Improved machine learning algorithm and its application in CPR prediction

Crack propagation rate (CPR) is important parameter in material speciality, establishing fatigue crack propagation rate is the key to forecasting structure fatigue lifetime, many approximate fitting models are used to calculate CPR, nine parameters fatigue crack propagation rate model and McEvily model are typical representatives, they are widely applied at present, but it is very complex to realize these models, partial derivative must be calculated and there is large deviation between fitted static parameter and actual value and physical conception isn´t clear. In this paper, in accordance with the disadvantage above methods, we presented optimal common machine learning algorithm (least squares support vector machine) method for fatigue crack propagation rate forecast, Complicated and strong nonlinear fatigue crack propagation rate curve was simulated by network design and conformation of least squares support vector machine learning algorithm and the optimized SVM parameters were selected. Two prediction examples of material A and material B are simulated to validate its efficiency, the experiment shows improved SVM had excellent ability of nonlinear modeling and generalization, the mean relative error is 0.1074% by calculating, it provided an economical, practical and reliable approach for material fatigue design.