Goodness-of-fit Tests Comparison for Statistical Process Control in an Automotive Industrial Unit

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

This paper aims to evaluate whether if the data that form several samples used for the statistical process control (SPC) control charts derive from a population with a normal distribution or not. A piece manufactured in a Portuguese small and medium-sized enterprises (SME) that operates in the automotive industry is used as an example. Knowing if the distribution is normal or not allows identifying what out of control tests should be applied and can also help finding precise false alarm rates. For this purpose, six goodness-of-fit tests are used and then compared. Some of these goodness-of-fit tests could be more sensible than others in detecting departures from normality. The results for two in three scenarios of the same dimensional feature show that some goodness-of-fit tests reject the null hypothesis and that the data of the measured samples do not derive from a population with a normal distribution.

Original languageEnglish
Title of host publicationICITM 2020 - 2020 9th International Conference on Industrial Technology and Management
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages161-165
Number of pages5
ISBN (Electronic)9781728143064
DOIs
Publication statusPublished - Feb 2020
Event9th International Conference on Industrial Technology and Management, ICITM 2020 - Oxford, United Kingdom
Duration: 11 Feb 202013 Feb 2020

Conference

Conference9th International Conference on Industrial Technology and Management, ICITM 2020
CountryUnited Kingdom
CityOxford
Period11/02/2013/02/20

Keywords

  • automotive industry
  • control charts
  • goodness-of-fit test
  • quality control
  • statistical process control

Fingerprint

Dive into the research topics of 'Goodness-of-fit Tests Comparison for Statistical Process Control in an Automotive Industrial Unit'. Together they form a unique fingerprint.

Cite this