ASME B89.4.10:2021 pdf download

ASME B89.4.10:2021 pdf downloadMethodsforPerformance Evaluation of Coordinate Measuring System Software
1 SCOPE
A critical issue in industrial coordinate metrology is the measurement ofa work piece to assure compliance with its dimensional requirements. When using a computerized coordinate measuring system (CMS), the usual practice is to correlate computer-calculated outputs with the dimensional requirements of the workpiece. This correlation is performed by various computer routines that process dimensional coordinate data sets consisting of measurement samples of the object being evaluated. The purpose ofthis Standard is to provide guidelines forevaluatingthe qualityofsolutions generated byCMS software and to define minimal documentation requirements for software providers. Additionally, this Standard gives default definitions for collections ofdata sets that span a variety ofreal-world measuring scenarios. These data sets are depen- dent on the fitting algorithm being tested. This Standard is concerned with testing the behavior of algorithm imple- mentation,notthetestingofalgorithms themselves. Thus,thesoftwareistreatedas ablackbox; onlytheinputandoutput are observed and evaluated. Itis notthe intentofthis Standard to endorse or rate any computational method or system. Software performance evaluation is useful because it (a) allows objective validation of software (b) reduces the possibility of error in software application (c) defines a method of comparing CMS software This Standard covers the following areas: input data, feature construction, software documentation, performance characterization, and test methodologies.
1.1 Assumptions
The assumptions inherent in this Standard are as follows:
(a) Measurement uncertainty in coordinate samples is not addressed.
(b) Methods to input predetermined samples to the computational system are available.
(c) Personnel have adequate experience and trainingto implementthe evaluation and understand the implications of the results.
1.2 Application
This Standard is one component required for the evaluation of CMSs. Other relevant documents can be found in Nonmandatory Appendix E.
1.3 References
The following is a list ofstandards referenced in this Standard. Unless otherwise noted, the most recent edition shall apply.
ASME Y14.5, Dimensioning and Tolerancing
ASME Y14.5.1, Mathematical Definition of Dimensioning and Tolerancing Principles
Publisher: The American Society of Mechanical Engineers (ASME), Two Park Avenue, New York, NY 10016-5990 (www.asme.org)
See Nonmandatory Appendix E for additional, informative references.
coordinate measuring system (CMS): any piece ofequipment that collects coordinates (points), calculates, and displays additional information using the measured points. datum: a theoretically exact point, line, or plane derived from a feature on a part. See ASME Y14.5M-2018. least-squares fit feature: a feature ofperfect form, corresponding to a set ofdata points, that minimizes the sum of the squared deviations between the feature and the individual data points. (Reference Nonmandatory Appendix C for addi- tional information.) This term is elsewhere sometimes referred to as the Gaussian associated feature. NOTE: In this Standard, unless otherwise indicated, the least-squares fitis understoodto be notweighted, i.e., each pointis given equal weight in the least-squares objective function, even if the points in the test data are not exactly evenly spaced. objective function: a function which is to be optimized by searching for a minimum (or maximum) as its parameters are varied.Adifferentobjectivefunctionisusedforeachtypeoffit,e.g.,aleast-squaresversusminimum-circumscribedcircle. reference evaluation: the evaluation of the substitute feature using a known implementation of an algorithm. reference feature: a substitute feature used as the basis for evaluating a test feature. substitutefeature: afeature ofperfectgeometricformthatcorresponds to asetofdata points andis intendedto minimize an objective function. test: a basic unit of evaluation, based on one or more related data sets, which are applied to one or more software implementations of an algorithm. test feature: a substitute feature computed by the software under test.