On The Control Charts For Gradational Qualities In High Yield Processes

One most important feature of the modern production processes is the high conformity rate in the production line, which is especially exemplified by the production processes of modern electronic products. Generally speaking,such kind of process is called high yield processes or high quality processes in the statistical process control. The quality control of the high yield processes is very important issue in the modern production processes. It has an instructional significance for the productional practices. However, how to break formerly bisection method of product quality (the product quality classified into conforming and non conforming) so as to make the product quality be more precise controlled during the quality control of high yield processes is a problem with much realistic meanings. This point is this thesis concern to.When applying traditional control charts to the high yield process, there are many flaws and deficiencies in the traditional P- chart or NP- chart,such as the large sample size or the concavo convex undulating control lines due to a lot of changes in samples,which will bring some unnecessary limitation to the control chart or will lead to the control chart invalidation. To rid of such deficiencies of the P-chart and NP-chart ,many researchers presented the geometric control chart, negative binomial control chart, geometric CUSUM control chart, geometric EWMA control chart with the number of consecutive conforming items. When utilizing these control charts, the products will be classified into two grades:conformity and nonconformity. However, in many reality cases, the product quality doesn't turn from conforming into nonconforming abruptly. It is a gradually changing process not an abrupt alteration. The qualities levels of the much products are somewhat between conforming and nonconforming in a high quality process. In order to make the strictly distinguishing and satisfying the needs for precisely controlling of product quality, this thesis brings forward to classify the product quality into three grades. At first, we introduce the several control methods and propose four kinds of three grade control charts. And finally, make the deep analysis to the evaluation standard of three grade control charts. This thesis also give a basic introduction to geometric control charts, analyze the application of these control charts, and by virtue of statistics software, such as mimitab, matlab to construct the charts, and discuss the process capability index of a high-yield process.