A combined technique of Kalman filter, artificial neural network and fuzzy logic for gas turbines and signal fault isolation

The target of this paper is the performance-based diagnostics a gas turbine for automated early detection components malfunctions. proposes new combination multiple methodologies single and failures on two-spool engine. aim technique to combine strength each methodology provide high success rate with presence measurement A KF (Kalman Filter), ANN (Artificial Neural Network) FL (Fuzzy Logic) used in research order improve rate, increase flexibility number detected methods have more robust solution. Kalman filter has his noise treatment, artificial neural network simulation prediction reference deteriorated performance profile fuzzy logic categorization flexibility, which quantify classify failures. In area GT (Gas Turbine) diagnostics, issues utilization 2-spool industrial engine not been investigated extensively. This reports key contribution component brief results quantification classification rate. tested constant deterioration increasing random deterioration. For nominal 0.4%, particular, above 92.0%, while 95.1%. Moreover, speed data processing (1.7 s/sample) proves suitability online diagnostics.