Fatigue improvement of steel bridges with high-frequency mechanical impact treatment

Sammanfattning: This thesis investigates the performance of fatigue-improved welds with high-frequency mechanical impact (HFMI) for application on new bridges. Fatigue strength improvement with HFMI can enable lightweight design of bridges and allow the utilisation of the benefits of high-strength steels. Studies of various bridge types were performed in this thesis showing that 20% material saving is possible in the main load-carrying members through post-weld treatment and the use of increased steel grades (fy > 355 MPa) where necessary. Limitations of the application of HFMI treatment on bridges were also identified, related to the degree of improvement and choice of steel grade. Experimental work of HFMI-treated joints with thick main plates relevant for bridges is scarce in the literature and comprehensive studies on the thickness effect are few. Therefore, the thickness effect was studied based on an established database of 582 fatigue test results of different types of HFMI-treated joints, collected from 28 studies. It was shown that the thickness effect becomes weaker than what is recommended for as-welded joints as a result of HMFI treatment. Fatigue experiments were conducted on a typical fatigue-prone detail in steel bridges with load-carrying plates of 40 and 60 mm which showed a significant fatigue strength improvement after HFMI treatment, exceeding recommended fatigue strengths given by the International Institute of Welding. Based on the fatigue experiments, a weak thickness effect was derived for non-load-carrying transverse attachment joints where the attachment and weld sizes are kept constant. The performance of HFMI-treated welds in composite steel and concrete road bridges was studied through a state-of-the-art review and simulations of variable amplitude in-service stresses in four case-study bridges in Sweden. It was shown that, in such bridges, very high and varying stress ratios are present due to a high portion self-weight stresses, which constitute up to 50% of the highest total stresses. Furthermore, it was revealed that the fatigue-critical locations in HFMI-treated bridges remain unchanged compared with conventional bridges and that compressive overloads pose no detrimental effect that requires additional attention in the fatigue assessment. Variable amplitude experiments with a bridge spectrum load from the case studies were conducted, including both low and high mean stress tests. The low mean stress tests performed equally or better than the constant amplitude fatigue strength, confirming that bridge loads do not pose any additional damaging effect for non-load-carrying transverse attachment specimens. The high mean stress tests clearly reflected the detrimental effect of high tensile self-weight stresses and enabled verification and development of approaches to consider these effects in design.