Supply Chain Optimization of Petroleum Refinery Complexes

Production planning is a very important step in managing the operation of a single refinery and their complexes. Despite its significance, few contributions have been found in literature and the existing ones rely on linear models. Pinto et al. (Comp. Chem. Engng. 24, 2259-2276, 2000), presented a superstructure that represents a general topology and allows the implementation of nonlinear process models as well as blending relations. The main objective of the present work is to extend the single refinery model to a corporate planning model that contains multiple refineries, which can be connected by supply pipelines in common. Intermediate streams also interconnect refineries in order to take advantage of each plant infrastructure. The model is optimized along a planning horizon resulting in a large scale Mixed Integer Nonlinear Program (MINLP). The non-linearity arises from blending equations and physical properties. The objective function maximizes net present value under raw material and product inventory level constraints as well as mass balance and operating constraints in each refinery. Finally, detailed analysis for different crude oil types and product demand scenarios is incorporated in the model. A real-world application is developed for a refinery network composed of three refineries. Different petroleum types are supplied to the refineries from a single oil terminal. Results show that the optimization of the supply chain presents clear advantages of the corporate planning with respect to multiple one-site refinery production planning.