Efficient planning of crude oil supplies through long-distance pipelines. (4th March 2019)

• Record Type: Journal Article
• Authors: Cafaro, Vanina G.; Pautasso, Pedro C.; Cerdá, Jaime; Cafaro, Diego C.
• Is Part Of: Computers & chemical engineering. Volume 122(2019)
• Page Start: 203
• Page End: 217
• Publication Date: 2019-03-04
• Subjects: Oil Supply -- Optimization -- Pipeline -- Planning; Chemical engineering -- Data processing -- Periodicals; 660.0285
• DOI: 10.1016/j.compchemeng.2018.06.028
• Languages: English
• ISSNs: 0098-1354

Abstract

Highlights: We address the planning of crude oil supplies through a long-distance pipeline. We propose a mixed-integer nonlinear formulation and a tailored solution strategy. Solution quality is assessed by solving a valid relaxation of the original MINLP. Very efficient solutions to industrial size problems are found in modest CPU times. The algorithm outperforms previous contributions, lowering costs and CPU burden. Abstract: This work introduces a mixed-integer nonlinear programming (MINLP) formulation for the efficient planning of crude oil supplies to a major refinery, through a long-distance pipeline. The proposed approach combines the potentials of slot-based and general precedence continuous-time representations to simultaneously address two problems that are typically decoupled: pipeline transportation and crude oil blending. General precedence sequencing variables coordinate incoming/outgoing flows to/from every tank, while crude oil batches are traced into the pipeline following a slot-based scheme. The model precisely monitors key component concentrations, keeping oil properties within admissible ranges. Since the proposed formulation is nonconvex, an efficient solution strategy is followed and a tailored relaxation is subsequently solved to assess the quality of the solutions achieved. Results show that the model is able to find very efficient solutions to real-world case studies in modest CPU times.