1、2008 ASHRAE 329ABSTRACT There have been a number of concerns regarding thebalcony spill plume equation provided in North Americanstandards and codes. These include: lack of verification by full-scale experiments and application of the equation for highatria, even though it was developed for low-heig
2、ht atria. As aresult of these concerns, the American Society of Heating,Refrigerating and Air-Conditioning Engineers (ASHRAE)initiated a project to evaluate the balcony spill plume equationused in North American codes and standards (RP-1247). Theresearch project included computational fluid dynamics
3、(CFD) modeling studies to investigate smoke entrainment inthe balcony area and in high atria. The primary researchconducted as part of RP- 1247 was a series of full-scale exper-iments conducted to investigate smoke flow in balcony spillplumes and the resulting mechanical exhaust requirements foran a
4、trium. The full-scale experiments included measurementsinside the fire compartment and in the opening between the firecompartment and the balcony area. They also includedmeasurements in a simulated atrium space. In Part 1 of thispaper, the results of the measurements in the fire compartmentarea are
5、discussed. The data were also analyzed to estimate themass flow rate through the compartment opening. The esti-mated mass flow rates are compared with algebraic equations,which are used to estimate the mass flow through a compart-ment opening.INTRODUCTIONAtriums have become popular elements in comme
6、rcial,office, and residential buildings because they can provideattractive, environmentally controlled, naturally lit spaces.Such spaces, however, present a challenge for fire protectionengineers because their height (typically greater than 20 m)decreases the effectiveness of automatic sprinkler sys
7、tems andbecause they lack the floor-to-floor separations that can limitthe likelihood of fire and smoke spreading from the floor of fireorigin to other areas of the building. Evacuation routes in atri-ums are of greatest concern because they become vulnerable tospreading smoke unless smoke managemen
8、t measures areused. As a result, specific requirements are included in buildingcodes for atria (ICC 2003, 2006; NFPA 2006; NRCC 2005).In recent years, approaches to smoke management in atriahave been developed and are provided in standards and engi-neering guides (NFPA 2005; Klote and Milke 2002; Mo
9、rganet al. 1999). Smoke management systems can be used toaccomplish one or both of the following (NFPA 2005): Maintain a tenable environment in the means of egressfrom large-volume spaces during the time required forevacuationControl and reduce the migration of smoke between thefire area and adjacen
10、t spacesNFPA 92B (2005) provides algebraic equations for thedesign of smoke management systems for three design firescenarios:The fire is located on the floor of the atrium, and smokeproduction includes the air entrainment into the plumeas it rises to the ceiling (axisymmetric plume).The fire is loc
11、ated in an adjacent space, and the smokeflows through a compartment opening and subsequentlyunder a balcony before entering the atrium space (bal-cony spill plume).A ventilation-limited fire is located in an adjacent space,Balcony Spill Plumes:Full-Scale Experiments, Part 1G.D. Lougheed, PhD C.J. Mc
12、CartneyMember ASHRAEG.D. Lougheed is a senior research officer, and C.J. McCartney is a technical officer in the Fire Research Program, National Research Coun-cil, Ottawa, Ontario, Canada.NY-08-039 (RP-1247)2008, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.as
13、hrae.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.330 ASHRAE Transactionsand the smoke enters the atrium through a wi
14、ndow,which is broken by the fire (window plume).Initially in North America, the design of atrium smokemanagement assumed that the fire was located on the floor ofthe atrium and that the smoke production could be estimatedusing the axisymmetric plume equation. It was assumed thatsprinklers would limi
15、t the size and, thus, the smoke producedby fires in adjacent spaces. In recent years, smoke manage-ment requirements for atria in the US building codes specifythat the design of atrium smoke management systems includesdesign fire scenarios in both the atrium and adjacent spaces(UBC 1997; ICC 2003, 2
16、006). Also, research including full-scale fire tests indicates that sprinkler-controlled fires cangenerate significant quantities of buoyant smoke (Morgan etal. 1999; Madrzykowski and Vettori 1992; Lougheed 1997;Lougheed et al. 2000; Lougheed et al. 2001; Ghosh 1997). Asa result, it has become neces
17、sary to consider design firescenarios involving fires in both the atrium space and adjacentspaces in the design of an atrium smoke management system.There have been a number of concerns regarding thebalcony spill plume equation provided in NFPA 92B (2005)and ICC (2003): the lack of verification by f
