Desempenho do gráfico de controle 3-D para a média na presença de alterações na média e desvio padrão do processo: uma abordagem prática

Abstract

Among the tools that can be used for monitoring production processes is the statistical process control (SPC), created by Walter A. Shewhart, for the construction of control charts. This method is based on production processes with a single production flow, however, in parallel processes, where there are several production flows, monitoring through traditional graphics becomes difficult and bureaucratic. It is also important to highlight that in parallel processes there are two different variability: longitudinal and transversal. An alternative, in these cases, is the construction of a 3-D control chart, which consists of the elaboration of three charts: one to monitor the mean, one to monitor the transversal variability and another one for the longitudinal variability. Therefore, the objective of this work is to determine the performance of the 3-D control chart for monitoring the mean in the presence of changes in the mean and standard deviation of the process in a practical approach. As a research methodology, data from the manufacture of metal stoppers from a processing industry in Ceará was used. This parallel process consists of beats of a metal plate of a machine, where in each beat 27 stoppers are produced, which originate from 27 punches. Furthermore, simulations were performed with changes in the process mean, in the process standard deviation and in the mean and standard deviation simultaneously. From this, the 3-D control chart proved to be efficient in detecting evidence that the process is out of control and has a good performance when these changes are present in production