Model based multi-loop predictive control scheme for multivariable processes

Industry applications of multi loop controller approaches are common because of their robustness, minimal tuning parameter, and straightforward design. A model based multi-loop predictive control method is presented. The proposed method directly uses the process model without approximation and reduction. A sequential relay based auto tuning procedure is applied for the tuning of the controller parameters. The efficacy of the suggested multi-loop control strategy is illustrated through comparative analysis and simulation. Simulation results show that the proposed method gives better response in terms of Integral Absolute Error value compared to multi-loop PI controller tuned by other methods. The proposed predictive method has IAE value of 28.9 whereas Biggest Log Modulus Tuning method has 55.34 for the Wood and Berry distillation column process. Alike, the proposed predictive method has IAE value of 350.26 whereas BLT method has 360.01 for the Orgunnaike and Ray distillation column process. In general, the multi-loop predictive PI scheme improves the overall performance.