Statistical Detection of Outliers in the Certification of NIST Reference Charpy Lots

Published

June 3, 2024

Author(s)

Enrico Lucon

Abstract

This report describes the procedures used by the NIST Charpy Machine Verification Program to detect statistically significant outliers within the test results obtained from the certification of Charpy reference lots. The evaluation of outliers is based on an array of rigorous statistical procedures. The process begins with the screening of potential/suspected outlier data points for datasets assuming an approximately normal distribution, followed by the establishment of their statistical significance using multiple statistical tests. When this examination identifies a data point as a significant outlier, the corresponding Charpy test and specimen are subjected to in-depth examination, in order to establish whether a test- or material-related problem has occurred, and could justify discarding the test result before calculating the certified absorbed energy of the lot and its associated uncertainties, and establishing whether the lot should be accepted or rejected. The same strategy can also be applied to suspicious values reported by customers. Practical examples are illustrated in appendices.

Citation

NIST Interagency/Internal Report (NISTIR) - 8526

Report Number

8526

NIST Pub Series

NIST Interagency/Internal Report (NISTIR)

Pub Type

NIST Pubs

Download Paper

Local Download

Keywords

Certification tests, Charpy reference lot, NIST Charpy Program, statistical outlier, statistical significance.

Citation

Lucon, E. (2024), Statistical Detection of Outliers in the Certification of NIST Reference Charpy Lots, NIST Interagency/Internal Report (NISTIR), National Institute of Standards and Technology, Gaithersburg, MD, [online], https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=957454 (Accessed July 17, 2024)