基于VMD样本熵和KELM的输电线路故障诊断

针对输电线路短路故障危害大,故障识别率较低的情况,提出基于变分模态分解(VMD)样本熵与核极端学习机(KELM)相结合的输电线路故障诊断方法,提高输电线路故障诊断的正确率。首先,采用VMD对故障后的三相电压进行分解,得到一系列三相平稳的模态分量;其次分别计算每组各分量的样本熵值,作为输电线路故障提取特征,组成样本库;以提取的输电线路故障特征输入到核极端学习机进行训练,获取诊断模型,然后比较其与极限学习机(ELM)、支持向量机(SVM)和BP神经网络的诊断效果。仿真结果表明,VMD样本熵+KELM的输电线路故障诊断模型精度高于其他3种算法,且运算速率更快,噪声鲁棒性更好。

Fault diagnosis of transmission lines based on VMD sample entropy and KELM

The fault diagnosis method based on the variational mode decomposition ( VMD) sample entropy and kernel extreme learning machine ( KELM) is proposed to improve the accuracy of the transmission line fault diagnosis, aiming at the damage of transmission line short-circuit fault and the low recognition rate. Firstly, the three-phase voltage after the fault is decomposed by VMD and three groups of stationary modal components are obtained. Secondly, the entropy of each component of each component is calculated, the feature extraction of the transmission line fault is completed and the sample library is formed. Finally, the KELM is used to diagnose transmission line fault and compared with the extreme learning machine ( ELM), support vector machine ( SVM) and BP algorithm. The simulation results show that the transmission line fault diagnosis of VMD sample entropy and KELM is not affected by the fault time, fault type and transition resistance, and the correct rate is higher than the ELM , SVM and BP algorithms. It can better realize the fault diagnosis of transmission lines, have better generalization ability, and the running time is far less than the other three kinds of algorithms. Experiments show that the fault diagnosis model of VMD sample entropy and KELM is more accurate than the other three algorithms, with faster operation rate and better noise robustness.