EXPERIMENTAL PROCEDURE 



The experiment was conducted in the combustion chamber of the Northern Forest Fire 

 Laboratory. The combustion chamber is connected to an environmental conditioning facil- 

 ity capable of maintaining humidity within ±1.5 percent and temperature within ±33° F 

 (±0.5° C) . The entire series of experimental burns was run at 80° F (27° C) and 20 

 percent relative humidity, resulting in a 5 percent equilibrium moisture content for 

 the fuel. The burns were conducted under an exhaust flue in the center of the combus- 

 tion chamber 44 by 44 by 66 ft (13.4 by 13.4 by 20.1 m.) 



A wire mesh basket (13 by 15 mm grid mesh size), 1 ft^ (929 cm^) in circular area 

 and 0.5 ft (15.2 cm) deep was used to contain the fuel. Square cut aspen wood excelsior 

 {Populus tremuloides Michx.), 0.07 cm in cross-sectional dimension (a fine fuel) was 

 used; the same fuel was used in the original spreading fire experiment to which the 

 burning basket results are to be compared. The particle density was 0.40 g/cm^. Dif- 

 ferent packing ratios were obtained by changing the load while holding the depth fixed. 

 A doughnut-shaped basket with an annular space of 0.5 ft (15.2 cm) and identical depth 

 and packing ratio surrounded the inner basket (fig. 1) to eliminate edge effects and 

 thus maintain a horizontal planar combustion front as the fire moves downward through 

 the inner basket. Separation between the baskets is highly important. An annular space 

 of one-fourth inch (0.6 cm) separating the inner and outer baskets was optimum for these 

 tests. A larger space will result in increased burning within the open annular space 

 deteriorating the horizontal combustion front, whereas a smaller space will result in 

 serious disturbance to the load measurement through connecting bridges occasioned by 

 heat warping the adjacent wire mesh walls of the baskets. 



A string presoaked in xylene, a highly flammable fluid, was laid back and fourth 

 over the upper surface of the fuel to insure uniform ignition. The load was monitored 

 by a load cell transducer located below the inner fuel basket and insulated from the 

 heat by baffles tliat also prevented airflow through the annular space between the center 

 basket and the outer ring of fuel. 



RESULTS 



Figure 3 is a representative set of fuel load histories for each packing ratio, 3. 

 The load-loss rate was obtained simultaneously with the load history by electronically 

 differentiating the signal from the load cell transducer. The differentiated signal 

 was also used to determine the first indication of weight loss for the curves in figure 

 3 because of its greater sensitivity to changes in the load. Except for 6 - 0.065, each 

 curve has an initial negligible load loss. 



3 



