Optimization model for the operational-level scheduling of multi-source pipelines with parallel batch injections

Abstract

Pipeline networks are the shippers' first choice for carrying large volumes of refined petroleum productsfrom oil refineries to distant distribution terminals. Optimization approaches for solving the pipelinescheduling problem proceed in two hierarchical stages: the aggregate and the detailed planning steps.The aggregate plan determines the batch sizes, the sequence of batch injections, and the allocation ofbatches to customers. The subsequent stage refines the aggregate plan to find the detailed schedule ofbatch input and output operations. This paper presents a mixed-integer linear programming (MILP) formulationfor the detailed scheduling of multi-source pipelines that accounts for parallel batch injectionsand simultaneous product deliveries to multiple terminals. It overcomes a critical drawback of previousmodels that assume single source configurations. Modeling multi-source pipeline networks is a greatchallenge, requiring a completely revised approach. The new model finds cost-effective solutions withremarkable efficiency.