Two‐dimensional iterative learning robust H∞ optimization and linear quadratic fault‐tolerant control design for uncertain batch processes

Abstract Traditional two‐dimensional control strategies often struggle to simultaneously achieve both dynamic robustness and tracking accuracy when addressing control challenges posed by partial actuator failures and uncertainty for batch processes. Therefore, this paper proposes a two‐dimensional H∞ linear quadratic iterative learning control (2D‐H∞‐LQILC) strategy. First, a novel state space model is constructed by integrating combining state increments along the time domain with tracking errors, providing greater design flexibility for subsequent controller design. Second, by introducing a differential operator along the batch direction, a two‐dimensional state space model was constructed, endowing the controller with iterative learning capability. Furthermore, an iterative learning control law with optimal robust performance index is designed. This control law effectively suppresses the impact of uncertainties and ensures tracking performance and convergence capability under partial actuator failures. Finally, validation results for both constant and time‐varying partial actuator failure scenarios demonstrate that the proposed 2D‐H∞‐LQILC strategy outperforms conventional two‐dimensional control strategies in terms of control accuracy, convergence characteristics, and robustness.