基于WSE-RVM的柔性多端直流输电换流器故障诊断

Fault Diagnosis of Converter in Flexible Multi Terminal DC Transmission Based on WSE-RVM

为了快速、准确地对柔性多端直流输电系统(VSC-MTDC)中电压源换流器(VSC)的故障进行诊断,提出了一种基于小波奇异熵(WSE)理论和相关向量机(RVM)的诊断方法。首先,在PSCAD/EMTDC中搭建了三端直流输电系统的仿真模型,并对电压源换流器的几种常见故障进行了仿真分析。其次,以VSC-MTDC直流侧电压作为故障信号源,利用小波奇异熵理论对仿真数据进行故障特征提取,将得到的故障特征向量作为相关向量机的输入样本。最后,利用相关向量机对换流器故障进行诊断,结果表明,小波奇异熵理论可以较好地提取电压源换流器的故障特征,相关向量机比同等条件下的支持向量机具有更高的诊断准确率和更短的诊断时间。

In order to diagnose the fault of voltage source converter （VSC） in VSC- MTDC more quickly and accurately, a fault diagnosis method based on wavelet singular entropy（WSE） theory and relevance vector machine （RVM） is proposed.Firstly, based on PSCAD/EMTDC, a simulation model of three-terminal DC transmission system is built, and several common faults of VSC are simulated and analyzed.Secondly, the DC side voltage of VSC -MTDC is used as the fault signal source, using the wavelet singularity entropy theory to extract the fault features of the simulation data. The obtained fault feature vector is used as the input sample of RVM.Finally, RVM is used to diagnose the fault of VSC.The final findings show that, the wavelet singular entropy theory can be well used to extract the fault features, and RVM has higher diagnostic accuracy and shorter diagnosis time than SVM under the same conditions.