基于小波奇异熵和SOM神经网络的微电网系统故障诊断

针对微电网系统运行方式灵活、拓扑结构多样的特点,基于对小波变换、奇异值分解和泛化信息熵基本理论的分析,揭示了小波奇异熵能够对故障信号给出确定的量度,将小波奇异熵与自组织特征映射(self-organizing feature map, SOM)神经网络相结合,提出一种能够适应微电网系统拓扑结构变化情况的故障诊断方法。 利用PSCAD4.2建立了微电网故障仿真系统,进行故障诊断仿真试验。 试验结果表明:该方法不受故障位置、故障时刻等因素的影响,在微电网系统拓扑结构发生变化的情况下,能实现有效的故障诊断。

Fault diagnostic method for micro-grid based on wavelet singularity entropy and SOM neural network

According to the diversity of micro grids topology, through analyzing the theories of wavelet transform, singular value decomposition and extended shannon-entropy, the wavelet singular entropy could measure the fault signal. A fault diagnosis method for the micro grid system was proposed by integrating the wavelet singular entropy with the self organizing feature map(SOM)neural network. A micro grid fault simulation system was established by PSCAD4.2. The simulation results proved that the proposed diagnosis method was insensitive to the location and the time fault occurs, which had strong adaptability to the variation in structure topology.