遗传算法在故障特征选择中的应用研究  

Application of Genetic Algorithm to Fault Feature Selection

作　　者：史东锋[1] 屈梁生[1] 

机构地区：[1]西安交通大学机械诊断与控制学研究所,西安710049

出　　处：《振动．测试与诊断》2000年第3期171-176,共6页Journal of Vibration,Measurement & Diagnosis

基　　金：中国博士后基金资助项目! (编号 :1 999-1 7)

摘　　要：针对传统组合优化方法用于特征选择的种种缺陷 ,提出了基于遗传算法的特征选择策略。经仿真及回转机械故障实例分析 ,证明该方法拥有较强的并行性和寻优能力 。Feature selection is an important step in fault diagnosis owing to its considerable reduction in redundant features to enhance the effect of diagnosis. In view of the deficiencies of traditional combination optimization method, a new method based on genetic algorithm is proposed in the paper. According to the results of computer simulation and practical diagnosis example, it is proved that this method possess excellent parallelism and optimization property.

关 键 词：特征选择 遗传算法 故障诊断 回转机械 仿真 

分 类 号：TH133[机械工程—机械制造及自动化] TP391.9[自动化与计算机技术—计算机应用技术]