电力系统及其自动化学报2017，Vol.29Issue(2)：39-44,6.DOI:10.3969/j.issn.1003-8930.2017.02.007
小波能量偏度神经网络的HVDC换相失败故障诊断
Fault Diagnosis of Commutation Failures in HVDC System Based on Neural Network of Wavelet Energy Skewness
摘要
Abstract
Commutation failure is a common fault in HVDC systems. In order to effectively diagnose the causes of com-mutation failures,a new fault diagnosis method of commutation failures in the HVDC system is proposed based on the neural network of wavelet energy skewness. After the 15-layer wavelet decomposition of acquired DC voltage fault signal in the inverter side is performed,the wavelet coefficients on every scale are obtained and the wavelet energy skewness is extracted to construct the eigenvector of wavelet energy skewness. With this eigenvector as fault sample ,three-layer BP neutral network is trained to implement the fault diagnosis of commutation failures. A ± 800 kV UHVDC system is taken as an example,and a variety of causes inducing failure fault are simulated and analyzed. This method is used to conduct wavelet analysis,fault feature extraction and BP neural network training. At last,the unknown fault is identi-fied. The results show that the method can accurately diagnose the fault cause of commutation failures.关键词
换相失败/小波能量偏度/神经网络/故障诊断Key words
commutation failures/wavelet energy skewness/neural network/fault diagnosis分类
信息技术与安全科学引用本文复制引用
朱艳,王渝红,丁志林,宋梁,李兴源,邱大强..小波能量偏度神经网络的HVDC换相失败故障诊断[J].电力系统及其自动化学报,2017,29(2):39-44,6.基金项目
国家高技术研究发展计划(863计划)资助项目(2011AA05A119) （863计划）
四川省电力公司检修公司科技资助项目(SGSCJXOOYJKJ1301058) （SGSCJXOOYJKJ1301058）
