48 / 2025-03-29 15:48:35

Enhanced Model Predictive Control for Oscillation Suppression in Dynamic Processes

Model predictive control (MPC), Continuous stirred-tank reactor (CSTR) system, Oscillation penalty functions

Traditional Model Predictive Control (MPC) struggles to suppress output oscillations in industrial processes due to disturbances and model uncertainties. This work proposes an enhanced MPC strategy incorporating first- and second-order output differential penalty terms into the cost function. These terms explicitly penalize output oscillation rates and accelerations, improving robustness. The optimization is efficiently solved via quadratic programming (QP). The two-CSTR case study validates the method’s superiority over traditional MPC and PID controllers, demonstrating smoother output responses and reduced fluctuations.