Abstract: This paper presents a knowledge base which is constructed on the basis of fuzzy neural network. It chooses the five-layer network: at first fuzzing the inputs, then using the RBF network to learn, finally the output is fuzzed in reverse. The fuzzy RBF network matches the on-line fault diagnosis.

摘要： 提出了一种基于模糊RBF神经网络建立故障诊断专家系统知识库的新方法,采用5层神经网络,先对输入量进行模糊化处理,然后对RBF神经网络进行学习,最后进行反模糊化处理。该模型非常适合复杂的异常炉况系统在线故障诊断。

关键词: 专家系统, 异常炉况, 径向基函数, 模糊神经网络

CLC Number: 

• TP182