基于BP网络的故障诊断方法及其在电站中的应用

电站机组可能会发生各种故障，有些故障没有明显征兆。为能诊断这类故障，需要新的方法。提出了根据当前技术条件下可以测量到的参数，而不一定是故障特征参数，进行故障诊断的思路。介绍了BP人工神经网络的结构和学习方法，提出了基于BP网络的模式识别能力，建立电站性能监测与诊断系统的新思路和方法。利用该方法，对电站设备性能下降故障的程度进行了成功诊断。图2表2参3

Fault Diagnosis Method and Its Application in Power Station Based on BP Network

There are different kinds of fault in a power station unit. Some fault has no obvious sign. To diagnose this kind of fault, new method is needed. A diagnosis way is provided, that is based on using the parameters detected by measuring elements but not characteristic sign. The structure and learning method of back propagation artificial neural network are introduced. A method of fault diagnosis is provided, which is based on the ability in pattern recognition of the back propagation artificial neural network. It can provide a new way for the performance monitoring and diagnosing in power station. By using this method the degree of the performance deterioration of the equipment can be successfully diagnosed. Figs 2, tables 2 and refs 3.