基于人工神经网络的定子绕组堵塞故障诊断新方法

New Method Based on Artifical Neural Network for on - Line Diagnosis of Cooling Circuit Blockage Fault on Turbogenerator Stator

本文提出了基于人工神经网络(Artificial Neural Network简称ANN)的定子绕组堵塞故障诊断新方法,首先分析堵塞故障的特征,计算各种故障时的温度分布,使定子绕组各测点的温升和不同故障之间,建立一一对应关系的数学模型,然后构造神经网络,从而提高故障诊断的准确性。

This paper present the diagnosis method of armature winding blockage with the ANN(BP). First calculating the temperature distributing of fault in the stator, the one by one parallelism connection is set up between the fault and the temperature distributing , then the diagnosis rules is set up with drilling the ANN. The method is able to find the fault wider and earlier than the temperature patrol measure instrument.