برنامه ریزی بازارمحور و بهینه سازی اقتصادی فرایندهای شیمیایی چند درجه ای مستمر

In this paper an approach for flexible production scheduling for continuous multi-grade chemical processes is proposed. The approach integrates the economics of production and of company–market interaction for single-machine multi-grade continuous processes. The operation of a continuous multi-grade plant is modeled utilizing the detailed inclusion of grade transitions and sales orders and opportunities. The added value of the operation, which is used as the criterion function, is modeled by analyzing the costs and benefits of plant operation. The model results in a mixed integer linear program which is solved to compute suitable short term production and sales decisions. The approach is demonstrated on a gas phase HDPE manufacturing plant.

During the past decades the chemical industry has been faced with a major change into a globally competing and demand driven mode of operation. Companies are required to respond quickly to changing market situations and must meet customer-specified product specifications. A main challenge and a key to demand-driven operation, is the allocation of production resources to comply with orders and physical constraints such as plant capacity and storage facilities. Such problems are generally referred to as scheduling problems, and their role in the internal supply chain of manufacturing companies is broadly acknowledged. An overview can be found in [11]. Although most of the scheduling literature focuses on batch operations, there is recent work on continuous process scheduling, see [8], [4] and [9]. The approach here will consider continuous, single-stage, multi-grade chemical plants, and differs from the work cited in several aspects. Our way of including the effect of process transitions on the material flows appears new, although in a recent paper [12], a formal way of modeling the effect of grade transitions on product flows has been presented in relation to a continuous time formulation cyclic scheduling problem for hybrid flowshop facilities. In Giannelos and Georgiadis [4] sequence-dependent changeover tasks are modeled, however their effect on the material balance is not considered. Further, most scheduling studies assume the order base to be fixed in advance and strive for ‘minimum makespan’, ‘minimum lateness’, production maximization or cost minimization. In our approach, the negotiation of sales orders and purchases is an integral part of the decision making that is supported by the scheduler, and to this end the scheduler selects a set of appealing purchase and sales transactions from a larger set of possible transactions (denoted opportunities). Making the schedule responsive to varying market situations, implies that it will not exhibit simple structures like for example a cyclic structure. The scheduling approach described in this paper is probably less general than formulation posed by Mockus and Reklaitis [9] and Lee and Lee [5], in the sense that for the moment, only single-machine chemical plants are considered. Nevertheless, with the detailed inclusion of grade transitions and sales orders and opportunities it is believed to capture some of the essential and distinguishing elements of the short term scheduling problem encountered in many high-volume chemical production plants. The scheduling approach considered here includes a formulation of the grade change problem that includes a truly economic objective. This way, a feasible interconnection of the short-term scheduler and the process control and optimization functions that operate on a faster time scale is achieved. The paper presents a survey of work available in detail in Tousain [14].

In this paper a scheduling approach has been presented that enables the production management of a multi-grade chemical plant to operate the plant according to a market-responsive and economics-based objective. For single-machine multi-grade processes, such a scheduler can be designed by capturing the purchasing, production and sales decisions and their effect on the company’s objective into a MILP which can be solved using standard software. Process transitions and their effect on material flows are included in the formulation without the need to introduce additional binary variables. Reasonable solution times were encountered for the HDPE production test problem that was considered in this paper. The application of the flexible short term scheduling approach to a gas phase HDPE manufacturing plant for a fictional market situation demonstrates a significant increase in added value compared to the traditional fixed and variable duration slate scheduling.