Fault Tolerant Control of an Industrial Gas Turbine Based on a Hybrid Fuzzy Adaptive Unscented Kalman Filter

Abstract

Gas turbines are generally used in power generation, the oil and gas industries, and as jet engines in aircrafts. Fault tolerance and reliability is important in such applications. Thus, accurate modeling and control system design is necessary. In this paper, first a nonlinear hybrid fuzzy model was developed for an industrial gas turbine, and then this model was used as the core of a fault tolerant control (FTC) system. The aforementioned model was trained by use of three months of operational data of a GE MS 5002 D gas turbine that is used for gas injection application, then it was fine tuned using expert knowledge and physical principles. A graphical user interface (GUI) was also developed to run various realistic operational scenarios of the gas turbine. The main point of the present work consists in introducing nonlinear fuzzy model schemes as the core of an adaptive unscented Kalman filter (AUKF) for fault diagnostic purposes. Analysis of the simulation results discloses that this FTC approach alleviates the effects of faults in two different scenarios such as sequential drift and bias in sensors/actuators and also in simultaneous faults that are a disastrous situation.