(285f) Integrated Crude Oil Blending and Scheduling for Petroleum Refinery

Crude oil scheduling plays the key role in refinery plant's operation. It is also the most complex part of refinery plant's operation. In practice, different types of crude oil can be blended in order to meet distillation and other facilities specifications. Crude oil blending extends feasible crude types, because a crude oil that was not suitable for processing can be blended with other oils to become processing viable. Obviously, crude oil blending makes the scheduling problem more complex.

In this paper, a crude oil blending and scheduling model was proposed to simulate and optimize crude scheduling problem under blending. The model covers crude oil purchase, storage, blending, crude distillation, and simplified cracking facilities. The time period is from one day to one month. Within this time period, the arrival time of crude oil can be accurately obtained from tank ship's schedule. Other raw material and production demands are obtained from monthly plan.

The model's objective is to maximize the plant profit. But firstly it should meet inventory and production demands, and maintain refinery plant's normal operations. These demands are expressed by model constraints. The scheduling model is a large-scale mixed integer nonlinear programming problem, which takes long time to solve. Thus, the scheduling model could not take every factor into account to avoid the model become too complex. To expedite the solving process, industrial experiences were integrated into the model as constraints. These constraints address API gravity and sulfur content of blended crude oils, and special crude oil selections for making lube oil and asphalt.

To demonstrate the efficacy of the developed blending and scheduling model, a case study for a refinery under uncertainties of crude oil changes was presented. It shows the model always helped the refinery plant select the most profitable crude oils.