(58b) A Process Integration Approach to the Optimal Scheduling of Unsteady State Material Recovery Networks

Industry at large and the chemical industry in particular are striving to improve productivity and enhance revenues in a highly competitive global economy. While much attention has been given to improving design, efficiency, and performance for individual units in chemical plants, less attention has been given to the overall integrated optimization of the processes and the overall management of resources and wastes within chemical plants. This work takes a process integration approach to the optimal scheduling of unsteady state material recover networks. The problem of scheduling the operation of a process receiving time-varying feedstocks to yield products with desired specifications is considered. A systematic procedure and methodology was developed to schedule and operate an unsteady state material recovery network with given tanks, pipelines, and interception units. The objective is to determine the storage, allocation, and separation of the various sources over time so as to achieve maximum profitability while meeting all process constraints. Due to the dynamic variation of the sources, the problem was transformed into a multi-period problem with discrete time intervals q spanning time period t. The sources, interceptors, and tanks were discretized into a number Nt of time intervals spanning the cycle time t. A structural representation was developed to embed the network configurations of interest for scheduling. Bellman's principle of optimality was used as the basis for a two-stage optimization formulation. In the first stage, parametric optimization is carried out to determine the optimal policies for operating each interception device. Next, these optimal policies are incorporated in an optimization formulation that seeks to optimally schedule the storage, mixing, and interception of sources. The mathematical formulations over time period t were solved simultaneously for the discretized periods to achieve total annualized profit of the network while determining key scheduling information. A case study involving the recovery of ethanol was solved to illustrate the usefulness of the devised procedure.