Petroleumrefinery is a leading segment of the entire petrochemical industry. Scheduling and planning of crude oil is very important to a petroleum refinery due to the potential realization of large cost savings. It is also the one of most challenging problems for refinery operations. However, it is not easy to simulate and optimize the front-end crude-oil and refinery operations at the same time. And in reality, the crude scheduling activities are highly vulnerable to disruptions brought by various uncertainties.
In this paper, a large-scale continuous scheduling model has been employed to simulate and optimize the front-end and refinery crude-oil operations. The scope of the scheduling problem includes crude oil unloading from vessels to the crude storage tanks at onshore berths, transfer of crude from these tanks to the charging tanks, charging crude distillation units with crude mixes from the charging tanks and further processing of crude inside the refinery, which includes crude distillation, cracking, coking, reforming, hydrotreating, product blending and component recovery.
The objective of the proposed crude scheduling problem is to maximize the total operational profit; meanwhile, it should meet inventory and production demands, and try to maintain refinery plant’s normal operations under different circumstances of uncertainties. Typical uncertainties come from volatile crude market, product demand changes, shipping delay, increasingly strict environmental regulations, etc. The scheduling model is a large-scale mixed integer nonlinear programming problem (MINLP). An outer-approximation based decomposition method is implemented to obtain the optimum solution of the developed MINLP scheduling model. The efficacy of the proposed rescheduling model has been demonstrated by different case studies.
See more of this Group/Topical: Computing and Systems Technology Division