Chemical Engineering Transactions (Sep 2013)
Optimization of PID Controller Parameters in the Case of Batch Styrene Suspension Polymerization
The highly exothermal polymerization reaction needs an efficient cooling system in order to avoid thermal runaways. The thermal agent flowrate or inlet temperature may represent the manipulated variable (MV) that the controller should adjust according to a set of laws to achieve a suitable goal. The laws the controller observes determine the class it belongs to: proportional (P), integral (I), derivative (D) or a combination of them. Due to its versatility and better performances, the PID controller was our choice. Its parameters can be set either by applying a set of rules, like Ziegler-Nichols’ or Cohen-Coon’s, or by searching for the optimal values fulfilling a performance criterion concerning the polymerization process, the optimization algorithm being suitably selected. One possible choice as performance criterion could be the distance between the actual reactor operating temperature and a set point. An interesting case study for applying this solution strategy is the batch reactor for the polymerization of styrene in suspension due to the two-step heat transfer: a) directly, from the monomer-polymer drops of given distribution to the continuous phase and b) indirectly, from the latter to the thermal agent, through the reactor wall. This way, the monomer/polymer drop temperature (which is function of drop’s size as well and determines the polymer characteristics) can be influenced by the way the system responds to the adjustments made by the controller. The drop-water interface acts like a heat sink, but the average temperature of the drop is maintained at a higher level as compared to water by the ongoing polymerization reaction. The larger the drops, the higher the temperature differences between them and the continuous medium, with implications, as is to be shown by this contribution, on the optimized PID controller parameters.