重型燃气轮机状态能观度分析及其在故障诊断中的应用  

States Observable Degree Analysis of a Heavy-duty Gas Turbine and Its Application in Fault Diagnosis

作　　者：蒲星星[1] 刘尚明[1] 蒋洪德[1] 

机构地区：[1]清华大学热科学与动力工程教育部重点实验室,北京市海淀区100084

出　　处：《中国电机工程学报》2012年第2期117-122,共6页Proceedings of the CSEE

基　　金：国家自然科学基金项目(60979014)~~

摘　　要：不同能观度的测量模型对重型燃气轮机故障诊断精度影响很大,对测量模型的优劣进行定量评价有重要意义。针对现有系统状态能观度分析存在的不足,提出一种新的基于能观矩阵奇异值分解的状态能观度分析方法,该方法在系统状态处于不同数量级时也可以准确计算系统各个状态的能观度。利用提出的能观度分析方法,计算得到了重型燃气轮机在设计工况下各状态的能观度和各测量对各状态能观度的敏感度排序。利用强跟踪滤波器对两种能观度测量模型下重型燃气轮机压气机故障诊断进行仿真,结果验证了文中方法的正确性,该方法可以用于测量模型优劣的定量评价。It has important significance for fault diagnosis accuracy of heavy-duty gas turbine to quantitatively evaluate the measurement model, different observable degrees of which would affect the fault diagnosis accuracy seriously. Considering the drawbacks of the current methods, a new method for observable degree analysis was proposed based on the singular value decomposition （SVD） of the observablity matrix. It can be used to analyze the observable degree for each state correctly even when the states of system have different order of magnitude. Based on the method, the observable degree for each state was obtained at the design point for heavy-duty gas turbine, and so the sensitivity order of each measurement to each state. Simulations of compressor fault diagnosis with the strong tracking filter （STF） were also conducted using two different observable degree measurement models, and the results prove the accuracy and the feasibility of the proposed method to quantitatively assess the merits of the measurement model.

关 键 词：能观度 奇异值分解 燃气轮机 故障诊断 强跟 踪滤波器 

分 类 号：TK39[动力工程及工程热物理—热能工程]