FATIGUE CRACK GROWTH PREDICTION BASED ON IPSO-PF ALGORITHM

The traditional Paris formula ignores the influence of various uncertain factors in the crack growth process, which leads to a big difference between the predicted crack growth process and the real crack growth process. In order to improve the prediction accuracy of fatigue crack growth, a fatigue crack growth prediction method based on the improved particle swarm optimization particle filtering (IPSO-PF) algorithm was proposed. Firstly, based on the framework of the particle filtering (PF) algorithm, the particle swarm optimization (PSO) algorithm was used to optimize some particles based on the updated observation information， keeping the state of particles with large weights unchanged, and particles with small weights tend to high likelihood region, and IPSO-PF algorithm was designed. Then， combining IPSO-PF algorithm with Paris formula, a fatigue crack growth prediction model based on Paris formula and IPSO-PF algorithm was constructed. Finally, the validity of the model was verified by using the open 2024-T351 aluminum alloy data set. The results show that compared with the traditional PF algorithm, IPSO-PF algorithm can improve the diversity of particles. The prediction error of the crack growth prediction model based on IPSO-PF algorithm is 2.6%, which is better than 9.2% based on PF algorithm.