基于相似性建模和模糊概率有向图的热力设备故障诊断

Thermodynamic equipment fault diagnosis based on similarity modeling and fuzzy probabilistic signed directed graph

在发电机组设备故障诊断中,相似性建模(SBM)通过设备的历史数据计算得到设备的运行特性,可以对设备进行实时监测;模糊概率有向图(FPSDG)通过建立完备的单元库和概率表实现故障诊断,可以对设备故障进行准确定位。本文利用SBM和FPSDG的优点,将两者融合得到基于SBM-FPSDG的设备故障监测诊断方法。以高压加热器为研究对象,利用其历史运行数据建立SBM模型,根据设备运行特性建立了故障单元库和概率表,可以实现高压加热器监测及故障诊断。对国内某600 MW机组1号高压加热器进行仿真计算,计算结果表明:该方法能够准确地估计高压加热器参数,具备良好的实时监测能力,能在设备故障发生早期给出报警并准确定位;该方法计算简单,适合于工程应用。

During the fault diagnosis of generator units,the similarity modeling method(SBM)can be used to obtain the equipment operation characteristics through calculating the historical data,which can monitor the equipment in real-time.The fuzzy probability signed directed graph(FPSDG)can be used for fault diagnosis of equipment by establishing a complete cell library and probability tables,which can accurately locate the equipment faults.Combing the advantages of SBM and FPSDG,the equipment monitoring and fault diagnosis method based on SBM-FPSDG is obtained in this study.Take the high pressure(HP)heater as the research object,the SBM model is established based on its historical operating data,and the fault cell library and probability tables are obtained according to the operating characteristics of the equipment.Then the HP heater monitoring and fault diagnosis can be realized.This method is used to calculate and simulate the No.1 HP heater of a 600 MW domestic unit.The results show that this method can accurately estimate the HP heater parameters,which has great real-time monitoring ability.Moreover,by this method,the fault can be alarmed and accurately located in the early stage of equipment failure.In short,this fault diagnosis method is simple and suitable for engineering applications.