When Zero Isn't Zero; How to Handle Lower Detection Limits

Why Should You Attend:

You will learn how to estimate process parameters (mean and standard deviation, or corresponding parameters of non-normal distributions that are common in these applications) when some of the data are below the lower detection limit of the gage or instrument.

Reliability studies must often deal with right-censored data, e.g. when items are tested to 1000 hours, or 20,000 cycles, and failure times or number of cycles are recorded for each item that fails. The analytical methods, which are readily available in statistical software packages, also account for the items that do not fail at all. While left-censoring is rarely if ever applicable in reliability studies, similar methods can be used to deal with situations in which, for example, an instrument cannot detect less than 0.2 ppm of an impurity or trace contaminant and returns a measurement of 0 or "not detected" under these circumstances.

Areas Covered in the Webinar:

• Awareness of real-world quality applications in which the gage or instrument cannot detect less than a certain amount of a trace impurity, contaminant, or pollutant
• Use reliability statistical methods for left-censored data to identify the parameters of the statistical distribution for the quality characteristic.
• The statistical distribution must be known from past experience or similar applications. It is unlikely to be a normal (bell curve) distribution.
• Perform tests for goodness of fit to the selected distribution.
• Use the parameters to:
• Calculate the nonconforming fraction of items above the upper specification limit, and convert this into a process performance index.
• Set the upper control limit and center line for a statistical process control chart for the quality characteristic in question.

Who Will Benefit:

• Quality managers, engineers, and technicians, and others with responsibility for application of statistical methods

William Levinson
Principal Consultant, Levinson Productivity Systems

William A. Levinson, P.E., is the principal of Levinson Productivity Systems, P.C. He is an ASQ Fellow, Certified Quality Engineer, Quality Auditor, Quality Manager, Reliability Engineer, and Six Sigma Black Belt. He is also the author of several books on quality, productivity, and management.

Follow us :