A significant number of high-performance engineering structures are repeatedly subjected to both thermal and mechanical loads, often in a combined fashion. However, because of the increase in the plasticity of metallic structures when they are loaded at high temperatures, the analysis become very complex. This presents a significant obstacle for the comprehension of both the growth of cracks and the thermo-mechanical fatigue performance of the material. Thermomechanical fatigue and thermal fatigue are characterized by external and internal constraining forces, respectively. The beginning and spread of thermal fatigue cracks are controlled by a variety of factors, including the modes of heating and cooling, the temperature range, the maximum temperature rates, and the holding times. The process of a crack beginning and the rate at which it spreads are both sped up when the temperature is raised. However, because of the development of powerful statistical learning algorithms as well as rapid advancements in computational power, there has been an increased adoption of machine learning approaches as well as other advanced computational analyses and numerical software for crack damage detection and damage severity. This has led to an increase in the use of these methods.