Fatigue Life Prediction Based on Data Driven Reduced Order Model
In order to solve the problem of effective prediction of aircraft structural fatigue life,a-chieve condition-based maintenance and improve the safety and reliability of aircraft,this paper takes the real key structural component of an aircraft——the wing-body fusion section as the specific re-search object and conducts in-depth research on the basis of fatigue life related theories,the crack propagation data at the hole edge of the section edge is obtained through simulation calculations.Ai-ming at the problems of complex fatigue damage modeling and weak generalization of the wing box section components,a feedback-based model is proposed.LSTM's fatigue life prediction method takes load spectrum,initial state and material parameters as input,and crack length.residual strength and service life as output to predict the entire process of fatigue life.Finally,a multi-dimensional e-valuation index is constructed and the rationality and superiority of the model described in this topic are verified in view of regression prediction accuracy,model generalization and noise robustness.
万方数据
维普
