A COMPARATIVE STUDY OF CONTROL CHARTS FOR MONITORING LOW NONCONFORMING PROPORTION

Ângelo Márcio Oliveira Sant’Anna
Department of Mechanical Engineering, Polytechnic School, Federal University of Bahia, Salvador/BA,
40210-910, Brazil. Email: angeloms@usp.br

Abstract

Control charts are important tools to monitor the quality of products in several industrial processes. The traditional Shewhart p-chart for monitoring non-conforming data was constructed following the binomial distribution with normal approximation. However, this scheme suffers inaccuracy when there is low nonconforming item leading to errors on monitoring process. Studies on control charts for nonconforming data seek to enhance the binomial to normal approximation to reduce risks to get false alarms on monitoring. In this paper, a new p-chart based on the Beta distribution is developed, which can substantially replace the approximation proposed in the literature. The proposed p-chart is compared with several normal approximation-based charts and the performance analysis for monitoring low nonconforming proportion is illustrated with real studies.

Keywords
Statistical quality control, nonconforming data, Beta distribution.

Scope and Topics
Statistical Process Control; Quality Engineering

2015 QPRC