Uma abordagem de otimização para a programação do transporte de derivados escuros de petróleo por uma malha dutoviária

• Ano de defesa: 2020
• Autor(a) principal: Bueno, Lucas
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• 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: Petróleo - Derivados - Transporte; Otimização matemática; Programação heurística; Programação (Matemática); Modelos matemáticos; Oleodutos de petróleo; Programação linear; Petroleum products - Transportation; Mathematical optimization; Heuristic programming; Programming (Mathematics); Mathematical models; Petroleum pipelines; Linear programming; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO; Engenharia Elétrica
• Link de acesso: http://repositorio.utfpr.edu.br/jspui/handle/1/23590

Due to its efficiency, the pipeline modal is commonly used to transport oil and its derivatives. However, since expanding the pipeline networks involves high costs, it is important to optimize the use of the existing resources. One step to optimize the use of pipeline networks is to make a proper schedule — an NP-complete problem. Thus, this thesis is about an optimization approach for scheduling heavy oils transportation through a Brazilian mesh-like pipeline network. In this network, seven pipelines connect eight nodes, four of which are refineries, three are intermediate depots and one is a harbor. Some characteristics of this problem differentiate it from similar ones, such as the need to perform oils quality degradation, oils blends and changes on flow direction (reversions), as well as the need to include plugs to avoid undesired interfaces and to consider oils heat loss. To generate solutions, the problem is decomposed into three steps: assigning, sequencing and timing. The assigning step is solved with an Mixed Integer Linear Programming (MILP) model, the sequencing step with a heuristic algorithm and an MILP model, and the timing step with a heuristic algorithm and an Linear Programming (LP) model. Quantitative and qualitative experiments were carried out, with which monthly schedules were obtained in a non-prohibitive computational time (minutes).