A MPC-Based Optimal Temperature Control Strategy for Final Cooling Towers in Coking Processes

Coking production processes are strongly coupled and time-varying. As a critical component in coking production, the final cooling tower's outlet gas temperature directly affects downstream stability and efficiency. Conventional PID struggles under disturbances like gas-flow and cooling-water fluctuations. We propose a Model Predictive Control (MPC) scheme for outlet temperature control. Based on the energy conservation with unidirectional heat transfer and effective temperature difference assumption, we build a heat-transfer model and apply frozen linearization to obtain an affine predictor between temperature and water-flow setpoints. Then, a quadratic objective with three weights-tracking, control-magnitude regularization, and increment regularization-is solved online as a convex Quadratic Programming (QP) in receding-horizon form, where only the first control move is applied. Furthermore, dual-layer constraints are imposed to enforce both rate and actuator limits. Experiments show that the proposed method achieves fast, accurate, and robust regulation, indicating strong engineering applicability and interpretability for stable, efficient operation.