摘要
针对高压断路器故障识别中样本数据少,算法易陷入局部最优、故障识别准确率不高的问题,提出了一种带有引力搜索算子的烟花算法优化支持向量机的高压断路器故障识别新方法。首先对高压断路器机械振动信号进行VMD分解,得到一系列反映振动信号特性的固有模态分量,求取各分量的排列熵作为特征输入向量。其次采用万有引力算法改进烟花算法,通过粒子间的引力作用使位置较差的粒子与优秀粒子进行信息交互,产生位置信息改进后的粒子,通过多次迭代,得到支持向量机(SVM)的最优参数惩罚因子c和核函数g,构建出高精度的高压断路器故障识别模型。最后将提取的特征向量按照2∶1分为训练集和测试集,输入到优化后的SVM模型中进行故障识别。实验结果表明:VMD分解提取排列熵能有效获取振动信号的特征,优化后的SVM模型在正确率、精度等方面高于传统的SVM、灰狼算法、遗传算法、烟花算法等故障识别模型。该方法能够在样本数据较少时,有效提取故障的特征信息,故障识别准确率达到100%。
In view of such issues as indufficient sample data in high voltage circuti breaker fault identification,the tendency of algorithms to fall into the local optima and low fault identification accuracy,a novel fault identification method for high voltage circuti breaker,which optimizes support support vector machines using a fireworks algorithm with a gravitional search operator,is proposed.First,the mechanical vibration signal of the high-voltage circuit breaker is given VMD decomposition and a series of intrinsic mode components reflecting the characteristics of the vibration signal are obtained and the arrangement entropy of each component is taken as the characteristic input vector.Then,the gravitational search algorithm is used to improve the fireworks algorithm.Through the gravitational effect between particles,the particles with weak position interact with the excellent particles generate particles with improved position information.Through multiple iterations,the optimal parameter penalty factor c and the kernel function g of the support vector machine(SVM)are obtined and a high-precision high-voltage circuit breaker fault identification model is constructed.Finally,the extracted feature vector is divided into the training set and the test set according to 2∶1,is input into the optimized SVM diagnosis model for fault identification.The experimental results show that VMD decomposition and extracting permutation entropy can effectively obtain the characteristics of vibration signal,and the optimized SVM model is higher than the traditional SVM,gray wolf algorithm,genetic algorithm,fireworks algorithm and other fault identification models in terms of accuracy.This method can effectively extract the fault feature information with few sample data,and the fault identification accuracy can reach up to 100%.
作者
赵岩
党康佳
孙江山
ZHAO Yan;DANG Kangjia;SUN Jiangshan(School of Electrical and Control Engineering,Heilongjiang University of Science and Technology,Harbin 150022,China)
出处
《高压电器》
北大核心
2025年第12期17-24,共8页
High Voltage Apparatus
基金
国家自然科学基金资助项目(51677057)
黑龙江省省属高等学校基本科研业务费项目(2021-KYYWF-1476)。
关键词
高压断路器
故障识别
振动信号
排列熵
烟花算法
支持向量机
high voltage circuit breaker
fault identification
vibration signal
permutation entropy
the fireworks algorithm
SVM
