|
Track 2 Tutorial 3
3:30 to 5:00 p.m. Thursday June 19, 2008
Using the Weibull Function to Validate
Product Life
Many products must pass a
reliability demonstration test to validate that they meet the life
requirement of their customers. A supplier must show that a very
large percentage of the product population will function properly
for a given period of time under severe, specified conditions. If a
lab test is used, questions quickly arise regarding how to properly
analyze a skewed failure distribution and how to handle a data set
that has some non failed items (suspensions). What should one do
with several failure modes? What if the test applies a mix of loads
over time? When production tools are moved to a new plant, is it
necessary to validate that the new plant makes the same or
equivalent product? This tutorial discusses how the Weibull
function and associated Weibull plots and contour plots can be
effective tools to answer these questions. Contour plots provide an
easily understood graphical method to identify either sizeable
differences or negligible differences.
Key Words: Weibull Analysis,
Life/Failure Distributions, Bp Percentile, Target Life, Timeline
Plots, Failed and Suspended Items, Weibull Probability Plots,
Contour Plots, Confidence Bounds
Jerry Roslund
GLR Consulting
Rochester Hills, Michigan |
 |
|
|
|
