An efficient approach to the operational scheduling of a real-world pipeline network

Oil-derivatives pipeline logistics using discrete-event simulation

“…It permits to easily find alternative detailed schedules by using different operational criterion. Several approaches were proposed to study pipeline scheduling problems, including rigorous optimization models, knowledge-based techniques (Sasikumar et al 1997), discrete-event simulation (Mori et al 2007, García-Sánchez, Arreche, andOrtega-Mier 2008), and decomposition frameworks (Hane and Ratliff 1995, Neves et al 2007, Boschetto et al 2008. Rigorous optimization methods generally consist of a single MILP (Mixed Integer Linear Programming) or MINLP (Mixed Integer Nonlinear Programming) mathematical formulations and are usually grouped into two classes: discrete and continuous, depending on the way that volume and time domains are handled.…”

“…However, a recent paper of Rejowski and Pinto (2008) assumes that each pipeline segment is composed by packs with equal or different pre-specified volumes to account for reductions in the pipeline diameter, and the horizon length comprises time intervals of adjustable duration to allow changes in the pump injection rate. Neves et al (2007) presented a decomposition approach for the planning of pipeline operations over a monthly horizon. The decomposition relied on a heuristic pre-processing block that accounts for demand requirements, production planning, and typical lot sizes to determine a candidate set of product sequences.…”

“…Since seasonal costs of the electric energy are considered, the model includes binary variables just to avoid pumping operations during high-energy cost periods. Boschetto et al (2008) reformulated the hybrid approach of Neves et al (2007) using a different decomposition strategy now involving three blocks: (i) a resource-allocation block determining candidate sequences of batch injections, (ii) a pre-analysis block specifying the precise volumes to be either pumped from source nodes or received in destination nodes, and providing the earliest start/finish times for stripping operations in every destination node, and (iii) a continuous-time MILP model determining the exact timing of pump and delivery operations at each node.…”

Computação Evolucionária em Problemas de Engenharia

“…Uma melhor eciência no uso da rede, quando possível, é uma solução mais barata e mais rápida que investimentos na ampliação da malha de distribuição. De acordo com Neves-Jr et al (2007), é um grande desao a busca por modelos de scheduling para redes de dutos complexas integrando diversas áreas e transportando diferentes produtos que possam ser implementados na prática, considerando uma carga computacional aceitável. Atualmente a programação das operações em redes reais é realizada por especialistas experientes, entretanto técnicas de pesquisa operacional têm se tornado uma ferramenta eciente na otimização destes processos operacionais, auxiliando na modelagem e resolução de problemas combinatórios (Relvas et al, 2006).…”

Optimization model for the detailed scheduling of multi-source pipelines

“…, k19 , k22 ) supply products to multiple terminals at the same time. The model comprises 3071 variables (157 of which are binary)Best detailed pipeline schedule found for the supplementary example…”