Multi-objective Optimization of Low Density Polyethylene (LDPE) Tubular Reactor Using Strategies of Differential Evolution

Abstract

Multi-objective optimization of industrial low density polyethylene (LDPE) tubular reactor is carried out using improved strategies of multi-objective differential evolution (MODE) algorithm (namely, MODE-III and hybrid-MODE). Two case studies consisting of two-objective optimization and four-objective optimization are considered. In case-1, two objectives namely, maximization of conversion and minimization of the sum of square of normalized side chain concentrations are considered. A set of eleven decision variables, which consists of operating variables, namely, inlet temperature (T in), inlet pressure (P in), the feed flow rates of -oxygen (F o), -solvent (F S), -initiators (F I,1, F I,2), and the five average jacket temperatures (T J,1 - T J,5), are considered. Constraints on maximum temperature attained in the reactor and number average molecular weight are considered. The results of present study show that MODE-III algorithm is able to give consistent results for various control parameters. These results show the ability of the existing algorithm to produce more valuable and practical results that are important to the process plant engineer.KeywordsDifferential EvolutionPareto FrontMultiobjective OptimizationPareto Optimal SolutionTubular ReactorThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.