模糊BP神经网络及其在故障诊断中的应用

Fuzzy BP Neural Network and Its Applications in Fault Diagnosis

概括总结了传统故障诊断系统的缺陷 ,介绍了模糊神经网络技术的优点及其在故障诊断中的优势 ,简述了采用模糊BP神经网络解决故障诊断的关键问题 ,包括构建模糊BP神经网络 ,以及生成推理和解释机制的基本方法。以空调系统故障为例 ,描述了该系统的故障诊断模型、各输入量的隶属函数、输出的判定依据 ,以及依照BP算法建立的知识库。最后概括分析了模糊BP神经网络的优点 ,指出了其在技术实现中存在的主要问题。

In this paper, we summarize the disadvantages of the traditional fault diagnosis systems, introduce the advantages of fuzzy neural network, and its superiority in fault diagnosis. The key problems of fuzzy BP, including principles of establishing the fuzzy BP model, rules of inference and interpretation are presented. Taking the example of the air condition system, we describe the fault diagnosis model, fuzzy sets, BP algorithm and the knowledge base. Finally, the merits and main problems of furry BP are pointed out.