基于神经网络及模糊理论建立回热系统故障诊断层次模型

Fuzzy Logic and Artificial Neural Network Applied in Evaluating Malfunction Model of Thermodynamic System

针对热力系统故障发生的模糊性，提出将模糊辨识理论和人工神经网络相结合，建立回热系统故障诊断层次模型．该模型根据回热系统以加热器为中心的特点，分层诊断，缩小了知识库，减少了计算量，并利用了神经网络“边学习、边辨识”的优点，使故障诊断迅速、准确．经仿真实验，证明了模型辨识效果良好．

Combining the fuzzy logic and artificial neural networks, the approach of establishing fault diagnosis model of regenerative system has been put forward. With the heater being the center of the regenerative system, the fault diagnosis model based on this new approach makes fully use of the advantage of artificial neural networks that the model can be learning while it is identifying, reduces the calculation workload and the size of knowledge repository, which makes the fault diagnosis rapid and accurate. The simulation result shows the model satisfactory.