基于径向基核函数DAG-SVM的变压器故障诊断

Transformer Fault Diagnosis Based on Radial Basis Kernel Function DAG-SVM

本文将有向无环图(Directed Acyclic Graph,DAG)结构和支持向量机(Support Vector Machine,SVM)的分类能力相结合,提出一种基于径向基核函数DAG-SVM的变压器故障诊断方法。通过使用径向基核函数,DAG-SVM能够将非线性特征映射到高维空间,并在该空间中进行分类,从而更好地捕捉变压器故障的复杂模式和特征。数值计算结果表明,基于径向基核函数的故障诊断综合正确率为73.88%,均高于线性核函数、多项式核函数、S型核函数三种方法,所提基于径向基核函数DAG-SVM的变压器故障诊断模型具有较好的诊断效果。

This paper combines the structure of Directed Acyclic Graph(DAG)and the classification ability of Support Vector Machine(SVM)to propose a transformer fault diagnosis method based on radial basis kernel function DAG-SVM.By using the radial basis kernel function,the DAG-SVM is able to map the nonlinear features into a high-dimensional space and classify them in that space,thus better capturing the complex patterns and features of transformer faults.Numerical calculations show that the integrated correct rate of fault diagnosis based on radial basis kernel function is 73.88%,which is higher than that of the three methods of linear kernel function,polynomial kernel function,and S-type kernel function,and the proposed transformer fault diagnosis model based on radial basis kernel function DAG-SVM has better diagnostic effect.