Process Capability Analysis, and What to Do if it's Not a Bell Curve

Why Should You Attend:

Customers often request process performance metrics from their suppliers, and this may in fact be mandatory in the automotive sector (IATF 16949:2016 clause 9.1.1.1). Even if it is not mandatory, it is extremely useful to assess the process' ability to meet specifications. It is therefore vital to know more than just the mathematical mechanics of their computation. These metrics can be highly misleading, and even off by orders of magnitude in terms of the process' nonconforming fraction, because of:

• Incorrect selection of the rational subgroup (a sample that reflects all variation sources in the process). As an example, within-batch variation, as might be measured by a sample from a single lot or batch, does not reflect between-batch variation.
• A non-normal process distribution. The traditional calculation relies on the assumption that the process data conform to a bell curve.

This webinar will cover not only the traditional calculations and their practical implications, but also how to handle the frequent real-world situations in which the bell curve assumption is not met.

Areas Covered in the Webinar:

1. Relationship between variation and accuracy, and process yield.
• Accuracy is the degree to which the process is centered on the nominal, which is usually halfway between the specification limits for a normally distributed process.
• Variation is the spread in the critical to quality (CTQ) characteristic. Less is better, just as a rifle has less variation than a musket.
2. Short term versus long term variation
• Short term variation occurs, for example, in the same production lot or batch where all the parts are subject to essentially identical conditions. This variation is the source of the process capability index.
• Long term variation such as between-lot variation is not accounted for by the short term variation estimate. It, plus the short term variation, is reflected in the process performance index.
• If the process capability index is significantly larger than the process performance index, the rational subgroup (a sample that reflects all the variation sources in the process) has not been selected correctly.
3. Non-normal distributions
• The traditional calculations for capability and performance indices rely on the assumption that the process follows a normal or bell curve distribution. Application of the same mathematical formulas to non-normal data will deliver grossly inaccurate results.
• Off the shelf methods are however available with which to calculate meaningful performance indices for these processes.
4. How to perform a process capability study
• The process must be in control.
• The rational subgroup must be selected correctly.
• An adequate data base is mandatory.
• Always test the assumption that the data follow the normal (or other selected) distribution before reporting results.

Who Will Benefit:

The following titles across the manufacturing industry will benefit from this training:

• Manufacturing
• Production
• Quality
• Engineering
• Product Management
• Project Management
• Technician
Instructor Profile:



William Levinson
Principal Consultant, Levinson Productivity Systems

William A. Levinson, P.E., FASQ, CQA is the principal of Levinson Productivity Systems, P.C. , and a leading authority on Henry Ford's business system; later adopted by Japan as the Toyota production system. He is also the author of The Expanded and Annotated My Life and Work: Henry Ford's Universal Code for World-Class Success.

Topic Background:

Process capability and process performance indices are vital process metrics that are often submitted to customers. The traditional calculation of these indices, however, relies on the assumption that the process follows the normal or bell curve distribution. It is therefore vital to identify the underlying statistical distribution of the critical to quality (CTQ) characteristic before reporting a process performance index.

Follow us :