Optimal Robust Design for Wood and Berry Distillation Column Using Multiobjective Genetic Algorithm Tuned Model Predictive Controller

Distillation column is the most common and frequently used process in chemical engineering. Improved control strategy can help reducing the cost and improving the quality of the product. Main concern in distillation column is the purity of output product. An optimal robust design has been implemented for Wood and Berry distillation column, a multivariable system with dead time. To achieve the desire robust response and tracking performance, infinity norm of the sensitivity function (S(jw)vertical bar)(infinity) and infinity norm of complementary sensitivity function (T(jw)vertical bar)(infinity) are minimized simultaneously using multi-objective genetic algorithm (MOGA). By tuning of the Model Predictive controller (MPC) parameters using MOGA, an optimal set of solutions are generated and the ideal solution is selected from the Pareto optimal set using level diagrams. From the simulation results it is clear that MOGA tuned MPC is robust in nature and it also performs the proper tracking of the desired distillation composition and bottom composition in wood and berry system.