基于最小二乘支持向量机建模的电液伺服系统故障检测方法

Fault Detection Approach for the Eletro-hydraulic Control System By LS-SVM-based Modeling

基于最小二乘支持向量机建模的方法,研究了电液伺服系统的故障检测问题。介绍了基于最小二乘支持向量机进行建模的基本原理,分析了电液伺服系统所存在的非线性和故障模式,给出了基于最小二乘支持向量机建模进行故障检测的方法,试验结果表明,由支持向量机模型预测输出与实际输出相比较所形成的残差,能够准确地反映故障发生与否的情况;同时,与神经网络方法和普通的支持向量机方法相比,最小二乘支持向量机方法更适合工程应用,效果更好。

The fault detection based on least squares support vector machines （ KS - SVM） modeling for an eletro - hydraulic control system was studied. The basic theory and process of modeling based on LS - SVM was introduced. The non - linear and fault model of an eletro - hydraulic control system was analysed. The fault detection approach based on LS - SVM modelling was given. The result shows that the residual can show whether the fault occurs accurately, and the method is fit to the engineering, has better effect than the neural network and common SVM.