基于混沌粒子群优化支持向量机的变压器故障诊断

Fault Diagnosis of Transformers Based on Support Vector Machine with Improved Particle Swarm Optimization

为了提高电力变压器故障诊断的准确性,也要克服人工神经网络(ANN)中存在的收敛速度慢、容易陷入局部极值等缺陷,提出一种混沌的粒子群优化支持向量机的变压器诊断方法,该方法不仅具有很强的全局搜索能力,而且适用于支持向量机(SVM)参数优化,提高算法的鲁棒性.首先利用混沌的粒子群算法优化支持向量机的参数,把气体的特征参数代入优化的支持向量机分类模型中进行诊断,能够准确地分类变压器故障,从而达到故障诊断的目的.实验结果与常规方法比较,该方法能简单有效,诊断速度快,诊断正确率高.

In order to improve the accuracy of fault diagnosis of power transformer and solve the problem of local optimal solution and low convergence rate existed in artificial neural networks(ANN),we presented a method based on support vector machine(SVM)with chaos particle swarm optimization(CPSO)for fault diagnosis of transformer.The method not only has strong global search capability and robustness,but also is very easy to implement.First,the chaos particle swarm optimization is suitable to determine free parameters of SVM.Second,the CPSO-SVM classifiers with the dissolved gas analysis can achieve diagnostic accuracy.Finally,fault diagnosis of power transformer was realized.The experimental results indicate that the method is simpler,faster and more accurate compared with the traditional algorithm.