Detecção e diagnóstico de falhas em alto - forno - um estudo de caso: um estudo de caso

• Ano de defesa: 2002
• Autor(a) principal: Robercy Alves da Silva
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Engenharia elétrica
• Link de acesso: http://hdl.handle.net/1843/BUOS-8CVNNC

This work consists in a system developed to detect scaffold in blast furnace. The scaffold is a failure, which can cause big money losses, and injury employees, and is very difficult to be detected by blast furnaces operators. Using computing intelligence techniques, this system isallowed to detect, in real time, the scaffold growing, alarming and showing the scaffold position inside the blast furnace. How this is a detection and diagnoses problem in a difficult modeling dynamics system, it will be solved using pattern classification theory. To pattern classification, will be proposed a new neurofuzzy topology, which is an ANFIS topology variation. To training this network, was used data acquired from the plant. To validate, was used data from the same plantduring the last eight. During this time, the system detected seven real scaffold situations. These preliminary results certify this system as a useful solution to solve problems of scaffold detection in blast furnaces.