Prediction and Analysis of Fatigue Life of Power Installation Tools Based on Improved Particle Swarm Optimization

In recent years, electricity has increasingly become an indispensable energy source. Compared with the traditional thermal power generation, the emerging wind power generation has the advantages of pollution-free power generation and wide area. At present, the internal failure mechanism of power devices caused by junction temperature fluctuation has been deeply studied in China, and on this basis, the fatigue life prediction method of power modules and the reliability of power devices in complex environments have been developed. In this paper, IGBT power module is taken as the research object, and detailed research is carried out in the aspects of module electro-thermal model construction, fatigue life prediction and working reliability evaluation. The relationship model between material property parameters, stress levels, remanufacturing process influencing factors and fatigue life is established by using the backpropagation neural network (BPNN) optimized by the second-order particle swarm optimization algorithm, and the life prediction is conducted for remanufactured workpieces. By combining the improved particle swarm optimization algorithm and BP neural network, the life prediction model of electronic components is established. The model is used to predict the life of electronic components under the same stress conditions, and the life prediction under the stress acceleration conditions is also made. Experiments show that the algorithm has stronger nonlinear fitting ability and higher accuracy.