基于改进鱼群算法和支持向量机的变压器故障诊断

电力变压器在整个体系中处于十分重要的地位,部件的运行概况和整个电网的稳定性具有密切联系。对电力变压器的故障诊断,工程实践中广泛采用的是油中溶解气体法,由于变压器故障样本比较少,属于小样本数据,而支持向量机能够较好地解决小样本的多分类问题,因此提出利用改进鱼群算法对支持向量机寻优得到全局最优解,得到具有最佳参数的支持向量机模型。通过数据实例分析得出,改进鱼群算法故障诊断模型比粒子群算法故障诊断模型和改良三比值法分类准确率高。

Transformer Fault Diagnosis Based on Improved Fish Swarm Algorithm and Support Vector Machine

Power transformer holds a very important position in the whole system,for its operational status is closely related to the stability of the whole power grid.Oil dissolved gas method is widely applied in engineering practice for fault diagnosis of the power transformer.Since transformer fault samples with a small sample database are quite limited in number and the support vector machine (SVM) is a good solution of the multi-class problem of small samples,this paper proposes that the improved fish swarm algorithm should be used for SVM optimization to obtain a global optimal solution and a SVM model of optimal parameters.Analysis of data instances shows the fault diagnosis model of improved fish swarm algorithm can achieve a higher accuracy than the particle swarm optimization fault diagnosis method and the improved three-ratio approach.