Logo of the Physikalisch-Technische Bundesanstalt
Production sequence of Si-spheres and interferometrical determination of the sphere volume

New methods for determining measurement uncertainty for tactile coordinate measurements


The EMPIR project „Standards for the evaluation of the uncertainty of coordinate measurements in industry” (17NRM03 EUCoM) has the goal of developing new methods for determining measurement uncertainty and is about to finish. The new methods A and B (analogous to type A and B uncertainties from GUM – Guide to the expression of Uncertainty in Measurement) should offer new ways of estimating the measurement uncertainty of a tactile coordinate measurement on a workpiece.

Schematic representation showing the workpiece being rotated by 90° about each major axis, for subsequent measurements
Figure 1: Measurement of a workpiece in several orientations for method A

Method A (“a posteriori”) takes an empirical approach: The uncertainty is based on a large amount of measurement data. The workpiece is measured in several different orientations (Figure 1), to estimate the effects of repeatability and geometry of the coordinate measuring machine (CMM). In addition, length and sphere standards are measured to evaluate scale and probing errors. (Figure 2) This has the advantage of being able to estimate measurement uncertainties without the need for modelling of CMM, workpiece, etc. The trade-off is a relatively work-intensive measurement.

Column chart showing, as an example, a breakdown of a method A estimates, highlighting the different components
Figure 2: Individual uncertainty contributions from method A. Each example is taken from an independent measurement on a different coordinate measuring machine

Method B (“a priori”) includes two separate approaches (B1 and B2). Both are based solely on modelling, and use prior knowledge to derive the uncertainties, for example machine specifications, normed tests and expert knowledge. Among others, this includes MPE values (Maximum Permissible Error), nominal points and vectors (method B1) or measurements of the length measurement error E according to ISO 10360-2 (B2)

The methods are currently being validated as part of the project. The 14 project partners and collaborators have carried out measurements on several workpieces. The workpieces cover a range of different features, including freeform surfaces. There are also qualitative differences between workpieces and the CMMs used. All of this was done to ensure that the results are widely applicable, particularly in an industrial setting. The individual uncertainties are estimated based on these measurements and the associated specifications. The different method uncertainties are compared to each other and to independent, established estimation approaches (e.g., the virtual coordinate measuring machine, VCMM). The aim is to test whether these uncertainties are comparable – both with different measurements and different methods – and how robust they are. If validation is successful, these methods are also to be included in the ISO 15530 series of standards.



Physikalisch-Technische Bundesanstalt
Bundesallee 100
38116 Braunschweig