Multi-objective optimization of integrated biodiesel production and separation system

Abstract

Biodiesel is a mixture of fatty acid methyl esters (FAME), which can be produced by the transesterification reaction of vegetable oils with methanol. The production of biodiesel is comprised of a series of stages such as reaction, methanol separation, water washing, decantation, and unreacted oil separation followed by the biodiesel purification. Each of these stages involves certain performance measures which need to be optimized for making the overall process cost-effective and productive. Therefore, a commercial base catalyzed batch transesterification process is optimized first with an objective: maximization of the fatty acid methyl esters concentration in final batch time. Dynamic optimization is used with the hot water flow rate passing through the reactor jacket as the decision variable. Orthogonal collocation on finite elements (OCFE) method is employed to convert the dynamic optimization problem into a non-linear programming (NLP) problem. The integrated reaction-separation steps of the biodiesel production process is modeled and simulated in Aspen Plus. Total two multi-objective optimization (MOOP) problems are formulated and solved to capture the trade-offs among the performance indices of the reactor and separation units collectively. Compared to the dynamic optimization problem result, MOOP-1 produced an optimal biodiesel concentration of 0.97 kmol/m3 at an optimal reactor energy cost of 0.12 $/l FAME produced. MOOP-2 determined that the energy expenditure in the reactor jacket is 1.064 times the total energy usage of all the reboilers of the distillation columns. In comparison to MOOP-1, the Pareto-fronts of MOOP-2 produced more cost-effective results in terms of comprehensive trade-offs among the objective functions used.