410
we were not aware of any such standards. However, in the course of writing this paper, we discovered an effort led by James S. Bigelow (1992) of Exxon Chemical Company, Polymers Group, based in Houston, Texas. He is chairing an ANSI-Z1 Committee on Quality Assurance that is currently developing exactly such a standard. At the time of this writing, the tentative title of Draft 0.3 of that standard is 'Standard Method for Calculating Process Capability and Performance Measures.” Besides establishing a standard for process capability, the proposed standard also discusses process performance. Process performance indices represent a measure of historical performance and generally cover longer time frames than capability indices.
In this section we will discuss the concept of process capability and review the commonly used process capability indices of Cp, Cpl, Cpk, ,
and C^. The major elements of process capability, and thus of the capability
indices, are the product specifications and the process characteristics. Neither of these is known with certainty, and care must be taken in the definition of each.
Product specifications are usually given in terms of the limits (USL and LSL) and the target (7). These specification limits, presumably the result of customer driven needs, are initially stated in terms of the product performance requirements. Specific component or subsystem specifications are then deveIoped to meet these objectives and are sometimes deve!oped through statistical tolerancing calculations. Thus, specifications should not be viewed as “written in stone,” but rather as negotiated values.
Whereas the specification limits are derived from customer needs, the process characteristics are dependent upon the process producing the product. The process characteristics are often given in terms of the process mean, p, and the process variability or reproducibility, which is stated in terms of a standard deviation, a. These characteristics are measured on products produced by the process. The interval p ± 3a is often used to characterize the process. The
length of this interval, 6ct, is termed the natural process tolerance. The statistical definition of the mean and the standard deviation are readily understood by most quality practitioners. However, their interpretation in the context of process capability is not straightforward.
Juran and Gryna (1988, page 16.14) define process capability as: the measured, inherent reproducibility of the product turned out by a process. He