【作者】 廖犬发；

【导师】 刘会金；

【作者基本信息】 武汉大学 ， 电力系统及其自动化， 2004， 硕士

【摘要】 配电网故障区间诊断，故障隔离是配电网自动化的重要方面，当配电网发生故障时要求能够根据配电网的实际拓扑结构，快速地定位故障区间并且通过遥控开关隔离与故障点相连的用户，然后通过转换联络开关和分段开关的状态，恢复对完好区域的供电。 传统的配电网故障区间诊断算法采用统一矩阵法，在实时信息序列中存在畸变信息时有可能错判和误判。本文提出基于改进BP网络的配电网故障区间诊断算法，当实时信息中出现畸变信息时，能自动纠错和进行准确定位，并具有高度的容错性能。 本文对BP网络的理论进行了较为详细的阐述，对于BP网络的层数，隐含层单元数目，初始权值，初始学习率及目标误差的选取进行了分析探讨，对配电网进行了分析，建立了配电网的故障区间诊断的模型，并在对配电网络故障区间诊断问题上进行了运用和仿真分析，证明了该方法具有一定的学习能力和容错能力。 本文用MATLAB作为仿真工具，对故障诊断的改进BP算法进行了实例仿真，分析和算例的结果表明该改进算法是有效的。更多还原

【Abstract】 Fault Section Detection in distribution network is an important aspect of distribution automation. When there is a fault in the distribution system, it can fast locate the fault and separate the load which connect the fault by operating switches according to thereal-time structure of the distribution network. Then it can find an operating pattern by changing the status of the tie or sectionalizing switches.The fault sections location in distribution networks based the unified matrix algorithm may result in the false results under the information mutation mode; The method of fault section diagnosis of distribution networks based on new neural network is presented in this paper. The method has high fault-tolerance performance.This paper comprehensively states the principle and conception of ANN, the author makes a deeper research on such aspects about BP ANN as its algorithmic theory,constructure, initial value, and a model about Fault Section Detection in distribution networks based on the analysis of the distribution network is proposed. This method has high fault-tolerance performance and the test results have verified its correctness.The author studies the case of BP ANN applying in engineering of Fault Section Detection in Distribution network.The simulation system is constituted with MATLAB. The results of the analysis and simulation denote that the method proposed in this paper is efficient and reliable.更多还原