Predictive control of a low density polyethylene plant

The paper aims to present the possibilities of controlling a multizone tubular reactor for producing low-density polyethylene using the Model Based Predictive Control (MBPC) strategy, specifically Extended Predictive Self-Adaptive Control (EPSAC) approach. The possibilities of using “solidary predictive control” and "selfish predictive control" approaches to the Multi-Input Multi-Output (MIMO) case are investigated. The results are compared to the classical PID control. The results clearly show that MBPC can successfully deal with this kind of chemical process and gives better performances than PID for both SISO and MIMO cases.