Forecast model applied to quality control with autocorrelational data

• Autor(a) principal: Souza, Adriano Mendonça
• Data de Publicação: 2013
• Outros Autores: Klidzio, Regiane
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Revista Produção Online
• Texto Completo: https://www.producaoonline.org.br/rpo/article/view/1217
• topic: Time Series. Linear Models. Autocorrelation. Residues. Control charts.; Séries temporais. Modelos lineares. Autocorrelação. Resíduos. Gráficos de controle.

This research approaches the prediction models applied to industrial processes, in order to check the stability of the process by means of control charts, applied to residues from linear modeling. The data used for analysis refers to the moisture content, permeability and compression resistance to the green (RCV), belonging to the casting process of green sand molding in A Company, which operates in the casting and machining, for which dynamic multivariate regression model was set. As the observations were auto-correlated, it was necessary to seek a mathematical model that produces independent and identically distribuibed residues. The models found make possible to understand the variables behavior, assisting in the achievement of the forecasts and in the monitoring of the referred process. Thus, it can be stated that the moisture content is very unstable comparing to the others variables.