smokemanagement system used. Guidelines for the design ofmechanical smoke exhaust systems require that the maximumvelocity of the make-up air velocity be restricted to 1 m/s, acriterion that for some buildings causes great d
3、ifficulties todesigners. In this study, a CFD model was used to evaluate thiscriterion for make-up air velocity and to determine if this limitis appropriate. The study considered different fire sizes in vari-ous size atria equipped with smoke exhaust systems. The resultsof the analysis indicate that
4、, for some cases, increasing thevelocity limit may have a negative impact on the fire plume andthe hot layer height.INTRODUCTIONMost atria have a large undivided space, designed forcreating visual and spatial appeal. One of the concerns asso-ciated with atria is fire safety. When a fire occurs in an
5、 atrium,smoke can fill the atrium and the connected floors blockingexit routes and endangering occupants. In North Americamechanical exhaust systems are used to extract smoke fromthe atrium to maintain the smoke layer at the desired height. Design requirements for smoke exhaust systems areprovided b
6、y NFPA 92B 1 and Klote and Milke (1992) 2.Current design requirements set a maximum make-up airvelocity of 1 m/s to prevent disruption of the plume. Accordingto NFPA 92B 1, the 1 m/s criterion is based on limitedresearch into the effect of wind on flames. The work is citedin the SFPE Handbook of Fir
7、e Protection Engineering 3 4.Many designers have stated that meeting the 1 m/s require-ment is often costly and presents a hardship. For example: tomaintain a 40-m clear height in a 50-m tall atrium, with a firesize of 1 MW, the mass flow rate of the smoke exhaust shouldbe 288.91 kg/s. With a make-u
8、p air velocity of 1 m/s, the areaof the opening providing this air should be 347 m2. If thisopening is put at ground level, and assuming a height of theopening of 3 m, the length of the opening should be 115 m. Formany buildings such an opening may not be feasible. There is,therefore a need to inves
9、tigate the 1 m/s make-up air velocitycriterion to determine whether it is too conservative.This paper investigated the impact of make-up air velocityon the fire plume and the interface height. A computationalfluid dynamics (CFD) model was used to evaluate the impactof make-up air velocity on the eff
10、ectiveness of an atriumsmoke exhaust system to determine whether the currentrestriction of 1 m/s is justified. Comparisons of results ofdifferent fire scenarios are presented in order to explain themechanisms of smoke flow in atria when an air jet impacts thesmoke plume.Mechanical ExhaustMechanical
11、exhaust systems use either a dedicatedexhaust system or the exhaust fans of the HVAC system. Kloteand Milke 5 present a method of analysis of a mechanicalexhaust system that is based on the following simplifyingassumptions:Evaluation of Atrium SmokeExhaust Make-Up Air VelocityGeorge Hadjisophocleous
12、, PhD, PE Jian ZhouMember ASHRAEJian Zhou is a student and George Hadjisophocleaous is a professor in the Department of Civil and Environmental Engineering at CarletonUniversity, Ottawa, Canada. NY-08-020 (RP-1300)2008, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.
13、(www.ashrae.org). Published in ASHRAE Transactions, Volume 114, Part 1. For personal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitted without ASHRAEs prior written permission.148 ASHRAE Transactions The only mass flow into the smoke la
14、yer is the fireplume.The only mass flow from the smoke layer is the smokeexhaust.The exhaust is removing only smoke and not any airfrom below the smoke layer.The smoke layer height is constant.The flows into and out of the smoke layer are at equi-librium.Heat transfer between the smoke layer and the
15、 surround-ings has reached steady state.The mass flow rate of an axisymmetric plume mpat differ-ent heights Z can be computed using Equation 1 1. (1)wheremp= mass flow rate of plume at height Z, kg/sQc= convective potion of the heat release rate, kWZ = hot layer height, m According to the above assu
16、mptions, to maintain the requiredclear height at steady state, the exhaust mass flow rate meshould be equal to the plume mass flow rate mpat the designheight.Make-Up Air VelocityFor an atrium smoke management system that involvesthe venting of smoke from the hot upper layer, a make-up airsupply must
17、 be provided. Once the exhaust rate for the smokecontrol system is identified, the atrium design must be capableof providing make-up air to the space so the atrium does notbecome a vacuum. Make-up air should be introduced into theatrium below the smoke interface level. Several researchershave tried
18、to determine if there are adverse effects from usinghigh make-up air supply velocities. Heskestad 6 and Mudanand Croce 7 suggest that velocities above 1 m/s alter thesymmetric smoke plume, which results in an increase in theamount of air entrained into the plume. NFPA 92B 1 specifically states that
