Uma abordagem de decomposição por heurísticas e programação matemática para o scheduling de operações de um poliduto com uma origem e múltiplos destinos: rede OSBRA

• Ano de defesa: 2016
• Autor(a) principal: Meira, William Hitoshi Tsunoda
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso embargado
• Idioma: por
• Instituição de defesa: Universidade Tecnológica Federal do Paraná; Curitiba; Brasil; Programa de Pós-Graduação em Engenharia Elétrica e Informática Industrial; UTFPR
• Palavras-chave em Português: Pesquisa operacional; Programação linear; Agenda de execução (Administração); Método de decomposição; Heurística; Petróleo - Refinarias; Modelos matemáticos; Engenharia elétrica; Operations research; Linear programming; Scheduling (Management); Decomposition method; Heuristic; Petroleum refinaries; Mathematical models; Electric engineering; CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA
• Link de acesso: http://repositorio.utfpr.edu.br/jspui/handle/1/1844

This work presents an approach to optimize the scheduling of a real pipeline network with one source and multiple destinations, the OSBRA network. The network consists of a refinery that supplies five terminals, which are connected by a single pipeline. Batches of different products are sent from refinery and received by terminals during a time horizon. Optimized scheduling decisions must be taken to attend inventory issues at the same time that operational constraints are respected. This work proposes a decomposition approach solution using Mixed Integer Linear Programming (MILP) combined with heuristic algorithms. The solution approach is divided into two phases: (i) allocation and sequencing and (ii) scheduling of deliveries. The allocation and sequencing phase is responsible for determining the order, volume, and flow rate of each batch to be pumped from the refinery to the terminals. This phase is composed of an MILP model that runs iteratively in a rolling horizon approach. In a following step, the scheduling of deliveries phase details how the batches' deliver will occur in each terminal, considering the repumping operation in the last pipeline segment. For the repumping, a mathematical model is executed iteratively, also using the concept of rolling horizon, setting a new sequence of batches to be pumped from Goiânia to Brasília, the farthest pipeline segment from the refinery. Similar characteristics of the real OSBRA network are considered, such as inventory control, flow limits, daily demand, tank maintenance, total or partial pipeline maintenance. Two study cases considering scenarios based on real data with a 30-day programming horizon are presented, in which the second case was subjected to a series of modifications in order to evidence the proposed approach functionalities. The result's analysis supports the conclusion that the solution approach provides adequate results for the operation of the network experts. Thus, the obtained results can aid the operational scheduling decision of OSBRA network.