Desempenho do gráfico de controle por bateladas para a média na presença de alterações na média e na variância do processo: uma abordagem prática

Abstract

The Statistical Process Control Chart (GCEP) owes its origin to Walter Shewhart at the beginning of the 20th century, consisting of an adequate tool to inspect and evaluate a production process. The graph initially proposed by Shewhart takes into account the mean and variability of the process. In a batch production process, variations may occur within each batch or batch and variability between batches. Differences between batches must be understood as part of the process and must be incorporated into the applied control chart. Within this context, the batch control charts are built from the (X¯) Chart to monitor the quality level and a (Rm) Chart to monitor the process variability. Therefore, the general objective of this work is to investigate the performance of the batch control chart for the mean in the presence of changes in the mean and variance of the process. As a research methodology, a database taken from Siqueira’s dissertation (2017) was used, referring to the concentration of a certain stored chemical, which is a herbicide named in the research as Active A 95%. In his work, it was presented an application of the batch control chart, however, did not address its performance. In this way, the present work carried out simulations in the process from changes in the mean, standard deviation and in the mean and standard deviation simultaneously and later calculated the probability of detection of failures and the average number of samples or inspections until the detection of the first failure in the process. As a result, the batch control chart showed good performance in detecting failures in the presence of changes in production.