Modelagem e simulação de sistemas a eventos discretos via redes de Petri estocásticas: aplicação em mineração

• Ano de defesa: 2015
• Autor(a) principal: Cesar Monteiro Ribeiro
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Simulação de sistemas a eventos discretos; Redes de Petri; Mineração a céu aberto; Engenharia elétrica
• Link de acesso: http://hdl.handle.net/1843/BUBD-9ZDK8E

Some operational production processes are typically complex, mainly because of their dynamic nature, that changes their state over time, and their stochastic nature, which is determined by variables associated with uncertainty. A discrete event system is a dynamic system that evolves with the occurrence of events at intervals generally irregular and uncertain, being suitable to model operational processes. Among the techniques of discrete event system, Petri nets are an elegant way of modeling, whose simulation is simple and direct, besides having a well-established formal specification and being widely disseminated. They are suitable for modeling systems that have parallel concurrent asynchronous non-deterministic activities. In this work, the concepts of Petri nets were used to model the operations of loading and hauling mining processes of an open pit mine. Two diferent Petri net models have been developed in this work. The models follow a chronological order in the work evolution, showing an increasing level of complexity and accuracy. The first model has only 43 places, 38 transitions and 113 arcs. The fourth model, the most complete, which has characteristics of a colored Petri net, has 255 places, 170 transitions and 606 arcs. In an instance with realistic values, the loading/unloading and hauling represent about 60% of the mine productivity value and operational stops and maintenance represent about 40%. The simulator shown to be approximately three times faster than the reference model constructed in SIMAN. The productivity measure converges in the first seconds of simulation. Production results presented errors below 4%.