Metodologia heurística na otimização da programação dos planos de manutenção preventiva e preditiva na agroindústria

• Ano de defesa: 2023
• Autor(a) principal: Obal, Jonatas Santana
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Tecnológica Federal do Paraná; Ponta Grossa; Brasil; Programa de Pós-Graduação em Engenharia de Produção; UTFPR
• Palavras-chave em Português: Heurística; Manutenção; Programação estruturada; Agroindústria; Heuristic; Maintenance; Structured programming; Agricultural industries; CNPQ::ENGENHARIAS::ENGENHARIA DE PRODUCAO; Engenharia/Tecnologia/Gestão
• Link de acesso: http://repositorio.utfpr.edu.br/jspui/handle/1/31247

Preventive and predictive maintenance are carried out to prevent failures, thus improving the performance of machines and industry production. However, the planning of preventive and predictive maintenance is a challenge in industries, as they must consider issues such as, the allocated cost to maintenance, the available maintenance time of each equipment and the previous survey of the manufacturer's considerations. In addition, programming maintenance plans is a humanly difficult task in terms of the distribution of tasks due to their criticality to technicians on working days, subject to restrictions on working time, day of execution, among others. In this sense, this work proposes a heuristic methodology to optimize the programming of preventive and predictive maintenance plans in the agroindustry. Considering maintenance plan data, their pre-processing verifies the availability of staff, both for plans that must occur on specific days and for those that can be carried out on any day of the week. With pre-processed data, the proposed heuristic should be executed weekly, and its output is a list of plans that will be carried out on each day of the week. The heuristic was applied in a test case and later replicated in a real case of an agroindustry. The application of the heuristic in a fictitious case allowed the verification of its operation, while its application in real data demonstrated its efficiency and answered questions that were beyond expectations. The real case data was extracted from 1680 SAP assets (from the German: Systemanalysis Programmentwicklung. From this application it was verified the need of thirteen weeks to update the plans, that is, no plan has a negative deadline. The average of plans in a week is 432, which is planned using around 3 seconds by the heuristic. The results were presented to the maintenance manager, who received with surprise that 75% of the maintenance plans were with negative deadline. Regarding the output, the speed of solution presented by the heuristic was a very positive point. This will drastically decrease the planner's planning time, which currently takes around two days to plan the week, while the heuristic does it almost instantly.