Section: New Results

Assessing metrological performance in experimental mechanics

Participant : Frédéric Sur.

In experimental mechanics, displacement and strain fields are estimated through the analysis of the deformation of patterns deposited on the surface of the tested specimen. Regular patterns such as grids are processed with spectral methods (the so-called "grid method"), and random speckle patterns are processed with digital image correlation (DIC). The scientific contribution obtained in 2017 concerns the comparison of the grid method and DIC. Since neither guidelines nor precise standard are available to perform a fair comparison between them, a methodology must first be defined. In [13], it is proposed to rely on three metrological parameters, namely the displacement resolution, the bias and the spatial resolution, which are not independent but linked.For the value of the bias fixed in the study of [13], the grid method features a better compromise than subset-based local DIC between displacement resolution and spatial resolution, in spite of its additional cost due to grid depositing. Work in progress concerns several aspects of DIC-based methods. In particular, we worked on synthetic speckle image rendering: ground truth databases are indeed crucial to assess the performance of the algorithms estimating displacement fields. It is, however, not straightforward to control any aspect of the rendering algorithm to ensure that performance assessment is not biased by the rendering algorithm. In addition, a popularization paper has been published in ERCIM News [15]. This work is part of a collaboration between Magrit project-team and Institut Pascal (Clermont-Ferrand).