基于模糊神经网络的燃气轮机热力部件故障诊断研究

A Study on Fault Diagnosis for Thermal Components of Gas Turbine Based on Fussy Neural Network

针对燃气轮机热力部件故障,提出了基于模糊神经网络的故障检测和诊断方法。在利用模糊规则描述系统故障状态的基础上,通过建立故障诊断目标函数,利用误差反向梯度算法实时修正神经网络连接权值和阈值。仿真结果证明与传统BP神经网络相比,模糊神经网络在对燃气轮机热力部件故障的识别中,具有更高的准确率。

A method of fault detection and diagnosis for thermal components of gas turbine is proposed on the basis of the theory of fuzzy neural network. By using fuzzy rules to describe the status of system fault and establishing the object function for fault diagnosis, it uses error back gradient algorithm to realize real-time correction of the connection weight and threshold of the neural network. Simulation results show that, compared with traditional BP neural network, the fuzzy neural network is more accurate in the fault recognition of thermal components of the gas turbine.