19、the supply velocityof make-up air at the perimeter of the atrium must be limitedto sufficiently low values so as not to deflect the fire plumesignificantly, which would increase the air entrainment rate, ordisturb the smoke interface. A maximum make-up supplyvelocity of about 1 m/s is required, base
20、d on flame deflectiondata. Where maintaining a smoke layer height is not a designgoal, plume disruption due to supply velocity might not bedetrimental.The same mass flow rate of air as the air exhausted fromthe top of the atrium needs to be supplied to the atrium belowthe smoke layer. The supply nee
21、ded to accommodate theexhaust may be provided naturally through openings or leak-age paths or by using supply fans. Klote and Milke 5 pointout that fan-powered make-up air is often sized at 90% to 97%of the exhaust airflow rate. The remaining make-up air entersthe atrium through leakage areas. Yi et
22、 al. 8, using a zone model; studied the impact ofdifferent positions of make up air supply on the performanceof a mechanical exhaust system. Three scenarios with differ-ent relative positions for providing make-up air duringmechanical exhaust were considered: smoke layer interface isabove, within an
23、d below the air inlet. The predictions by thezone model agreed well with the experimental findings. Theystate that when the position of the air supply is lower than thesmoke layer, a minimum smoke layer interface height could bemaintained for a given fire size and extraction rate. When theair supply
24、 is above the smoke layer interface, make-up airwould enter the smoke layer directly and mix with the smoke.Smoke temperature would be reduced significantly and a safesteady height of smoke layer could not be attained for this situ-ation. When the air inlet is at the interface height, the averagetem
25、perature rise of the smoke layer would be lower than thecase with the air inlet located below the smoke layer.Smoke Layer InterfaceThe purpose of an atrium smoke exhaust system is tomaintain a specified clear height in the event of a fire. Toaccommodate the ceiling jet that forms under the ceiling o
26、f theatrium, the maximum clear height (Z) that can be achieveddepends on the height of the atrium (H). Figure 1 illustratesa mechanical smoke exhaust system in operation with themaximum achievable layer height in an atrium, which main-tains the smoke layer height above the highest level whereoccupan
27、ts may be present. In this study, an upper limit for Z/H of 0.8 was chosen, which relates to the highest achievablesmoke level to ensure that the thickness of the ceiling jet iscovered. 80% is the prescriptive height noted in standardsassuming the ceiling jet is 10% of the height and the reflectedce
28、iling jet adds another 10%. It is the height required for thedesign height unless additional modeling is done In real fires and CFD predictions, the interface betweenthe lower cold layer and the upper hot layer is not clearlydefined. A transition zone exists between the two layers,within which the t
29、emperature or CO2concentration changesfrom the ambient values to those of the hot layer. The depth ofthe transition zone varies from case to case. Cooper et al. 9developed a method for defining the height of the interfacebetween the hot and cold zones produced by a fire based on alimited number of p
30、oint temperature measurements over theheight of a compartment. They assumed that the interface is atthe height where the temperature, Tnis given by:(2)whereTmax= the maximum temperature of the compartment (outside the fire plume), K.Tb= the temperature near the bottom of the compartment, KCn= interp
31、olation constant typically in the range of 0.15 to 0.2.mp0.071Qc13Z530.0018Qc+=TnCnTmaxTb()Tb+=ASHRAE Transactions 149A similar equation can also be used to determine the inter-face height based on CO2concentration. Lougheed 10, 11pointed out that the value of Cnequal to 0.2 gives a smokeinterface h
32、eight near the bottom of the transition zone while Cnequal to 0.8 gives a smoke interface height near the top of thetransition zone. For determining the interface height usingpredictions of CFD models, a value of Cnbetween 0.5 and 0.6is recommended 12. In this study, a clear height is definedusing E
33、quation (2) with Cnset at 0.6. Both temperature andCO2concentration profiles obtained from CFD were used inEquation (2), so two interface heights are defined, one basedon temperature and the other on CO2. As there is a variationbetween the temperature and CO2profiles at different loca-tions in the a
34、trium, the interface heights were computed ateach location based on both temperature and CO2concentra-tion values, and their average value was used to represent theatrium interface height.DESCRIPTION OF THE CFD MODELThis section describes the model used for the numericalsimulations, as well as some
