基于改进支持向量机的变压器故障诊断

Transformer Fault Diagnosis based on Improvement of Support Vector Machines

针对支持向量机训练时间长、分类速度慢、故障诊断率不高等问题,引入模糊算法对支持向量机进行优化.先利用模糊C均值求得样本中心,再利用支持向量机中的二分类法对故障进行准确定位,达到诊断的目的.仿真结果表明,相比于支持向量机、BP神经网络和改良三比值法,改进后的支持向量机的故障诊断准确率最高.

To address problems with support vector machines, such as long time training, slow classification and low rate of fault diagnosis, a fuzzy algorithm is introduced. Sample center is first obtained by using the fuzzy C mean value, and then the two classification methods in a support vector machine are used to locate fault accurately for purpose of diagnosis. The simulation results show that compared with the support vector machine, BP neural network and modified three ratio method, the improved support vector machine has the highest accuracy rate of fault diagnosis.