基于改进灰狼算法与最小二乘支持向量机耦合的电力变压器故障诊断方法

Fault Diagnosis for Power Transformers Based on Improved Grey Wolf Algorithm Coupled With Least Squares Support Vector Machine

电力变压器运行故障的准确诊断有利于提高变电设备状态检修和电网安全运行水平,为实现故障的准确分类,文章以油中溶解的5种典型气体作为故障诊断的特征量,提出一种基于改进灰狼算法与最小二乘支持向量机耦合的电力变压器故障诊断方法。该方法通过改进灰狼算法寻求最小二乘支持向量机中的最优惩罚系数C和核函数参数g,用以提高故障诊断的准确率。首先阐明最小二乘支持向量机和灰狼算法的改进点并将二者耦合,将其代入413组电力变压器的油中溶解气体检测数据来诊断故障类型,与其他诊断方法进行对比;其次研究惩罚系数C和核函数参数g对电力变压器故障类型识别准确率的影响规律;最后借助训练后的改进灰狼算法与最小二乘支持向量机耦合方法,通过两台不同电压等级的变压器故障实例分析,验证了故障诊断方法的有效性。研究结果表明:相较于单一使用最小二乘支持向量机和传统灰狼算法与最小二乘支持向量机耦合,改进灰狼算法与最小二乘支持向量机耦合方法对电力变压器故障诊断的准确率分别提高了14%和7%。此外,惩罚系数C和核函数参数g对电力变压器故障类型识别准确率的影响呈现非线性规律,凸显了通过智能算法找到最优解的便捷性、必要性、有效性。

In order to achieve the accurate fault classification,the article proposes a fault diagnosis method for power transformers based on the improved grey wolf algorithm coupled with the least squares support vector machine by using five typical gases dissolved in the oil as the characteristic quantity for the fault diagnosis.This method seeks the optimal penalty coefficient C and the kernel function parameter g in the least squares support vector machine through the improved grey wolf algorithm to improve the fault diagnosis accuracy.Firstly,the improvement points of the least squares support vector machine and the grey wolf algorithm are elucidated and coupled.They are put into the 413 sets of dissolved gas in oil detection data of the power transformers are substituted to diagnose the fault types and compared with the other diagnostic methods;secondly,the influence law of the penalty coefficient C and the kernel function parameter g on the accuracy of the fault type identification of the power transformers is investigated;finally,with the help of the coupling of the trained improved grey wolf algorithm and the least squares support vector machine,the effectiveness of the fault diagnosis method is verified through the analysis of two transformer fault examples under different voltage levels.The results of the study show that,compared with the least squares support vector machine and the traditional grey wolf algorithm coupled with the least squares support vector machine,the accuracy of the improved grey wolf algorithm coupled with the least squares support vector machine method for the fault diagnosis of the power transformers is increased by 14%and 7%respectively.Besides,the influence of the penalty coefficient C and the kernel function parameter g on the accuracy of the fault type identification of the power transformers shows a non-linear pattern,highlighting the convenience,necessity and effectiveness of finding the optimal solution through the intelligent algorithms.