采用神经网络和专家系统的变电站故障诊断系统

介绍了采用ANN和ES的变电站故障诊断系统，充分利用ES的推理能力和ANN的学习能力。系统首先采用ES对故障报警信息进行预处理，再用ANN方法确定故障情况，最后利用ES评价保护和开关的动作情况。ANN采用RBF网络，提高了训练速度和诊断能力；训练样本包括用ES自动生成基本故障样本，以及无确定规则的特殊样本，提高了系统管理样本的能力。本系统故障诊断快速，动作评价可靠，可减轻运行人员的工作量。

A SUBSTATION FAULT DIAGNOSIS SYSTEM USING ANN AND EXPERT SYSTEM

This paper proposes a substation fault diagnosis system based on artificial neur al network (ANN) and expert system (ES).It diagnoses the fault using ANN and ana lyzes the operations of protectors and breakers using ES.ANN adopts RBF network that reduces training time and improves diagnosis ability.The training set is co mposed of two parts:ES automatically creates some and the others are special cas es.The system can diagnose the fault accurately which can relieve the burden of operators.