2016 - Webinar on Metrology: Statistical Analysis of Measurement Uncertainty

Description :
The seminar begins with an examination of the fundamental vocabulary and concepts related to metrology. Topics include: accuracy, precision, calibration, and "uncertainty ratios". Several of the standard methods for analyzing measurement variation are then described and explained, as derived from AIAG's Measurement System Analysis reference book. The methods include: Gage R&R (ANOVA method, for 3 gages, 3 persons, 3 replicates, and 10 parts), Gage Correlation (for 3 gages), Gage Linearity, and Gage Bias.
The seminar ends with an explanation of how to combine all relevant uncertainty information into an "Uncertainty Budget" that helps determine the appropriate width of QC specification intervals (i.e., "guard-banded specifications"). Spreadsheets are used to demonstrate how to perform the methods described during the seminar.
All manufacturing and development companies perform testing and/or inspection that involves measurements of products, components, and/or raw materials. The output of those measurements is compared to design or QC specifications, to determine whether or not the measurements "pass" those specifications.
However, all measurement processes have some inherent variability; that is, a given measurement will likely not be exactly equal to the true value, because of variation from a number of different sources. Some of those sources are: person to person, equipment to equipment, time to time, and calibration to calibration. How much trust to place in a given measurement can be quantified by determining the magnitude of each of those sources; in effect, the larger the uncertainty of the measurement (i.e., the greater the measurement variation, in comparison to the size of the design or QC specification interval), the lower the trust that should be placed in a given measurement.
If the measurement uncertainty can be quantified, it can be applied to reduce the width of the design/QC specifications, so that the resulting "guard banded" specifications can be used without concern for measurement variation.
Areas Covered in the Session :
Fundamental Vocabulary & Concepts
Gage Repeatability and Reproducibility (ANOVA method)
Gage Correlation
Gage Linearity
Gage Bias
Uncertainty Budgets and Guard-banded Specifications
Who Will Benefit:
QA/QC Supervisors
Process Engineers
Manufacturing Engineers
QC/QC Technicians
Manufacturing Technicians
R&D Engineers