Planning the operation of a large real-world oil pipeline

Pipelines are the most effective and reliable means for transporting oil derivatives from refineries to local depots. This paper considers the tactical distribution problem for a large pipeline network. Solutions to this problem consist of volume transfers satisfying local demands, correctly stocking production volumes, and observing a number of operational constraints. This work defines the tactical planning problem and proposes a novel and compact network flow model to address it. Also, a procedure is given to decompose the flow solution into pumping operations that can be used as input for short-term schedulers. Our experiments were performed over real instances provided by a large oil company. These instances consisted of 30 pipelines in an extension of 7000 km, 14 depots, 192 tanks, and more than 60 products. The model was tested with real-world instances and showed significant improvements over manually generated plans, which is the current practice.

► We model the transportation of oil derivatives through a network of inland pipelines.
► We formalize the operational restrictions having the greatest influence on schedules.
► We propose a new network flow formulation to solve the tactical planning problem.
► We present a decomposition heuristic to get pumping orders from tactical solutions.
► Pumping orders can be input to other heuristics that get detailed pipeline schedules.