Determining Measurement Uncertainties of Feature Measurements on CMMs (Bepalen van meetonzekerheden bij het meten van vormelementen met CMMs)

The dimensions of technical products are often of major importance for t heir proper functioning. Dimensional tolerances define how much the true dimensions may deviate from the aimed dimensions. Dimensional quality c ontrol ensures that the dimensional properties of the product comply wit h the tolerances. For single and non-complex products instruments like v ernier calipers and micrometer screw gauges can be used, while for more complex inspection tasks and series measurements usually coordinate meas uring machines (CMMs) are used.Because measurement devices also show measurement errors, a measurement result will never be exact. This means there will always be a measurement uncertainty that should be take n into account while evaluating conformance of products to tolerances. I f measurement uncertainty is neglected, this can result in false rejecti on or false acceptance of products, with possibly far-reaching consequen ces. Reliable measurement uncertainties are indispensable for unambiguou s evaluation of tolerances.Measurement uncertainty determination for coordinate measuring machines is difficult because of the many unce rtainty contributors (CMM hardware errors, temperature, measurement stra tegy, ...) that are involved. Therefore conventional uncertainty calcula tion methods, based on analytical propagation of standard uncertainties, can not be applied. Measurement uncertainty determination based on Mont e Carlo simulations is a valuable alternative to conventional uncertaint y calculation methods. These methods can cope much better with the compl exity of CMM measurements.This thesis describes four contributio ns to the determination of measurement uncertainties for feature measure ments on CMMs by means of Monte Carlo methods: When considering measurement uncertainties for feature measurements on C MMs, the influence of feature form deviations is often neglected because feature form deviations are not known in advance and hence difficult to model or to account for. Form deviations of features, like e.g. circles , lines, planes and cylinders, can have an important influence on the me asurement uncertainty, particularly if only a limited set of points is s ampled. Neglecting the influence of form deviations results in less reli able or wrong uncertainty statements, that may lead biased acceptance or rejection of parts. This thesis describes a method to determine the inf luence of feature form deviations on the measurement uncertainty. The measurement accuracy of a CMM varies over the measurement volume. Th is is mainly related to the geometric errors of the CMM in combination w ith the Abbe-offsets to the scales. This thesis presents a method to mod el the geometric errors of a CMM. It is based on the kinematic model of the CMM and simulated motion errors of the different axes. In contrast t o other approaches the proposed method does not need an error mapping of the CMM. The only input needed for this method are the scale positions and a valid ISO 10360-2 specification of a CMM. This thesis also elaborates an uncertainty determination method to combi ne the influence of feature form deviations and CMM hardware errors on t he measurement uncertainty for feature measurements. This method relies on a two-step Monte Carlo method: a first step to determine the form dev iation uncertainty, a second step to determine the uncertainties of the other feature parameters. Measured features are often used to construct part coordinate systems (P CS) or other features (e.g. a constructed point as intersection of two m easured lines). This means there will also be uncertainties associated w ith part coordinate systems and constructed features. Similarly there wi ll also be uncertainties related to measured geometrical tolerance value s (perpendicularity, parallelism, coaxiality, ...). Uncertainty evaluati on software (UES) for CMMs should take into account these derived uncert ainties. This thesis describes the implementation of an uncertainty eval uation software that takes into account the influence of feature form de viations and CMM hardware errors on measurement uncertainties of feature s and derived parameters.status: publishe