重庆交通大学学报(自然科学版) ›› 2011, Vol. 30 ›› Issue (S2): 1152-1158.DOI: 10.3969/j.issn.1674-0696.2011.supp2.013
John W.Fisher
John W.Fisher
摘要: Fifty years ago,fatigue design was not considered a serious bridge performance issue in the US. Fatigue design utilized the concept of limiting the maximum stress at details by using the R-ratio of minimum/maximum stress and an assumption that the fatigue limit occurred at 2 million cycles. The HS-20 design truck was used for specific cycles based on the type of road and the average daily truck traffic ( ADTT) . Fatigue cracks were observed to develop in steel bridges in the late 1960’s and early 70’s. This development and the results of large scale beam tests on welded details that were carried out between 1967 and 1987 provided a statistically sound data base and established that stress range alone was the only statistical significant design stress. The fatigue provisions adopting stress range were changed by AASHTO in 1974. This meant that only the live load cyclic stress was important. The basic fatigue resistance curves provide a log normal lower bound design basis for fatigue life that is used world wide today. Experience with steel bridges also showed that a rule adopted after WWII which avoided welding to tension flanges,has resulted in extensive web gap cracking from out of plane distortions. During the last decade these small web gaps have resulted in unanticipated brittle fractures as a result of tri-axial stresses even in the absence of fatigue crack growth.
中图分类号: