A Method to Improve the Robustness of Gas Turbine Gas-Path Fault Diagnosis Against Sensor Faults

Abstract

Gas-path analysis (GPA) method has been widespreadly used to monitor gas turbine engine health status, and has become one of the key techniques in favor of condition-oriented maintenance strategy. GPA method (especially nonlinear GPA) can easily obtain the magnitudes of the gas-path component faults. Usually, it is essential to use correct measurement information to obtain correct fault signature for producing accurate gas-path diagnostic results. However gas-path components as well as sensors may degrade or even fail during gas turbine operations. The degraded sensors may produce significant measurement biases, which do not follow the Gaussian distribution, and misleading diagnostic results may be obtained. In order to solve this problem, a method to improve the robustness of gas turbine gas-path fault diagnosis against sensor faults was proposed for the typical nonlinear GPA method. The proposed method includes two steps: first, to locate suspicious degraded sensors based on Gaussian data reconciliation principle for all the gas-path measurements and second, to detect, isolate, and quantify the degradation rate of major gas-path components based on an extended nonlinear GPA method. The proposed method can effectively and accurately detect and isolate degraded gas-path components as well as sensors, and further quantify the component degradations.