汽轮发电机组振动故障的综合自动诊断方法研究

Study on the Comprehensive Automatic Fault Diagnosis Method for Turbine Generator Sets

利用振动信号频率成分结构和频率幅值研究了汽轮发电机组振动故障的综合自动诊断方法。根据振动故障的频谱特点 ,将振动故障分为 4个模式类 ,并确定故障类模式中心 ,在故障类的层次上做模式识别 ;利用频谱征兆和模糊关系矩阵 ,计算故障隶属度值以确定故障的严重程度 ;

Based on the discussion in advantages and disadvantages of two traditional automatic diagnosis methods using fuzzy theory or neural network, a comprehensive automatic diagnosis method, using two kinds of information of fault spectrum structure and fault spectrum element amplitude, is proposed to diagnose turbine generator sets faults. Firstly, according to the feature of vibration fault spectrum, four kinds of fault patterns are obtained for diagnosis of vibration faults at the level of fault class, rather than the concrete fault. Secondly, the fault extent is measured by the result acquired by spectrum symptoms and fuzzy relationship matrix, which shows the relation between vibration symptoms and faults. The new method is proved to be reliable and effective by two examples.