Model-Based Faults Diagnostics of Single Shaft Gas Turbine Using Fuzzy Faults Tolerant Control

This paper proposes a new fault-tolerant control approach for gas turbines, based on fuzzy techniques that allow real-time observation of their behavior. The approach consists of detecting and locating faults, and then determining the appropriate control action to keep the turbine in stable operation. This improves the efficiency and lifespan of the turbine, and reduces maintenance costs with optimal planning. The state space model of the turbine is subjected to a diagnostic procedure based on Type-1 and Type-2 fuzzy models, using the operational data of the different operating points studied. The Luenberger observer is used in the fault detection mechanism, to characterize turbine component malfunctions by comparing the observed real behaviors and the fuzzy models. The results show that the fault-tolerant fuzzy control has ensured the stability and availability of the turbine during fault occurrence, with good performance in detection, isolation and reconfiguration.