Modeling of fatigue crack behavior and microstructure in high carbon high strength Q/T steel welds

The most important failure cause in automotive industry is believed to be fatigue. Hence, predicting the lifetime of critical components would be of interest for manufacturers. Computerized methods for predicting the fatigue lifetime as well as microstructural predictions could always be helpful once utilized properly. Amongst numerous components used in automotive industry, limited numbers are subjected to torsion especially if they are welded, but their reduced quality could lead into catastrophic failure. As a matter of fact, the fatigue properties of structures subjected to torsion loading is less investigated than the properties under other loading conditions. Combining the ideas above, this work has been developed based on modeling, simulating and predicting plenty of characteristics in high strength steel welds subjected to alternating torsion loading, using a number of softwares. Numerous approaches for predicting fatigue lifetime in such components alongside of microstructural prediction and control in those welds were engaged to help the automotive industry grasp a deeper knowledge about such specific cases. A test specimen which can include both welded high strength steels and specifications for bearing torsion loading has been designed. For analyzing the stress distribution on such specimen under defined torsion loading, ABAQUS was used. Then the model was imported into FRANC3D which specifically looks into fatigue via Linear Elastic Fracture Mechanics (LEFM) method. Afterwards, several other methods and hypotheses were employed to do the same procedure for comparison. For instance; strain-base hypotheses, Elasto-Plastic Fracture Mechanics (EPFM), rain-flow analysis for spectrum loading cases and hot-spot analysis were employed. The other interesting part of prediction relates to weld’s microstructure. For predicting the microstructure, both cooling rates after welding and CCT diagrams should be available. JMatPro® was predicting CCT curves for those joined steels and a Matlab code was predicting cooling rates visually, using Rosenthal equations. Coupling Matlab and JMatPro® resulted in compatible results compared to what was gained experimentally. At the end, the conditions and influence of residual stresses have been studied