Estudo do gráfico de controle por grupos como uma alternativa ao modelo de Shewhart em processos paralelos

Abstract

The use of the Charts of Shewhart as tool to monitor processes, whose products happen of multiple streams of production (parallel processes), must be seen with caution, therefore, the samples can be being constructed with item of different populations. A solution is to use a chart of control for each stream (traditional model of Shewhart), however, this would become the difficult and bureaucratic control; another alternative is the adoption of the Group Charts (GC), that it allows the control of multiple streams through an only chart. Therefore, the general objective of this work is to develop a study of the graph of control for groups in terms of its performance and efficiency as an alternative to the model of Shewhart in parallel processes. The study it was carried through in relation to the alteration in the mean and the variance of the process individually and in the mean and the variance jointly. This work looked for, according to boarding of some authors, to contextualize the importance of the statistical control of process inside of the logistic production. As the one of the results, keeping opening of the control limits in 3 standard eviations, the tax of false alarms grows with the increase of the number of streams, however, adjusting the control limits to take care of to an average tax of a false alarm for each 370,4 samples, it is had that with the increase of the number of streams the GC loses its performance.