Research On Sensor Fault Diagnosis And Signal Reconstruction Of Gas Turbine Control System

The control system of gas turbine monitors the running state of gas turbine through sensor measurement signals.And the accuracy and reliability of the sensor measurement signals is the key factors that ensure the safe and stable operation of control system.In order to improve the stability,safety and reliability of the control system,it is of great significance to diagnose,isolate the sensor fault and reconstruct the fault signals.In this paper,a series of researches on fault diagnosis,isolation and fault signal reconstruction of sensors were carried out for a heavy-duty gas turbine.The main contents are as follows:Firstly,the basic principle of the CLM based on a heavy-duty gas turbine was studied.The modified Newton-Raphson algorithm was proposed to improve the computational efficiency of the nonlinear equations and ensure the convergence of the iterative results.On the basis of these,the corresponding SVM was obtained by least squares method.The dynamic simulation results proved the precision of the fitted linear model was high.Secondly,based on the SVM of gas turbine,a sensor fault diagnosis system based on a bank of Kalman filters was designed for single fault and multiple faults in the control system.The optimal value of the cost function was obtained by PSO algorithm to determine the best fault detection threshold.And simulation test verified the feasibility and effectiveness of the fault diagnostic system.The effect of sensor fault degree and measurement noise intensity on the diagnostic system were further analyzed,and it was found that there is a minimum fault which can be detected in the diagnosis system,usually an integer multiple of the standard deviation of measurement noise.Finally,by analyzing the correlation between sensor measurement parameters,the fault signal reconstruction method based on multivariate linear regression was studied.Simulation test showed that the reconstruction accuracy of this method was high in the case of a few sensors failed simultaneously.However,the method has been confronted with certain limitations when the number of faulty sensors increased.Under the circumstance,a fault signal reconstruction method based on KPCA,which was suitable for simultaneous faults of multiple sensors,was further studied.The reconstruction precision was higher and the application range was wider.