Fault tolerant control using evolving fuzzy modeling

This paper proposes a fault-tolerant control (FTC) approach using evolving fuzzy modeling. FTC is performed in two steps: fault detection and fault accommodation. Fault accommodation uses evolving Takagi-Sugeno fuzzy models, and fault detection uses a model-based approach also based on fuzzy models. Information from fault detection is used for fault accommodation in a model predictive control (MPC) scheme. The evolving fuzzy modeling approach increases the control performance when the process is with faults. The proposed approach continuously evaluate the control performance and perform on-line clustering, if it is necessary. Evolving FTC is used to accommodate two simulated faults in a distillation column process. The considered faults are the load process fault (variation in feed composition) and the change in heating (variation of re-boiler temperature). The fault tolerant control using evolving fuzzy modeling was able to accommodate the simulated faults.