Figure 32. — Model of bridge approach span (42-inch size) burning in wind tunnel. 



These two cases simulated two possible 

 levels of ignition. Although the ignition was 

 different, the circulation patterns in the cavi- 

 ties are identical. In the photograph, figure 32, 

 a similarity to the expected eddy patterns 

 shown in figure 31 is evident. There is great 

 likelihood then that eddy patterns were cre- 

 ated in the cavities below the deck, enhancing 

 the fire activity at the bridge. 



Creosote preservative on the wood struc- 

 tures of the bridge and vegetation near the 

 bridge approaches certainly contributed to 

 the ignition process; however, the wind and 

 the associated eddies produced within the 

 cavities under the bridge are of greater impor- 

 tance. If there had been no cavities, it is likely 

 that burning would have been greatly reduced 

 except at the edges, resulting in much slower 

 burning and possibly more unbumed residue, 

 rather than the complete burnout of the ap- 

 proach spans that actually occurred. 



Winds at the bridge site were primarily fire 

 induced and generated in the drainage in an 

 area remote from the bridge. They were "in- 

 draft" winds resulting from strong convective 

 upward flow at the generating source. These 

 winds were sustained as other areas in the 

 drainage burned, creating additional genera- 

 tors. Furthermore, there were enough other 

 areas to sustain winds for the equivalent of at 

 least four residence times, thus accounting for 

 1 hour of intense fire activity at the bridge 

 site. This was certainly enough time for the 

 bridge to have achieved sustained combustion 

 without the necessity of continued sources of 

 wind and heat. It is very likely that these in- 

 draft winds were present at the bridge because 



the creekbed acted as a relatively unob- 

 structed trough for the flow of air. 



Damage to Steel Runways 



Heat transfer to the steel runways was suffi- 

 cient to cause a large reduction in the tensile 

 strength as evidenced by the distortion of the 

 plates shown in figure 30. Considering the den- 

 sity of steel we may assume that it would de- 

 form appreciably under its own weight if it lost 

 three-fourths of its peak strength. Examination 

 of the variation in steel strength with tempera- 

 ture indicates that a reduction of strength of 

 that extent would be achieved at 1200° F. This 

 is in no way indicative of the temperature out- 

 side the immediate bridge area. Contrary to 

 news reports, there is no evidence that the steel 

 had reached the melting point, 2,800° F. 



Conclusions 



An examination of the bridge and the 

 simulated bridge fire suggests that: 



1. The bridge approaches were ignited by 

 ground fuels. 



2. Once ignited, the bridge continued to 

 bum independent of the heat supplied by sur- 

 rounding forest fire. 



3. The bridge design contributed to the 

 fire damage, since eddies formed by the wind 

 normal to the bridge produced and captured 

 pockets of burning gas on the underside of 

 the bridge. 



4. The distortion of the steel plates merely 

 indicates that a fire was present long enough 

 to heat the plates to 1,200° F. or more (but 

 not up to the melting point, 2,800° F.) 



5. The bridge fire is not an indicator of 

 the intensity of the surrounding forest fire. 



36 



