摘要
Abstract
To solve the rupture life evaluation for steam turbine materials,a machine learning method based on support vector machine(SVM)is introduced,following by test data pretreatment and feature optimization.In this paper,by testing of various mechanical and durability on a large number of C422 materials,a sample durability correlation dataset was obtained.Through data preprocessing,introduction of feature dimensionality,feature and density analysis,a support vector classification model was established to partition the durability life of the material samples.The results show that the learned SVM model has accuracy,precision,and recall rates of 87.8%,88.9%,and 97.0%,respectively.Which is roughly 9.1%higher compared with the K-nearest neighbor algorithm accuracy.Furthermore,the SVM model provides a potential method for diagnosis of material rapture life.关键词
持久寿命评估/支持向量机/特征优化Key words
rupture life evaluation/support vector machine/feature optimization分类
金属材料
