Process Integration: Multiple Single Product vs. Multiproduct Plants

GABRIELA CORSANO,; MARCELO MONTAGNA,; AGUIRRE, PIO; OSCAR IRIBARREN,

Rio de Janeiro

Congreso; 15th European Symposium on Computer Aided Process Engineering (Escape 15); 2005

Process synthesis and design of batch plants must consider scheduling strategies simultaneously, to obtain a feasible design. In this paper NLP models are presented for getting the optimal design and operation of a batch plant where products follow different production routes sharing some but not all of the stages. The structure of the plant is decided: the alternative of multiple single product plants is compared with a multiproduct plant. Also for some batch stages, a superstructure is considered in order to obtain the best number of units and their configuration. A separate plant provides material and energy resources for these processes, with the residuals of one becoming raw material for others. The objective function employed is the maximization of the total profit given by the earnings of selling products and savings due to unused resources, minus the total annualized investment and operation costs. The multiproduct plant is designed by first solving a relaxed model, with the objective of finding the optimal campaign sequence and ratio between the number of batches of each product. Next a NLP model is solved to obtain the optimal feasible design, implementing a novel representation of the scheduling constraints. This representation arises from the zero-wait policy adopted, but is not the classical multiproduct plants flow shop formulation. In order to assess the multiproduct plant design, a single product multiplant complex was designed for the optimal production targets obtained with the multiproduct model. The design and operation cost for this complex is 22% larger than the one obtained with the multiproduct scheme. Optimal design and operation variables, plant structure and recycle streams for both schemes, and the comparison of both strategies are reported. Keywords: Process Integration, MultiproProcess Integration, Multipro