On the basis of analytical and experimental studies, revisions to the acceptance criteria for imperfections in steel bridge welds are proposed. The analytical work, based on linear elastic fracture mechanics, was concerned with evaluating combinations of imperfection size and stress range which would lead to fatigue failure. The experimental effort involved measurement of fatigue crack growth rates under a variable amplitude load spectrum realistic for highway bridges. These tests on AASHTO structural steels M222 and M244 (ASTM A588 and A514) investigated crack growth rates in a regime not previously investigated, namely, at delta K between 1.5 and 10 ksi/in., an appropriate range for evaluating fatigue in steel highway bridges. Fatigue studies were also conducted, using the same load spectrum, on specimens containing deliberately introduced weld imperfections, porosity and entrapped slag. Although the fatigue lives in these tests were longer than would be predicted for sharp cracks of similar initial size, seven of the eight specimens showed evidence of crack growth originating from weld imperfections. Whether or not a weld imperfection will be the origin of a fatigue failure is strongly influenced by the magnitude of the effective stress range; and the proposed weld acceptance criteria incorporate scaling of the allowable weld discontinuity size based on the anticipated stress range. (A)
Abstract