1st Edition

Statistical Process Control for Real-World Applications

By William A. Levinson Copyright 2011
    272 Pages 136 B/W Illustrations
    by CRC Press

    The normal or bell curve distribution is far more common in statistics textbooks than it is in real factories, where processes follow non-normal and often highly skewed distributions. Statistical Process Control for Real-World Applications shows how to handle non-normal applications scientifically and explain the methodology to suppliers and customers.

    The book exposes the pitfalls of assuming normality for all processes, describes how to test the normality assumption, and illustrates when non-normal distributions are likely to apply. It demonstrates how to handle uncooperative real-world processes that do not follow textbook assumptions. The text explains how to set realistic control limits and calculate meaningful process capability indices for non-normal applications. The book also addresses multivariate systems, nested variation sources, and process performance indices for non-normal distributions.

    The book includes examples from Minitab®, StatGraphics® Centurion, and MathCAD and covers how to use spreadsheets to give workers a visual signal when an out of control condition is present. The included user disk provides Visual Basic for Applications functions to make tasks such as distribution fitting and tests for goodness of fit as routine as possible. The book shows you how to set up meaningful control charts and report process performance indices that actually reflect the process' ability to deliver quality.

    Traditional Control Charts
    Variation and Accuracy
    Statistical Hypothesis Testing
    Control Chart Concepts
    Setup and Deployment of Control Charts
    Interpretation of x‑bar/R and x‑bar/s Charts
    X (Individual Measurement) Charts
    Average Run Length (ARL)
    z Chart for Sample Standard Normal Deviates
    Acceptance Control Chart
    Western Electric Zone Tests
    Process Capability and Process Performance
    Goodness-of-Fit Tests
    Multiple Attribute Control Charts

    Nonnormal Distributions
    General Procedure for Nonnormal Distributions
    The Gamma Distribution
    The Weibull Distribution
    The Lognormal Distribution
    Measurements with Detection Limits (Censored Data)

    Range Charts for Nonnormal Distributions
    Traditional Range Charts
    Range Charts with Exact Control Limits
    Range Charts for Nonnormal Distributions

    Nested Normal Distributions
    Variance Components: Two Levels of Nesting

    Process Performance Indices
    Process Performance Index for Nonnormal Distributions
    Confidence Limits for Normal Process Performance Indices
    Confidence Limits for Nonnormal Process Performance Indices

    The Effect of Gage Capability
    Gage Accuracy and Variation
    Gage Capability and Statistical Process Control
    Gage Capability and Process Capability
    Gage Capability and Outgoing Quality

    Multivariate Systems
    Multivariate Normal Distribution
    Multivariate Control Chart
    Deployment to a Spreadsheet: Principal Component Method
    Multivariate Process Performance Index
    Control Charts for the Covariance Matrix

    Appendix A: Control Chart Factors
    Appendix B: Simulation and Modeling
    Appendix C: Numerical Methods


    William A. Levinson is the owner of Levinson Productivity Systems in Wilkes-Barre, Pennsylvania.

    With this book, the author provides a useful contribution to the literature for dealing with methods for describing the quality of a nonnormal process through the use of appropriate control charts and/or process performance indices. … I found this book interesting and valuable in discussing topics related to SPC that are not found in traditional textbooks on the subject. In order to get the same level of information as provided in this book, one would have to search for numerous journal articles for details and applications.
    —Connie M. Borror, The American Statistician, November 2011

    … a useful addition to the library of any industrial statistician, process engineer, or quality engineer engaged in improving processes. It represents material that is not typically covered in conventional books on statistical process control. In particular, with its focus on nonnormal distributions, this text provides a measure of reality that many quality and manufacturing professionals have known for years—that not all process distributions follow the normal distribution. In that respect, this book is a refreshing addition to the literature and I highly recommend it.
    —Dean Neubauer, Journal of Quality Technology, Vol. 43, No. 4, October 2011