A particle swarm optimization-based support vector machine strategy for fault diagnosis of transformer has been put forward in this paper aiming at the deficiency in traditional diagnosis based on DGA. The transformer fault diagnosis model is established with support vector machine classification, and the particle swarm optimization algo-rithm is adopted to realize parameter optimization. Support vector machine classification machine is trained on the MATLAB platform using libSVM toolbox, and the well-trained SVM will be adopted to diagnose the transformer fault condition for 110kV Lixing substation. Results show that, using the particle swarm optimization of transformer for fault diagnosis based on support vector machine agree with the reality. This method can improve the accuracy of fault diag-nosis of transformer.%针对基于DGA的变压器故障诊断方法在变压器故障诊断中存在的不足，提出了基于粒子群优化支持向量机的变压器故障诊断方法。建立支持向量机分类机的变压器故障诊断模型，并用粒子群算法优化参数，利用libSVM工具箱在MATLAB软件平台上训练支持向量机分类机，用训练良好的支持向量机诊断110 kV立星变电站变压器故障状况。结果证明，采用基于粒子群优化支持向量机的变压器故障诊断结果与实际相符。此方法能够提高变压器故障诊断的准确率。
展开▼
