Gain scheduling model predictive controller based on extended state Kalman filter for fluidized bed combustor

• A GS–ESKF–MPC strategy is proposed for fluidized bed combustor control. • The ESKF enables joint estimation of system states and disturbances. • Hysteresis–based gain scheduling ensures smooth model transitions. • IAE reduced by 50–70% vs PI and 20–40% compared with conventional MPC. Fluidized bed combustors (FBCs) are widely used in low-grade coal-fired power generation due to their fuel flexibility, high combustion efficiency and low emissions of pollutants. However, their strong nonlinear dynamics, multivariable coupling, and frequent disturbances pose significant challenges for control design. This study proposes a gain scheduling model predictive control strategy enhanced by an Extended State Kalman Filter (GS‑ESKF‑MPC) to achieve efficient and robust control of FBC systems. The nonlinear FBC model is linearized at multiple operating points to construct a library of local linear models, and a hysteresis-based scheduling mechanism ensures smooth transitions across the wide load range. The ESKF jointly estimates system states and lumped disturbances, providing disturbance-aware feedback to the controller. This framework transforms the nonlinear control problem into a sequence of tractable quadratic programs, enabling real-time implementation. Simulation results under representative scenarios demonstrate that the proposed GS‑ESKF‑MPC consistently outperforms both conventional PI control and gain‑scheduled MPC with standard Kalman filtering, reducing the integral of absolute error by about 50%–70% compared with PI and by 20%–40% compared with Conventional MPC, it also achieves faster disturbance rejection, smoother actuator behavior, and improved fuel economy. Overall, GS-ESKF-MPC provides a robust, efficient, and adaptive solution for precise control of fluidized bed combustors.