小波模糊神经网络应用于配电网输电线的故障测距

在小电流接地系统单相接地故障特征分析的基础上，提出了一种基于故障后稳态及暂态电气量的小波模糊神经网络的故障测距方法。单相接地故障时的暂态分量故障特征非常明显，且故障暂态高频分量受故障前负荷的影响较少，故可以采用故障暂态分量描述故障模式特征并进行故障定位，鉴于已有的小波神经网络模型不适合于故障测距，作者从广义的小波神经网络概念出发，结合模糊控制理论，提出了适合于电力系统故障暂态和稳态信号分析的小波模糊神经网络方法，并将该方法应用于小电流接地系统直配输电线路的故障测距。理论分析及大量的EMTP仿真结果表明：本文所提出的小波模糊神经网络理论，模型及算法具有较好的故障测距性能，并可应用于电力系统的故障分析。

APPLICATION OF WAVELET FUZZY NERUAL NETWORK TO FAULT LOCATION OF SINGLE LINE GROUNDING IN DISTRIBUTION LINES

On the basis of analysis on the characteristics of single phase grounding fault occurred in small current neutral grounding system, a fault location method using wavelet fuzzy neural network (WFNN) based on post-fault transient and steady state electrical quantities is put forward. When single line grounding (SLG) faults occur in distribution lines, the fault characteristics of transient components are very obvious and the transient high frequency components are less influenced by pre-fault load, therefore, it is possible to depict the features and characteristics of fault modes by transient components occurred under faults and to locate the faults. Because the existing wavelet neural network model is not suitable to fault location, thus, from the concept of generalized wavelet neural network and integrating with fuzzy control theory, a wavelet fuzzy neural network method suitable for the analysis of the steady state signals in power networks and the transient ones under the faults is established and applied to the fault location ofdistribution lines in small current neutral grounding system. Theoretical analysis and the results from a large amount of simulations by EMTP show that the proposed theory of wavelet fuzzy neural network, its algorithm and models possess better performance of fault location, besides, this method can also be applied to power system fault analysis.