基于人工神经网络和专家系统结合运用的电网故障诊断技术的研究

人工神经网络（ＡＮＮ）和专家系统（ＥＳ）的结合是智能系统研究的发展方向。本文对ＡＮＮ和ＥＳ结合用于电力系统故障诊断问题进行了研究，提出一种新的电力系统故障诊断方法。该方法综合了ＥＳ和ＡＮＮ各自的优点，充分利用ＥＳ的推理能力和ＡＮＮ的学习能力，由ＥＳ根据故障报警信息搜索电网数据库，提出故障假设，再由ＡＮＮ对故障假设进行诊断。该方法教授给ＡＮＮ的知识是电网通用的故障判断知识，具有较强的能用性。研究表明ＡＮＮ和ＥＳ结合用于电网故障诊断是非常有效的。

STUDIES OF INTEGRATING ARTIFICIAL NEURAL NETWORKS WITH EXPERT SYSTEMS FOR FAULT DIAGNOSIS IN POWER SYSTEM

This paper studies the integration of artificial neural networks (ANN ) and expert systems (ES) for fault diagnosis in power systemes. A new approach for fault diagnosis is pwoposed in this paper. This method integrates advantages of ANN and ES, in which the ES is employed for searching the general database of power system and presenting fault hypothses. and the ANN is employed for verifying the fault hypothses, The training of the ANN is based general knowledges of fault diagnosis, therefore, the ANN is general. The simulating test proves this method can effectively distinguish fault elements and is robust.