Diagnóstico de acidentes da usina nuclear de Angra 2 baseado em agentes inteligentes de aquisição em tempo real, e em um modelo de árvore lógica

• Ano de defesa: 2017
• Autor(a) principal: Paiva, Gustavo Varanda
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Rio de Janeiro; Brasil; Instituto Alberto Luiz Coimbra de Pós-Graduação e Pesquisa de Engenharia; Programa de Pós-Graduação em Engenharia Nuclear; UFRJ
• Palavras-chave em Português: Engenharia de transportes; Diagnóstico de acidentes; Sistema especialista; Usina nuclear; CNPQ::ENGENHARIAS::ENGENHARIA NUCLEAR
• Link de acesso: http://hdl.handle.net/11422/10384

This work aims to create a model, using the Python language, which with the application of an Expert System uses production rules to analyze the data obtained in real time from the plant and help the operator to identify the occurrence of transients / accidents. In the event of a transient the program alerts the operator and indicates which section of the Operator's Manual should be consulted to bring the plant back to its normal state. The generic structure used to represent the knowledge of the Expert System was a Fault Tree and the data obtained from the plant was done through intelligent agents that transform the data obtained from the plant into Boolean values used in the Fault Tree, including using Fuzzy Logic. In order to test validate the program, a simplified model of the Almirante Alvaro Alberto 2 Nuclear Power Plant (Angra 2) manuals was used and with this model, simulations were performed to analyze the program's operation and the reduction in the transient identification time when compared with manual methods. The results of the tests presented significant reduction in the time and great accuracy, demonstrating the applicability of the model to the problem.