A Multi-objective Model Predictive Control for Temperature Control in Extrusion Processes

In this paper, we consider the temperature control problem of an extrusion process with both heaters and coolers. For the purpose of energy saving and avoiding frequent switch between the heaters and coolers, the coolers are only used to guarantee the barrel temperature below a given safety bound. When this safety constraint is satisfied, the heaters take actions for reference tracking. This scheme is formulated as a multi-objective optimization problem in the framework of model predictive control. Different objectives have different priority. The safety constraint is of the highest priority, and formulated as a constraint in the optimization. Minimization of the inputs corresponding to the cooler is of the second priority, and incorporated into the objective function with heavy penalty weight. Minimization of the tracking error is of the lowest priority. Thus, this term is also incorporated into the objective function, but with light penalty weight. In this way, energy consumption can be reduced and frequent switch can be avoided. Moreover, a polytopic invariant set is developed to guarantee recursive feasibility of the proposed MPC. Simulations are also conducted to show the effectiveness of the proposed method.