Consistent process monitoring criterion based on control data for an alternative attribute for narrowed tolerance

Abstract

The process capability indices C p , C pk are widely used in the modern quality management as statistical measures of the ability of a process to produce output X within specification limits. The customer's requirement to ensure C p ≥ 1.33 is often applied in contracts. Capability indices estimates may be calculated with the estimates of the mean m and the variability 6s, and for it, the quality characteristic in a sample of pieces should be measured. It requires, in turn, using advanced measuring devices and well-qualified staff. From the other hand, quality inspection by attributes, fulfilled with limit gauges (go/no-go) is much simpler and has a higher performance, but it does not give the numerical values of the quality characteristic. The described method allows estimating the mean and the variability of the process on the basis of the results of limit gauge inspection with certain lower limit LCL and upper limit UCL, which separates the pieces into three groups: where X < LCL, number of pieces is n 1 , where LCL X < UCL, n 2 pieces, and where X ≥ UCL, n 3 pieces. So-called Pittman-type estimates, developed by the author, are functions of n 1 , n 2 , n 3 and allow calculation of the estimated m and s. Thus, C p and C pk also may be estimated without precise measurements. The estimates can be used in quality inspection of lots of pieces as well as in monitoring and control of the manufacturing process. It is very important for improving quality of articles in machining industry through their tolerance.