Documents / Document
This is the final highway accident report (NTSB-HAR-71-1) from the National Transportation Safety Board, adopted in December 1970. It covers the collapse of the U.S. 35 bridge between Point Pleasant, West Virginia, and Kanauga, Ohio, on December 15, 1967, which killed 46 people. The Board found that the cause was a cleavage fracture in eyebar 330 at joint C13N. That fracture grew from a flaw produced over 40 years by stress corrosion and corrosion fatigue, in a spot that could not be seen or found by inspection.
“Cooper”1 page
Read from the scan by GLM-OCR; expect the odd misread word.
(b) The material was cold worked near the hole surface, increasing its susceptibility to corrosion-fatigue, probably to a greater degree than its susceptibility to stress-corrosion cracking. (c) Contaminant concentrations in the field were low relative to those used in the laboratory to establish stress-corrosion susceptibility. (d) A variable stress level was present, although it was small. If corrosion fatigue was responsible, it is logical to ask why such cracks did not develop first in the eyebars between U0 and U7 of the side span, where the range of live load stress contributing to fatigue was much larger and the dead load stress related to stress-corrosion cracking was smaller. The influence lines (see Figure 34) for these members have negative segments for loads within the side span, due to their participation in the stiffening truss. The member loads are shown in Table 7. It is also interesting to examine the probable frequency with which these maximum design live load stress ranges will be reached. As noted previously, the probability of 100 percent of design stress is quite remote. The live load at the time of collapse in eyebar C11-C13 was only 237.3 kips, or about 41 percent of design live load. This was due to a load extending from the west end of the structure to just beyond the center of the main span. The load in bar U5-U7 at the instant of collapse was only 142.5 kips, or about 17 percent of full live load tension. For this bar, however, the range of stress was greater than in C11-C13, since while the loads were accumulating only on the Ohio side span, the stress was of opposite sign. The total load in the side span was 125.1 kips or 330 lbs/ft., which is 47 percent of design live load and would have produced a compressive load in U5-U7 of about 200 kips. Thus, the load range for this bar was 342 kips. Loadings which did not extend very far into the center span produced even greater differences in stress range. Moreover, trains of vehicles which move continuously across the structure while maintaining a total length about equal to the length of the main span would produce approximately 85 percent of maximum range (495 kips) in C11-C13, but 96 percent of maximum range (120 kips) in U5-U7. Thus it would appear that the general range of stress in U5-U7 would be about twice that of C11-C13 for any given probability of occurrence. This would reduce the fatigue life by a factor of ten or more, or in other words, make the occurrence of a fatigue fracture in U5-U7 ten times more probable. It therefore appears that the stress-corrosion was the dominant mechanism. in spite of the fact that there is evidence
Document, cited by the archive. The PDF is mirrored here; the original link is above. The text was read from the page images by GLM-OCR; expect the odd misread word. 202 pages are in the text index: search them above, or from the library's search.