Controle estatístico do processo de torneamento duro na presença de variável de ruído

• Ano de defesa: 2019
• Autor(a) principal: Evangelista, George Sousa
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal da Paraíba; Brasil; Engenharia de Produção; Programa de Pós-Graduação em Engenharia de Produção; UFPB
• Palavras-chave em Português: Design de experimentos; Controle estatístico do processo; Análise multivariada; Cartas de controle; Design of experiments; Statistical process control; Multivariate analysis; Control charts; Torneamento do aço - Controle de ruído; Variáveis de ruído - Estudos; CNPQ::ENGENHARIAS::ENGENHARIA DE PRODUCAO
• Link de acesso: https://repositorio.ufpb.br/jspui/handle/123456789/17179

The hard turning process has characteristics that favor its choice in relation to traditional turning. It presents better results of surface integrity, short process time, possibility of working without the use of lubricating fluid. Surface integrity results are the focus of studies of machining processes in general. The surface roughness is observed in relation to several controllable input parameters. Among the inputs can also be identified noise variables. These variables are uncontroled, but they affect process responses. In addition, there is more than one roughness response. Multivariate studies allow the simultaneous analysis of correlated responses. Thus, this research aims to perform univariate and multivariate analysis by principal components of the hard turning process. This analysis seeks to observe the effect of the noise variable on the responses and verify process stability and capability. For such analysis a hybrid method of experiment design and statistical process control was used. The AISI 52100 hardened steel turning was selected for the study application. Process scenarios were developed with a factorial arrangement. For this scenarios, regression models, control charts, process performance, sigma level and parts per million estimation were calculated. The performance (Ppk) results for the multivariate scenario were between 0.18 to 1.11. The best scenario presented had a cutting speed of 170m/min and a flow of lubricating fluid of 3 l/min.