Otimização do scheduling de movimentações de um duto com uma origem e vários destinos

• Ano de defesa: 2012
• Autor(a) principal: Ribas, Paulo Cesar
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Tecnológica Federal do Paraná; Curitiba; Programa de Pós-Graduação em Engenharia Elétrica e Informática Industrial
• Palavras-chave em Português: Otimização matemática; Programação (Matemática); Agenda de execução (Administração); Heurística; Programação linear; Engenharia elétrica; Mathematical optimization; Programming (Mathematics); Scheduling (Management); Heuristic; Linear programming; Electric engineering
• Link de acesso: http://repositorio.utfpr.edu.br/jspui/handle/1/400

This thesis presents a model to support decision making of scheduling activities in a sequential pipeline system, composed by one refinery source and multiple destinations. This model is inserted into a system and used to develop a case study in the real world pipeline system called OSBRA that links the largest refinery in Brazil, REPLAN, to five terminals located in four different states. The proposed system relies its method in the iterative execution of a model of mixed integer linear programming, with the time horizon being moved in order to achieve a schedule for a period of at least one month. The main operational features and restrictions of OSBRA are considered in this model, as the flow variation due to the entry or exit of bleeding receiving, always within the limits of both the flow sections of pipelines as of bleeding. Others features of the real system contemplated in the model are the receiving by the intermediate bases exclusively by bleeding, meaning that only part of the batch is received thus, avoiding downstream stretches from the pipeline to stop, daily variance in the demand and the use of stamp batches to prevent contamination between adjacent derivatives. Eighteen scenarios of OSBRA’s system were submitted to the model and high quality solutions were obtained, reaching a satisfactory level of service in most scenarios.