基于粒子群算法和支持向量机的故障诊断研究

Research on Fault Diagnosis Based on PSO and SVM

支持向量机是采用结构风险最小化原则代替传统统计学中的基于大样本的经验风险最小化原则的新型机器学习方法,具有出色的学习分类能力和推广能力,广泛地应用于模式识别和函数拟合中;支持向量机中核函数的参数选择非常重要,它决定着故障诊断的精确度;为了提高电气设备故障诊断的精度和效率,将粒子群优化算法和最小二乘支持向量机相结合,提出了一种基于粒子群支持向量机的故障诊断方法,能够实现对核函数的σ参数进行快速动态选取,提高故障诊断的准确率和效率;实验表明,该方法能够有效地找出合适的核参数,并能取得较好的分类效果。

Support Vector Machine （SVM） have excellent learning, classification ability and generalization ability, which use structural risk minimization instead of traditional empirical risk minimization based on large sample. SVM is widely used in pattern recognition and function fitting. Kernel parameter selection is very important and decide the fault diagnosis precision. In order to enhance fault diagnosis preei sion, a new fault diagnosis method is proposed by combining PSO and LSSVM algorithm. It is presented to Choose a parameter of kernel function on dynamic, which enhances preciseness rate of fault diagnosis and efficiency. The experiments show that the algorithm can efficiently find the suitable SVM parameters, which result in good classification purpose.