Volume 78, Issue 6 -

The calibration of a reference measurement standard is described for contour measurements using an micro-coordinate measuring machine F25. First results are presented for calibrating working contour standards with the aid of reference contour standards and contour measuring devices.

This paper presents an improved algorithm for estimating line segments in noisy range data. An estimator´s performance depends on correctly taking the measurement errors into account. Most sensors measure range and angle independently and thus it is a sound assumption to model the measurement error by independent noise processes in radius r and azimuth angle φ . However, this leads to correlated errors in Cartesian coordinates. Simple estimators do not consider this and therefore yield poor results. Two more recent approaches based on Maximum-Likelihood-Estimation are discussed in this paper and an improved formulation is presented.

We present a method for the simultaneous measurement of crack tip position and stress intensity factors (SIF) at cracks in thin-walled, plain or curved structures under fatigue crack growth. As a condition the structures consist of an isotropic material that complies with the concept of linear elastic fracture mechanics. For the measurement an orthotropic piezoelectric polyvinylidene fluoride film (PVDF film) is applied on the cracked area of the structure. The electrical values of the piezoelectric film signal the position and loading of the crack. By the electrical signals measured the determination of crack position and SIF results from the solution of the corresponding inverse problem. The method is tested hypothetically on the example of a crack in an infinite plate. Beyond the previous publication [1] now we are able not only to determine the exact solution of the K I -factor but also of the K II -factor.

This article focuses on inpit-output (IO)-linearized measurement chains with uncertain parameters and nonlinear characteristics. An analytic yet concise form of the measurement error is derived consisting of a parameter perturbation term and a linearity deviation for each subsystem only. Hence, the influence of nonlinearities on errror propagation is quantified without tedious analytic transformations. This can also be used for system design since changing the form of chararacteristics can offer an alternative way to meet a given desired measurement error which may be easier than reducing tolerances.

With the guide GUM (Guide to the Expression of Uncertainty in Measurement) a framework is set to describe uncertainties internationally comparably. An example shows, how measurement uncertainty can be determined from data sheets in accordance with the rules of GUM.