Results 



Isotherms obtained from longitudinal and transverse members of a unit cell within 

 the fuel array are shown in figures 5 and 6, respectively. Two fuel sizes are repre- 

 sented: 0.6 cm. and 1.3 cm. The transverse and longitudinal members are analyzed 

 separately; no attempt is made to merge isotherms because no temperature data are 

 taken in the region of the connection. Isotherms in this region are extrapolations. 

 Consequently, the isotherms in the longitudinal members do not show masking by the 

 transverse member near their intersection. 



Areal temperature distributions from figures 5 and 6 were used in equations (6a) 

 and (6b) to obtain the nonuniform heat absorbed by the individual unit cell members up 

 to ignition (table 1) . No longitudinal member was instrumented during the first test 

 at the 0.6-cm. sample size; however, two adjacent longitudinal members were instrumented 

 in the second test at the 0.6-cm. sample size. 



Because final results are limited in figure 7, more than one function will fit the 

 data. However, the exponential form, 



e = exp (-4. 53/a) , 



is favored because: 



1. The data lie close to the exponential curve; 



2. the curve point (e=1, 1/a = 0) acts as a constraint on the possible 



curves that can pass through the data points; and 



3. the region of the curve covering small particle sizes (near e=1) exhibits a 

 linear relationship with 1/a, as suggested earlier by Thomas (1967) with 



A oc i/a. 



Although data are confined to fuels small in cross section, extrapolation to 0.6 cm. is 

 dependable. This judgment is based on reasonable results obtained by using the above 

 exponential relationship in Rothermel's (1972) fire spread model. 



11 