35、of the approaches used inmodeling.The CFD computer model FDS (Fire Dynamics Simula-tor) Version 4 13 was used for this study. FDS is used exten-sively for fire applications. FDS can model fires in a singlecompartment, multi-compartments, as well as, atria and ware-houses. Atrium GeometryThis study c
36、onsidered a rectangular atrium with the fireon the ground floor. The characteristics of the atrium and thefires considered are the following:The atrium has a square cross sectional area with widthsranging from 10 m to 40 m and heights from 10 to 60 m.The atrium has an opening on one side for make-up
37、 air,with an area that is variable to provide the necessarymake-up air velocity of 0.5, 1, 1.25, and 1.5 m/s.Fire heat release rates considered are: 1, 2.5, and 5 MW.The location of fire: For atria with widths of 10 or 15 m, the fire waslocated at 0.25 L (L: the width of atrium) fromthe wall opening
38、. For atria with widths of 20, 30, and 40 m: thefire was located at 5 or 2.5 m from the opening.There are exhaust fans located at the top of the ceiling,which provide the necessary smoke exhaust. Smokeexhaust openings were uniformly distributed over theentire area of the ceiling to minimize the effe
39、ct of theceiling jet. The fan exhaust flow rate was computedusing Equation (1).Boundary Conditions The following boundary conditions were used in thesimulation:Solid wall: The walls of the atrium were modeled assolid walls covered with gypsum boards. Floor: The floor was modeled as 200 mm thick conc
40、rete. Ceiling vent: A constant mass flow rate was definedthroughout the ceiling area based on the mass flow raterequired to maintain the interface height at 0.8H, whereH is the height of the atrium.Wall openings: The make-up air opening on the wallwas assumed to be a passive opening.NUMERICAL SIMULA
41、TIONSThe study considered different make-up air velocitiesranging from 0.5 m/s to 1.5 m/s, three fire sizes 1, 2.5, and5 MW at different locations from the opening and five differ-ent atrium sizes. Figure 2 shows a sketch of the atrium consid-ered. The make-up air opening is placed on one wall of th
42、eatrium at ground level and the fire was located in front of theopening at different distances. The parameters considered forthese simulations are shown in Table 1.Fire was modeled using the heat release rate per unit areafeature of FDS (HRRPUA). The area of the fire was varied tokeep the HRRPUA at
43、1 MW/m2. This heat release rate densityis typical of fires from common fuels. The impact of thisparameter on the fire plume is described in 14. For all simu-lations, the location of the fire was in front a the opening at thedistance shown in Table 1. The openings for make-up air forall simulations w
44、ere placed at ground level. The size of theopenings was varied to achieve the required make-up airvelocity. The mass flow rate was calculated using the correla-tions in NFPA 92B such that the smoke layer height is main-tained at 0.8H, where H is the height of the atrium. This flowrate was used as a
45、boundary condition in FDS at the ceilinglevel. To minimize the impact of the exhaust characteristics onthe conditions in the atrium, the whole ceiling area was usedto exhaust smoke. For all simulations, the convective heatrelease rate was assumed to be 65% of the total heat releaserate. This is conf
46、irmed in Figure 3, which shows the total andradiative heat release rate for the 2.5-MW fire.Figure 1 Minimum smoke layer thickness.150 ASHRAE TransactionsTo consider all parameters shown in Table 1, 96 simula-tions were performed. Due to space limitations, the results ofthe simulations for the 10-m
47、atrium are discussed in detail andonly summarized results are presented for the other atriumheights. Results for 10-m Tall AtriumThe fire for the 10-m tall atrium simulations was placed2.5 m from the opening. The mass flow rate of the exhaust forthe 1-MW fire simulations was computed using Equation
48、1 tobe 20.85 kg/s. This value was used as the boundary conditionat the ceiling of the atrium. Four simulations were performedto consider the four make-up air velocities. The requiredvelocities were obtained by varying the area of the make-up airopening. To obtain the velocities of 0.5, 1.0, 1.25 and
49、 1.5 m/s, the area of the opening was set to 35.36 m2, 17.67 m2, 14.13m2and 11.79 m2respectively. Figure 4 depicts temperature variation with time at thequarter point of the atrium (Point 7 of Figure 2) and threedifferent heights. The figure shows that steady state conditionsin the atrium were reached at about 100 s. Based on this, thesimulation time for all simulation for this atrium size was setto 200 s.Figure 5 shows the temperature distributions at 200 s ona vertical plane passing through the fire centerline and theTable 1. Parameters Used for t
