< back to main site

Publications

The comparison of algorithm for the assessment of Type A1 surface texture reference artefacts.

Forbes, A B; Harris, P M; Leach, R K (2003) The comparison of algorithm for the assessment of Type A1 surface texture reference artefacts. NPL Report. CMSC 33/03

[img] Text
cmsc33.pdf

Download (901kB)

Abstract

The dimensional assessment of a Type A1 surface texture reference artefact involves estimating the depth of a groove by fitting a pair of nominally parallel lines to measurement data and then calculating the separation between the lines. The evaluation of the uncertainty associated with this estimate is important and needs to take into account i) the statistical model for the measurement data including correlation between the measured x and z values and ii) the algorithm for determining an estimate of the depth. In this report, we analyse a range of least squares regression algorithms that can be applied to determining the geometry of these artefacts for a range of data uncertainty structures. The analysis shows that while for completely general uncertainty matrices the algorithms can produce markedly different results, for realistic data and uncertainty models, their behaviour is generally very similar. As a consequence, we can conclude it is safe to use the more straightforward approaches, so long as appropriate attention is given to the uncertainty associated with the fitted parameters. The report describes a number of analysis techniques and algorithm testing methods that have quite wide application.

Item Type: Report/Guide (NPL Report)
NPL Report No.: CMSC 33/03
Subjects: Engineering Measurements
Engineering Measurements > Dimensional
Last Modified: 02 Feb 2018 13:16
URI: http://eprintspublications.npl.co.uk/id/eprint/2899

Actions (login required)

View Item View Item