Improved design of constrained model predictive tracking control for industrial processes against unknown uncertainties and partial actuator failure

Abstract For complex industrial processes in which model uncertainties, external disturbances, and partial actuator faults coexist, conventional robust model predictive control (MPC) has difficulty in achieving satisfactory control performance. To address this issue, this paper proposes a robust constrained model predictive control strategy for discrete‐time systems with input and output constraints. First, a tracking error is introduced to the original state space model to construct an extended state space model that simultaneously incorporates process dynamics and tracking characteristics. Building upon this, a linear matrix inequality (LMI) condition ensuring robust stability is derived using a quadratic Lyapunov function. Input and output constraints are unified into LMI form, enabling simultaneous computation of feedback gains and predictive control laws. This approach achieves robust guarantees for closed‐loop stability and constraint satisfaction under unknown disturbances and partial actuator failures. Using injection speed control of an injection moulding machine as a case study, simulation comparisons with traditional robust MPC under model mismatch and partial actuator failure conditions demonstrate that the proposed strategy achieves smaller overshoot, smoother control inputs, and superior overall tracking performance while satisfying constraints. This validates the method's effectiveness and engineering application potential in real industrial scenarios.