An Intelligent Fatigue Life Prediction Method for Aluminum Welded Joints Based on Fatigue Characteristics Domain

摘要：

　　In order to reduce the dispersion level of fatigue test samples of aluminum alloy welded joints and further improve the fatigue life prediction accuracy,the concept of fatigue characteristic domain is proposed.Based on the attribute reduction of neighborhood rough set,the division method of fatigue characteristic domain is determined.According to the attribute reduction,the key fatigue life influencing factor set of welded joint is used to divide all fatigue test samples into several sub-domain sets,and each sub-domain set corresponds to a fatigue character sub-field.Support vector machine has the characteristics of simple calculation,perfect theoretical foundation and suitable for small sample problems.Based on its characteristics,an improved support vector machine model for fatigue life prediction of aluminum alloy welded joints is established in the determined fatigue characteristics domain.The experimental results show that compared with the least squares fitting method,the support vector machine model has stronger anti-noise ability and higher prediction accuracy.

关键词： Fatigue life Welded joints Fatigue characteristics domain Intelligent prediction

作者: Li Zou Hongxin Li Wei Jiang

作者单位: Software Institute,Dalian Jiaotong University,Dalian 116002,China;Sichuan Provincial Key Lab of Proc Software Institute,Dalian Jiaotong University,Dalian 116002,China

会议类型: 国际会议

会议名称: 2019机械设计国际会议暨第20届机械设计学术年会

会议地点: 浙江湖州

会议语种:英文

页码: 977-989

在线出版日期: 2019-08-12（万方平台首次上网日期，不代表论文的发表时间）