Desenvolvimento de uma carta de controle combinada beta-cusum binomial para monitoramento de produtos não conformes

Abstract

Statistical process control charts are tools widely used to monitor the quality of products, processes and services. Proposals for Shewhart in 1924 are traditionally applied because of their ease of implementation and analysis. Over the years, various control codes are allowed in the Shewhart type, with increased performance or performance due to process automation and high quality product use. However, these letters have a disadvantage of not detecting small changes within productive processes, where the use of cumulative sum letters (CUSUM) proposed by Page in 1954 is more appropriate. When a magnitude of change within a process is not known, one of the solutions is an application of combined cards as characteristics of Shewhart cards with CUSUM cards. The objective of this work is the development of a combined control chart as a control chart features, with beta distribution and accumulated total attributes chart, based on binomial distribution. Binomial Beta-CUSUM combined chart development procedures and their application to discrete and continuous data are demonstrated. The results show that a combined chart performs similarly to the Shewhart-CUSUM Binomial chart when applied to discrete data and performs better when applied to continuous data. Thus, a proposed Beta-CUSUM Binomial letter is valid and is a new combined letter alternative for monitoring nonconforming items.