Design of an Actively Cooled Plate Calorimeter for the Investigation of Pool Fire Heat Fluxes*

For final qualification of shipping containers for transport of hazardous materials, thermal testing in accordance with regulations such as 10CFR71 must be completed. Such tests typically consist of 30 minute exposures with the container fully engulfed in flames from a large, open pool of JP4 jet engine fuel. Despite careful engineering analyses of the container, testing often reveals design problems that must be solved by modification and expensive retesting of the container. One source of this problem is the wide variation in surface heat flux to the container that occurs in pool fues. Average heat fluxes of 50 to 60 kW/m2 are typical and close the values implied by the radiation model in 10CFR71, but peak fluxes up to 150 kW/m2 are routinely observed in fues (Keltner, et al, 1990). Heat fluxes in pool frres have been shown to be a function of surface temperature of the container, height above the pool, surface orientation, wind, and other variables (Nicolette and Larson, 1990). If local variations in the surface heat flux to the container can be better predicted, design analyses will become more accurate, and fewer problems will be uncovered during testing. The objective of the calorimeter design described in this paper is to measure accurately pool fire heat fluxes under controlled conditions, and to provide data for calibration of improved analytical models of local flame-surface interactions. The calorimeter design consists of an actively cooled plate as shown in Figure 1. The initial configuration consists of a water cooled flat plate that is 1 m square. Later configurations may be tried to simulate different surface geometries as shown in Figure 2. The purpose of the water cooling is twofold: first, it permits approaching steady state surface temperatures during the frre, and, second. by measuring water temperature rise arrt flow rate, it allows determination of the heat flux to the cooled surface. Segmentation of the surface into zone.. permits some local resolution of surface heat fluxes. The vertical flat plate geometry was chosen for the initial experiments because it matches a geometry already analyzed by one of the authors (Nicolette and Larson, 1990). Water cooled calorimetry also has some advantages over methods used for previous similar experiments. In the past (Gregory, et al. 1989, Gregory, et al. 1987, Nelsen 1986, Longenbaugh, et al. 1990) transient inverse beat conduction methods have been used to estimate surface temperatures and heat fluxes. The inverse technique consists of monitoring temperature rises at internal calorimeter or shipping container locations, and then solving the heat conduction problem \"backwards\" to estimate surface heat fluxes and temperatures that are consistent with the internal temperatures. Such tests have shown (Keltner, et al. 1990) that \"massively thermal\" objects behave differently in frres than smaller objects. Indications are that the object size or surface temperature of the container can play a role in determining local heat fluxes that are beyond the effects predicted from the simple radiative heat transfer laws. The analytical model described briefly here and in Nicolette and Larson, 1990 can be used to understand many of these characteristics. Unlike the previous experiments that provide only a brief time at each surface temperature as the calorimeter heats up, the current approach will allow a more careful near steady-state investigation of the effect of surface temperature and other variables. The technique also lends itself to the easy inclusion of other diagnostic methods such as radiometers, intrinsic thermocouples, heat flux gauges, and fiber optic probes. By performing the initial tests in the Smoke Emissions Reduction Facility (SMERF), a wind shielded facility, a major source of test-to-test experimental variation will be removed. Later tests in open pools will be used to assess wind effects.