18、ull-scale experiments andthe application of the equation for high atria eventhough it was developed for low-height atria.As a result of these concerns, the American Society ofHeating, Refrigerating and Air-Conditioning Engineers(ASHRAE) initiated a project to evaluate the balcony spillplume equation
19、 used in North American codes and standards(RP-1247). For this project, three research activities wereundertaken: 1. Full-scale experiments2. Computational fluid dynamics (CFD) modeling to inves-tigate smoke entrainment below the balcony and at thebalcony edge as the plume spills into an atrium3. CF
20、D modeling of smoke entrainment into a balcony spillplume for high atriaThe CFD modeling studies were undertaken to addresstwo concerns with the experimental program: The distance between the balcony and the ceiling in theexperimental facility was limited (5 to 7 m). This dis-tance exceeds the dista
21、nce required in the initial requestfor proposals. However, in comparison to scenarios inmany North American atria, this distance was a concern.The second concern was the inability to fully investigatethe effect of the parameters that affect air entrainment inthe balcony area using full-scale experim
22、ents. To address these issues, an effort was made to verify aCFD model using the National Institute of Standards andTechnologys (NIST) Fire Dynamic Simulator (FDS) soft-ware (McGrattan et al. 2002a, 2002b) for determining smokeentrainment into a spill plume. Detailed results of the model-ing studies
23、 are provided in master degree theses by Ko (2006)and McCartney (2006) and summarized in Ko et al. (2008) andMcCartney et al. (2008). The primary research conducted as part of RP-1247 wasa series of full-scale experiments that investigated smoke flowin balcony spill plumes and the resulting mechanic
24、al exhaustrequirements for an atrium. An extended set of full-scaleexperiments was conducted for a fire located in a compart-ment. Parameters that were varied included the width of thecompartment opening and the fire size. Tests were conductedwith and without a downstand in the compartment openingan
25、d with and without draft curtains used to channel the flowbelow the balcony.The full-scale experiments included measurements insidethe fire compartment and in the opening between the firecompartment and the balcony area. They also includedmeasurements in a simulated atrium space. In Part 1 of thispa
26、per, the results of the measurements in the fire compartmentarea are discussed. Also, the data were analyzed to estimatethe mass flow rate through the compartment opening. The esti-mated mass flow rates are compared with algebraic equationsthat are used to estimate the mass flow through a compartmen
27、topening. The results for the measurements in the simulatedatrium area are discussed in Part 2 of this paper (Lougheed etal. 2008).EXPERIMENTAL STUDIESThere have been several experimental studies of balconyspill plumes. Most of these were conducted at the BuildingResearch Establishment (BRE) in the
28、UK using 1/10 scalemodels. Experiments were also conducted by the University ofCanterbury, New Zealand, using salt-water modeling (Yii1998) and a 1/10 scale model similar to that used at BRE (Harri-son 2004). Reviews of the experimental studies are provided byHarrison (2004) and Lougheed et al. (200
29、7). A brief summaryof the experimental studies is provided in this section.Morgan and Marshall (1975)Morgan and Marshall (1975) conducted a series of 1/10scale model experiments using the scaling principles providedby Thomas et al. (1963). The physical model simulated thesmoke flow from a single sho
30、p with a width of 0.7 m (3.3 ft)and a double shop with a width of 1.4 m (4.6 ft). The firecompartment was 0.5 m (1.6 ft) deep and 0.5 m (1.6 ft) high.A 0.4 m (1.3 ft) deep balcony extended along the full width ofthe shop and had draft curtains located at both ends. Experi-ments were conducted with t
31、he front of the shop fully open,and a limited number of tests were conducted with a 0.16 m(0.5 ft) deep downstand running the full width of the opening.In addition, experiments were conducted with the doublewidth shop and variable openings. For these tests, the draftcurtains were located at the edge
32、 of the opening, and the down-ASHRAE Transactions 331stand was included. The heated flow from the apparatusflowed into the laboratory.Temperature data were used to develop the BRE spillplume method. Good agreement was obtained between theexperimental results and the theoretical predictions. Law(1986
33、) also used the maximum temperatures measured abovethe balcony for the initial development of a simplified equa-tion for a balcony spill plume.Morgan and Marshall (1979)Morgan and Marshall (1979) conducted a second series of1/10 scale experiments primarily using the 0.7 m (2.3 ft) widemodel shop. Fo
34、r these experiments, the smoke flowed into alarge box with the smoke extracted from the top of the boxusing a mechanical exhaust system at an approximatelyconstant rate. The heat source was a heater with a heat outputbetween 1 and 4 kW (0.95 and 3.78 Btu/s). Most experimentswere conducted with the d
