Study And Implementation Of Control Charts For High Yield Processes

Nowadays companies have greatly improved their operation management and product and process quality, thus the statistical quality control of low defects or near zero-nonconformity processes becomes more and more important. In such high quality processes, how to choose efficient and accurate control charts in order to conduct online statistical process control and ensure the stability of production process, and improve product and process quality, becomes the most imminent task of quality engineers and research scholars.The traditional control charts such as np and n charts are greatly limited due to their requirement of large sample size, inability to detect process improvement and reaction latency. In order to solve such problems, a new type of control charts monitoring the number of consecutive conforming products are introduced and evaluated. Such charts include geometric chart, negative binomial chart, geometric CUSUM chart and geometric EWMA chart. Geometric charts are introduced on the basis of literature study and some basic settings and practical issues in its applications are brought up and analyzed. Then a detailed study of geometric CUSUM and EWMA is given and some suggestions on their practical implementations are brought up on the basis of numerical computation and computer simulation. At last, a comprehensive comparison is made on the efficiency of the five kinds of charts including geometric chart, negative binomial charts when the parameter n equals 2, geometric CUSUM and EWMA chart using the universal standard of ANI. Some suggestions are given which are useful for choosing the best control chart under different deviations of process nonconformity levels.