A significant portion of manufacturing operations in the chemical process industry (CPI) and almost all the operations in the electronics industry and discrete parts processing industry are noncontinuous (batch/semicontinuous) in nature. These processes are economically very attractive for low-volume specialty products with high profit margins. A major advantage of noncontinuous processes is that they are flexible and can be multipurpose in nature, i.e. they are ideal for producing multiple products by sharing of common equipment. They provide flexibility in accomodating the variable nature of feed stock materials, the large number of products using similar recipes, the nature of processing steps (e.g. complex reactions, bacterial culture) and the short allowable inventory life of the products. Often dedication of separate facilities for different products can not be justified. Significant improvements in the cost-efficiency, productivity and flexibility of such plants can be achieved by using a systematic design methodology. The objective of the proposed research is to develop a computer-aided methodology for optimally designing a multipurpose plant in its most general form. A general multipurpose plant will employ a common pool of units of different types and sizes to produce a set of products. A unit may perform tasks for multiple products. Tasks may be assigned to a group of single or multiple nonidentical units, thus fully parallel production of different products can be achieved. Also a type of equipment may have items of different sizes. The optimal design problem will be to determine the number of batch/semicontinuous units, their types, their sizes, their time usage among different products and the production schedule of products. A mixed integer nonlinear programming formulation will be developed for the general design problem. Several rules/procedures will be devised for the systematic development, tightening and size-reduction of the formulation. An efficient procedure based on an outer- approximation/relaxation strategy will also be developed for the solution of the nonconvex optimization problem. This project will establish a systematic engineering design base for general batch plants, which will enable one to design them in a multipurpose form having high flexibility, cost-efficiency and equipment utilization. The project is also expected to lead to advances in the optimal solution of nonconvex mixed-variable optimization problems.
