基于Kohonen神经网络的燃气轮机故障诊断

Kohonen Neural Network-based Gas Turbine Fault Diagnosis

针对基于热力参数的燃气轮机8种典型常见故障,根据Kohonen神经网络诊断的工作原理、诊断特征,研究了基于Kohonen神经网络方法在燃气轮机故障诊断中的应用方法,得出了Kohonen模型具有自学习功能,运算速度快,类型识别能力强的优点,是一种适合于燃气轮机分类故障较好的具有特色的神经网络。

With respect to eight kinds of thermodynamic parameters-based gas turbine typical and common faults studied are the Kohonen neural network-based methods used for diagnosing gas turbine faults on the basis of diagnostic working principles and specific features of the Kohonen neural network. It has been found that the model of Kohonen network has the following merits: self-learning function, rapid operating speed and strong pattern- recognition ability. The Kohonen network is a relatively good neural network with characteristic features suitable for diagnosing various gas turbine faults.