【摘要】 分析了故障诊断中基于符号表示的常规推理技术和基于神经连接的神经网络推理技术的不同特点并进行了优劣比较;详细考察了用于推理的模糊神经网络基本构造及逻辑基础;给出了基于纯模糊逻辑系统和Takagi Sugeno模糊逻辑系统的几种常用模糊神经网络推理模型;最后应用一种网络结合max min算法分析了一个故障诊断实例,并得出较好的结果。更多还原

【Abstract】 The paper discusses and compares two technologies of inference in fault diagnosis which are based on symbol and nerve connection. Basic structure and logic foundation of fuzzy neural network are inspected in details which can be used in inference. Several inference models of fuzzy neural network are given based on the pure and TakagiSugeno fuzzy logical systems. In the end, a neural network combined with maxmin algorithm is applied to a fault diagnosis instance and better result is reached.更多还原