一种基于粗糙集理论的燃气轮发电机组故障诊断新方法

A Gas Turbogenerator Fault Diagnosis New Approach Based on Rough Set Theory

燃气轮发电机组的故障诊断 ,实质上是一个模式分类问题。本文以振动特征频谱为依据 ,提出了一种基于粗糙集理论的燃气轮发电机组故障诊断新方法。该方法不但可直接从完备的故障特征频谱样本集中导出正确的诊断结论 ,而且还能从不完整的故障特征频谱样本集中导出满意的诊断结论 ,它揭示了故障特征频谱信息的冗余性。本方法为在不完整征兆信息下的燃气轮发电机组故障诊断提供了新的思路。

Gas turbogenerator fault diagnosis can be treated as a pattern classification task.According to vibration characteristic spectrum,a gas turbogenerator fault diagnosis approach based on rough set theory is proposed.The accurate diagnosis results can be obtained directly from the set of complete fault spectrum samples,and the satisfactory diagnosis results can also be got from the set of incomplete fault spectrum samples by using the approach.The inherent redundancy in the spectrum information is revealed.The approach presents a new idea to gas turbogenerator fault diagnosis on the incomplete characteristics.Examples show that the proposed approach is very effective.