Fault diagnosis of the gas turbine based on self-adapting weighting evidence fusion

The application of self-adapting weighting(SAW) evidence fusion algorithm in the fault diagnosis of the gas turbine is discussed and a new multi-level information fusion model is proposed. Then address to minimizing the senor measurement uncertainty(MU), the model firstly adopts the self-adapting weighting fusion algorithm for the congenetic data fusion. Then the fault evidence is calculated based on data fusion results, in addition, the evidence preference weight(EPW) are solved through the minimal measurement uncertainty. Finally the advanced Dempster-Shafer(D-S) evidence theory is proposed for overall fusion of fault evidence. The experiments of fault diagnosis for one gas turbine are carried out, which demonstrates that the model could effectively diagnose the gas faults of the gas turbine and avoids the vile effect of the measurement uncertainty in a great deal.