Multiobjective Optimization of an Industrial Ethylene Reactor Using a Nondominated Sorting Genetic Algorithm

Ethylene is produced in the largest volume among the monomers, and hence, any improvement in its production process can bring important benefits to both industry and consumers. In the present paper, an industrial ethylene reactor has been studied with a multiobjective optimization technique to find a scope for further improvements and to detect a range of optimal solutions. An industrial reactor unit using ethane as the feedstock was modeled, assuming a detailed free-radical mechanism for the reaction kinetics coupled with material, energy, and momentum balances of the reactant−product flow along the reactor. To carry out the multiobjective optimization for two and three objectives, the elitist nondominated sorting genetic algorithm, or NSGA-II, was chosen. Instead of a single optimum as in traditional optimization, a broad range of optimal design and operating conditions depicting tradeoffs of key performance parameters such as conversion, selectivity and ethylene flow rate was successfully obtained. The effects of design and operating variables on the optimal solutions are discussed in detail, and the generated results are compared with industrial data.