Benders decomposition with integer sub-problem applied to pipeline scheduling problem under flow rate uncertainty

Important issues in a pipeline system are energy efficiency, reliability and throughput flexibility. Practically conventional pumps are not capable of operating at the highest attainable efficiency for long running time. This deficiency has prevented a pipeline system to operate close to its predefined program. A possible remedy is to take into account uncertainty due to pumps operations. In this paper, a stochastic two-stage mixed integer programming (MILP) model is developed for the multiproduct pipeline-scheduling problem under flow rate uncertainty. The problem arises in a number of settings, and the real-world applicability discussed and demonstrated. The stochastic MILP model involves many discrete variables that make it intractable for real-life cases. As a solution method, the sample average approximation is combined with a three-step solution approach based on Benders decomposition. The modeling and solving approach is evaluated in some case studies including a real-life problem from NIOPTC. (C) 2019 Elsevier Ltd. All rights reserved.