Uncertainty of digital image correlation measurements applied on structural concrete tests

Author: Georgios Zinas
Language: English

Abstract

Digital Image Correlation (DIC) is an optical measurement technique that provides continuous meas-urements of the displacement field for an area of interest within the tested specimen. Although the use of DIC for structural experiments is continuously growing, there are hardly any standardised pro-cedures to quantify the measurement uncertainty. This work aims at the systematic assessment of 3D-DIC uncertainty and its most influential parameters for large-scale experiments on concrete members.

Therefore, two general test procedures are chosen:
The zero-displacement-test quantifies the noise level in the displacement and strain field for a speci-men that does not exhibit any movement nor deformation. This testing procedure mostly delivers information on the influence of stable parameters from the test setup such as the lighting conditions, speckle contrast or random phenomena such as the Schlieren effect or vibrations.
The zero-strain-test determines the level of measurement uncertainty for increasing imposed dis-placements but no deformations of the specimen. This testing procedure can reveal biased systematic errors such as a bad calibration procedure, which lead to larger uncertainty with increasing displace-ments.
In the conducted zero-displacement-tests, the distribution of noise within the area of interest shows that the noise level is mostly dependent on the image contrast. The typical upper limit values of the noise level found in the literature (0.01 of pixel) are reached only under excellent image contrast and rise very significantly for moderate contrast (0.03 of pixel). Therefore, homogeneous lighting, lighting intensity and speckle quality are of most important to improve the measurement sensitivity.
The zero-strain-test emphasizes the importance of a good calibration, which needs a complete repre-sentation of the field of view and sufficient movement within the measuring volume to properly cali-brate the stereo-camera-system. A 0.05 pixel measurement uncertainty of the displacement is reached at an in place displacement of around 100 mm.
This Master’s thesis introduces the most common issues with measurement uncertainty using DIC. The most influential parameters from the test setup and random phenomena from the environment were detected and qualitatively examined. A final quantification procedure for the assessment of measurement uncertainty for DIC systems still requires further research.