Ensemble Learning Algorithms For Fatigue Crack Propagation And Life Prediction

The connection of porous cracks in the aircraft skin structure will cause damage to the structure,which belongs to the multiple part damage problem.At present,the porous mechanism research method is mostly used to solve this problem.The research on predicting the crack growth length between three holes based on the law of cracks between two holes provides a new idea for studying this kind of problem.Since the mechanism of fatigue crack growth is not clear,data-driven method has become an important research direction.Based on the data-driven method,the intelligent algorithm model of aviation fatigue crack growth is established,and the crack growth length between three holes is predicted based on the rule of cracks between two holes.The main work is divided into three parts:The first part is to carry out MTS fatigue crack growth test and data acquisition.At present,the research on multiple site damages of aviation aluminum alloy plate is still a key and difficult problem in this field.Therefore,MTS fatigue test machine is used to test the fatigue crack growth of aluminum alloy plate with two or three holes,and the crack length growth data under different test conditions are obtained.For the original crack growth test data obtained,the standardized data suitable for machine learning algorithm processing is obtained by using the standardization method in data preprocessing.The second part is the prediction of crack growth length by support vector regression(SVR).At the same time,the grid search algorithm is used to optimize the parameters of SVR.The established SVR model for predicting crack length has good prediction effect.Through the SVR model,the crack law between two holes can be predicted with the same load form.This SVR aviation crack growth prediction model can not only predict the crack length through the number of cycles,but also predict the number of cycles through the crack length,and then predict the remaining service life of the aviation aluminum alloy plate.The last part is the research of Heterogeneous Boosting Regression(HBoost.R)algorithm.The Heterogeneous Boosting(HBoost)algorithm is applied to the regression problem.A regression entropy measure is proposed,and HBoost.R algorithm is given.And the HBoost.R crack length prediction model with good prediction effect is established.Under the same load pattern,the crack growth prediction model of HBoost.R aviation aluminum alloy can be used to predict the three-hole inter hole crack rule according to the rule of two-hole inter hole crack.Figure 47;Table 12;Reference 56...