Long-life variable amplitude fatigue behavior of steel bridges

The principal objective of this study is to investigate the long-life fatigue behavior of bridge details subjected to service stresses.;Variable amplitude fatigue tests were performed on three types of steel bridge details under gross vehicle weight (GVW) loading, with few or no stress ranges exceeding the constant amplitude fatigue limits. Effects of the fatigue limit, heavy trucks and stress ratio on long-life fatigue strength were determined. The variable amplitude fatigue lives were predicted with the S-N curve approach (equivalent stress range and simplified models) and the fracture mechanics approach. Guidelines are recommended for fatigue design of steel highway bridges. The findings are as follows: (1) The gross vehicle weight (GVW) spectrum provides a more realistic testing spectrum for highway bridge details than does the Rayleigh distribution. (2) There is a variable amplitude fatigue limit and it should be placed a factor of two below the constant amplitude fatigue limit. (3) A few high stress ranges proportional to GVWs greater than 445 kN prolonged fatigue life rather than shortened it. However, overloaded trucks should not be allowed on highways because they damage the pavement and concrete decks. (4) Stress ratio had no statistically significant effect on the fatigue strength. (5) As equivalent stress ranges decreased, the fatigue lives exhibited a runout trend. The equivalent stress range model overestimates the fatigue lives, while the straight-line model of the AASHTO specifications underestimated them. The simplified model predicts the mean trend of data very well. (6) The fracture mechanics approach predicted the mean trend of data well for one detail (A588 steel stiffener), but underestimated fatigue life of another (A572 steel stiffener).