ABSTRACT Mechanical components are susceptible to fatigue cracks when subjected to prolonged and complex external alternating loads, which can ultimately lead to component failure. The introduction of advantageous compressive residual stress can counteract external tensile stress, thereby enhancing the fatigue life of mechanical components. However, when these components are subjected to mechanical loads and high temperatures, residual stress relaxation (RSR) may occur, which can affect the fatigue performance of the components. In recent years, RSR has been extensively studied by experiments and simulations. In this paper, we summarize the influencing factors of RSR in terms of cyclic loading, temperature, work hardening, and initial stress level, as well as discuss the intrinsic mechanism of RSR in terms of four types of dislocation motion: annihilation, proliferation, rearrangement, and slip. Given the practical needs of industrial production, data‐driven machine learning methods show potential in residual stress prediction, but data quality and model interpretability are still key challenges. In this regard, we propose the future focus of research in machine learning for residual stress prediction.
Bao et al. (Fri,) studied this question.