采用模糊神经网络进行凝汽器的故障诊断。根据某电厂的运行规程及运行经验构造了进行凝汽器故障诊断所需的全部 17个隶属函数 ,利用这些隶属函数可以对各个输入进行模糊化处理 ,使之成为量化输入。对模糊 BP网络进行训练 ,可以得到模糊 BP网络的知识库结构。在此基础上 ,对一个凝汽器实际故障进行了诊断 ,得出了令人满意的结果 Using fuzzy neural network to diagnose the fault of condenser. According to the operating rules and operation experience of a power plant, all the 17 membership functions needed to diagnose the condenser fault are constructed. These membership functions can be used to blur each input and make it a quantitative input. The training of fuzzy BP network can get the fuzzy BP network knowledge base structure. On this basis, the actual fault of a condenser was diagnosed, and satisfactory results were obtained