35、raft curtains at 0.7 m (2.3 ft).However, two experiments were conducted with the draftcurtains at 1.4 m (4.6 ft), and another two were conductedwithout draft curtains.Thermocouple trees located outside the plume were usedto measure the temperature distribution in the model atrium.These results were
36、used to determine the thickness of thesmoke layer. There was no clear demarcation between thesmoke layer and the ambient air. An effective layer depth wasdetermined by integrating the temperature-height curve.The BRE spill plume method was used to compare thetheoretical mass flow rates with those pr
37、oduced in the exper-iments. The results of this study gave rise to the developmentof the effective layer depth correction to allow for the temper-ature variation beneath the smoke layer. These experimentsshowed that the channeling by draft curtains was effective inreducing the amount of smoke produc
38、ed by the spill plume.Law (1986) used the spill plume results from these experi-ments to develop a simplified calculation method.Hansell et al. (1993)Hansell et al. (1993) conducted a series of 1/10 scaleexperiments with a model atrium. The model atrium was high(3.06 m) (10.04 ft) compared with the
39、model used by Morganand Marshall (1979). It had a plan area of 3.3 m2(35.5 ft2), anda mechanical system was used to extract air at the top. The experiments were conducted to address specificissues regarding the horizontal flow of smoke toward an open-ing, air entrainment into both free and adhered p
40、lumes, and theeffect of balcony depth. Law (1995) used the results of theseexperiments to modify the earlier correlation (Law 1986) forthe mass flow rate for a balcony spill plume and for the effec-tive width of the spill plume without draft curtains.Marshall and Harrison (1996)Marshall and Harrison
41、 (1996) conducted five series oftests using 1/10 scale physical models. Each series of testsused a different atrium physical model to investigate specificissues, including the effect of atrium size, combustion airflowthrough the compartment opening, and plume end effects onair entrainment into the p
42、lume. These experiments were usedby Poreh et al. (1998) to develop a simplified spill plumeexpression.Yii (1998)Yii (1998) conducted a study on spill plumes using salt-water modeling and a laser induced fluorescence (LIF) flowvisualization technique using a 1/20 scale model. These exper-iments inves
43、tigated the qualitative features of the smoke flowbut did not provide any quantitative information.Harrison (2004)Harrison (2004) conducted 1/10 scale physical modelexperiments. The fire compartment was 1 by 1 by 0.5 m (3.3by 3.3 by 1.6 ft) high with a 0.3 m (1.0 ft) deep balcony. Thecompartment ope
44、ning extended the full height of the compart-ment, and the width of the opening was varied using inserts.Full-height draft curtains were located at the edge of the open-ing.A series of 55 tests were conducted with a 0.6 m (2.0 ft)wide opening. The primary parameters for this series of testswere the
45、heat release rate, the compartment opening width,and the depth of the downstand at the spill edge. Four testswere conducted with the downstand located at the compart-ment opening. A single fire size was used for these tests, andthe opening width and downstand depth were varied.Harrison (2004) compar
46、ed the measured spill plume massflow rates with those determined using the BRE method. Analgebraic equation was also developed for the smoke flow ina balcony spill plume equation for the scenario with a down-stand at the spill edge. Also, an empirical relationship wasdeveloped for air entrainment be
47、neath a balcony.FULL-SCALE EXPERIMENTAL ARRANGEMENTExperimental DesignThe main objective of the experimental program was todetermine balcony spill plume mass flow rates produced bysteady fires in a compartment with an attached balcony. Theatrium smoke layer elevation was the major dependent vari-abl
48、e to be measured. However, in reviewing the literature usedto develop the present correlations for air entrainment intobalcony spill plumes, it was noted that some key parametersthat affect the mass flow rate in the balcony spill plume alsoaffect the conditions within the fire compartment and the ma
49、ssflow rate through the compartment opening. These parametersinclude the fire size (heat release rate), the width of thecompartment opening, and the presence of a downstand in theopening. In Part 1 of this paper, the experimental results areused to investigate the effects of these parameters on the condi-tions inside the fire compartment and on the smoke flowthrough the opening. The air entrainment in the simulatedatrium space is discussed in Part 2 of this paper.332 ASHRAE TransactionsThe parameters that were investigated in the full-scaleexperiments that affected the conditions within t
