Controle, supervisão e suporte à tomada de decisões em processos avícolas

• Ano de defesa: 2018
• Autor(a) principal: Schmidt, Lucas
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Tecnológica Federal do Paraná; Pato Branco; Brasil; Programa de Pós-Graduação em Engenharia Elétrica; UTFPR
• Palavras-chave em Português: Indústria avícola; Sistemas de controle supervisório; Inteligência Artificial; Automação; Poultry industry; Supervisory control systems; Artificial intelligence; Automation; CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA; Engenharia Elétrica
• Link de acesso: http://repositorio.utfpr.edu.br/jspui/handle/1/3252

The production of broilers has assumed a role of great socioeconomic importance for Brazil. For the success in the development of the poultry, an adequate control of temperature and humidity inside the aviary is essential, since the thermal comfort defines the relation between the feed consumption and the meat production. Current control methods, commonly used in Brazilian aviaries, do not have the desired efficiency. Although there are semiautomatic devices, they are dependent on the perception and intervention of the human expert, which tends to be inefficient and prone to errors. This research proposes an integrated approach to control, supervision and support to decision making for application in poultry management. Two control approaches presented in the literature will be analyzed and a hybrid integration approach will be proposed. The first one deals with a controller based on Supervisory Control Theory (TCS), able to observe the set of events of the plant and orchestrate the actuators in a minimally restrictive, controllable, non-blocking manner and according to a set of specifications. The second concerns a machine learning model that, through an artificial neural network (RNA), provides action plans for poultry management, prioritizing poultry welfare and productivity. The two approaches will be refined and then integrated, which will provide the operator with three process management options: purely logical control (TCS); control through artificial intelligence (RNA) strategy; and the logical control integrated to the action plans provided by the expert system. Once the integration stage is over, real management situations is= simulated, and responses are obtained for the three control options. Finally, a comparative analysis is presented between the results obtained through each approach.