A Multi-Level Simulation Approach for the Crude Oil Loading/Unloading Scheduling Problem

摘要

This paper presents an integrated optimization approach for crudeoperations scheduling and production for refineries. The production process iscomposed of a pre-fractionator, crude, and vacuum distillation columns. It ismodeled as an NLP. The scheduling problem is composed of unloadingoperations and simultaneous blending and charging of CDUs. It is modeled asa MILP. The nonlinear simulation model for the production process is used toderive individual crude costs for the two crudes considered (Dubai andMasila). This is performed using multiple linear regressions of the individualcrude oil flow rates around the crude oil percentage range allowed by theproduction facility. These individual crude costs are then used to derive a linearcost function that is optimized in the MILP scheduling model, along withlogistics costs. Results show that this integrated approach can lead to a costfunction in the crude scheduling problem capable of including production levelcost.