MODELLING OF QSTE340TM STEEL LIFETIME UNDER CONSTANT AMPLITUDE LOADING BY EMPLOYING MACHINE LEARNING METHODS
Keywords:
machine learning, fatigue lifetime, random forest, decision trees, support vector machines
Abstract
In the article, there was modelled the fatigue life of QSTE340TM steel under constant-amplitude loading using supervised methods of machine learning. Modeling was carried out by the method of random forests, decision trees and the method of support vectors. Dependencies of predicted and actual crack length were obtained for four stress ratio R=0.1; 0.3; 0.5; 0.7. There were built the dependences of the predicted and experimental crack length a on the number of load cycles N. It was found that the best results were shown by random forests method and decision trees.
