多层次信息融合技术在燃气轮机故障诊断中的应用  

Fault diagnosis of gas turbine based on multi-level information fusion

作　　者：周密[1] 刘永葆[1,2] 黄树红[2] 贺星[1] 

机构地区：[1]海军工程大学船舶与动力学院,武汉430033 [2]华中科技大学能源与动力工程学院,武汉430074

出　　处：《海军工程大学学报》2009年第6期45-49,61,共6页Journal of Naval University of Engineering

基　　金：国家自然科学基金资助项目(50775083)

摘　　要：探讨了信息融合技术在燃气轮机故障诊断中的应用,构建出一种燃气轮机多层次信息融合故障诊断模型。在模型的数据融合层,采用了自适应加权融合算法,以最小化传感器测量不确定度为目标对同源测量数据进行了融合,其结果再经特征层进行故障特征的提取和融合,最终通过决策层给出诊断判定。数值实验结果证明,该故障模型能有效减小传感器测量不确定度对于诊断的不良影响,所得诊断结果的合理性和精度均得到了提高。Based on analyzing the application of the multi-level information fusion to the fault diagnosis of the gas turbine,a new multi-level information fusion model was established.The self-adapting weighting fusion algorithm was used at the first data fusion level,which could minimize the uncertainty in measuring gas turbine sensors.Then the data fusion results of the multi-senor were further determined at the second character fusion level.Finally the fault diagnosis results were obtained through the decision-making level.The results of the experiment prove that the poor effect of the measurement uncertainty on diagnosis can be effectively reduced so as to improve fusion results rationally and ideally.

关 键 词：燃气轮机 多层次信息融合 故障诊断 自适应加权融合 

分 类 号：TK472[动力工程及工程热物理—动力机械及工程]