An MILP approach for scheduling of tree-like pipelines with dual purpose terminals
- PDF / 2,826,683 Bytes
- 29 Pages / 439.37 x 666.142 pts Page_size
- 64 Downloads / 135 Views
An MILP approach for scheduling of tree‑like pipelines with dual purpose terminals Mehrnoosh Taherkhani1
Received: 27 December 2017 / Accepted: 20 April 2018 © Springer-Verlag GmbH Germany, part of Springer Nature 2018
Abstract Pipelines represent the most reliable and economical mode of fluid transportation in the petroleum supply chain. They are often multi-product systems and are extensively used to carry different types of petroleum derivatives from refineries to distribution depots. This paper addresses the optimal scheduling of a treelike pipeline that connects several refineries to multiple depots and that meets the customer demands over a multi-period planning horizon. A continuous time scheduling formulation based on a mixed integer linear programing framework is presented which allows intermediate nodes to act as dual purpose stations. The problem goal is to satisfy local market requirements on time while keeping the inventory levels at depot tanks within feasible ranges. Solutions to three case studies show remarkable reductions in the CPU time with regards to previous contributions. Keywords Pipeline scheduling · Continuous time · MILP · Inventory management
1 Introduction In the oil supply chain, shipments of crude oil derivatives (kerosene, gasoil, gasoline, and jet fuel) via pipelines provide the most reliable and cost-effective transportation mode over land. In contrast with other transportation means (e.g., road, railroad, and vessel), pipelines can work continuously and convey large amounts of petroleum products between production and demand points (Rejowski and Pinto 2003; Mostafaei et al. 2015a, b, c). Pipelines can be arranged to three classes based on their configurations: straight lines with no branches, tree-like pipelines, and
* Mehrnoosh Taherkhani [email protected] 1
Department of Industrial Engineering, Zanjan Branch, Islamic Azad University, Zanjan, Iran
13
M. Taherkhani
mesh-structure pipelines, in which oil products can be conveyed to distribution centers through alternative paths. One of the main characteristics of pipelines is that they must be completely full of products at any time, and hence when a volume of material enters a pipeline, the same volume should be withdrawn at the other extremity. Pipelines are often multi-product systems, meaning that a variety of petroleum products can move back to back in the same duct. As a consequence, a mixing volume at the interface of two products is inevitable. The main decisions of pipeline scheduling concern the optimal sequence, timing and amount of injections (aggregate schedule) and deliveries (detailed schedule) in order to meet product demand at minimum total cost. The literature has used several optimization tools to solve the pipeline scheduling problem, such as knowledgebased heuristic techniques (Sasikumar et al. 1997; Chen et al. 2017a), decomposition frameworks (Hane and Ratliff 1995; Magatão et al. 2012, 2015; Boschetto et al. 2010; Fabro et al. 2014; Stebel et al. 2012) and rigorous optimization met
Data Loading...
