基于遗传算法的故障特征优化研究  

Research of Fault Feature Optimization Based on Genetic Algorithm

作　　者：刘运[1] 黎明[1] 董华[1] 

机构地区：[1]南昌航空大学,江西南昌330063

出　　处：《计算机仿真》2010年第4期192-196,共5页Computer Simulation

基　　金：国家自然科学基金(60475002);江西省教育厅科学技术研究项目(GJJ08209)

摘　　要：故障原始样本特征集中往往存在着冗余和不相关的分量。为了提高故障诊断系统的快速性和准确性,有必要采取适当的方法对故障诊断系统进行特征优化。为此提出了一种基于遗传算法的故障特征优化方法。方法对基本遗传算法的种群初始化方法,并对运算子进行了改进,同时利用支持向量机分类识别率构造遗传算法的适应度函数。通过对汽轮机减速箱轴承故障样本数据特征优化实例分析,结果表明方法能够得到更有利于分类的特征子集,使得故障诊断的准确率和计算效率都得到了改善,为关键设备状态监控与故障诊断提供了更为可靠的依据。Many redundant and unrelated features are included in the array of the original fault features. In order to increase the rapidity and validity of the malfunction diagnosis system, an appropriate approach must be taken for feature optimization in the fault diagnosis system. In this paper, a new approach based on genetic algorithm is pro- posed for it. The method of initializing the population and the operators have been improved, meanwhile the classifi- cation accuracy of the SVM is also utilized to format the fitness function of the genetic algorithm. The improved genet- ic algorithm is applied to the example of the turbine gearbox bearing for feature optimization simulation. It is proved that this method can obtain the subset of the features which contribute to pattern classification. With the result that fault diagnosis accuracy and computational efficiency have been improved. It lays a sound foundation for key equip- ment condition detection and fault diagnosis.

关 键 词：故障诊断 特征优化 遗传算法 支持向量机 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置]