Influence of small wheel defects on the risk of subsurface-initiated rolling contact fatigue of railway wheels

Abstract

Due to improved quality control in wheel manufacturing and maintenance actions in railway operations, subsurface-initiated rolling contact fatigue rarely occurs these days. However, in some cases, fatigue poses a risk to trigger accidents such as derailments. Therefore, it requires careful monitoring in wheel maintenance. The thesis employed Finite Element simulation (FE-simulation) and impact loads from wheel impact load detectors to study subsurface-initiated rolling contact fatigue. First, the influence of finite boundary effects on the subsurface stress field was studied by comparing stress evaluations from elastic Finite Element simulations with the analytical solutions for semi-infinite bodies. The study included an investigation of the lateral contact load position. It demonstrated that the finite boundary effects from the finite wheel body and laterally shifted contacts were insignificant. Second, the fatigue initiating defect sizes along depths were estimated using the Dang Van multiaxial criterion and Murakami’s theory. The estimation derived the Dang Van equivalent stress along depths with two cases, (1) extreme load and (2) wheel impact loads corresponding to wheel impact load detector measurements. It was found that a wheel defect at the depth from 1 to 10 mm below the wheel tread would more likely initiate subsurface rolling contact fatigue. The minimum fatigue initiating defect size (the critical defect size) occurred at depths 3 ∼ 4 mm beneath the wheel tread in both cases. The critical defect size was around 1.2 mm and 2.9 mm in each load case with non-detrimental defect size, do = 40 μm. Finally, the thesis estimated the fatigue life of a wheel containing a defect at a depth of 6 mm from the peak and mean impact loads measurements. The estimation derived a reduced equivalent Dang Van Wöhler curve (often called SN curve) due to material defects. The reasonable fatigue life estimates were around 191×103 and 330×103 km for a defect size 1 mm and non-detrimental defect size, do = 40 μm. The thesis confirmed the influence of a small defect (defect size of less than 1 mm) on subsurface-initiated rolling contact fatigue with given realistic impact loads. However, the analysis results were highly dependent on non-detrimental defect size, do, equivalent fatigue limit, σeDV, and fatigue limit cycle, Ne. Therefore, it needs reliable sources for the selection of fatigue properties and durable assumptions to obtain more accurate results.