多层次信息融合技术在燃气轮机故障诊断中的应用

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.

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

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.