DIC Experiment - Verifying the Robustness of Stereometric Grid Digital Image Correlation Software - In Collaboration with SAMWELL TESTING INC.

Keywords: Digital Image Correlation,Stereo Grid DIC, Displacement and Strain Measurement, Analysis Software Robustness

Main Text:

The quality of measuring displacement fields using digital image correlation depends on many factors, which can include:

•Internal to the software (algorithm implementation, shape functions used to describe displacement, image interpolation methods, etc.);

•External to the software (speckle pattern quality, camera resolution, calibration quality, etc.)

To verify the robustness of the digital image correlation software algorithms, we specifically compared the software results with other reference measurements obtained independently under the same test conditions (the same external factors of the software). In this process, we can use traditional digital image correlation software to compare the entire displacement field or use physical sensors to compare in localized areas.

To compare theEikoTwin DICresults with other gauges, a test was conducted atSamwell Testing Inc.testing laboratory. This test involved studying the bending of a cantilever plate bolted to a rigid frame. The load was applied at the center hole and increased with each new image.7.8g(see Figure1).

Figure 1 – CAD of the studied part (on the left) and loading principle scheme (on the right)

The test was conducted using a stereo correlation digital image device equipped with two12 MPcameras (4000×3000resolution). It was also equipped with a dial gauge and a non-contact displacement sensor to measure displacement at the same location (see Figure2).

Figure 2 – Photography of the sensors instrumentation during the test

To be able to compare all results, we created digital sensors in two digital image correlation software that were located at the same position as the two physical sensors (see Figure3).

Figure 3 – Location of the DIC sensors in both software

After post-processing the images with these two software, we exported the normal displacement of the part's surface measured at the virtual sensor locations. Now we can compare the displacements measured by the two physical sensors and the twoDICsensors during the test (see Figure4).

Figure 4 – Plot of normal displacement as a function of “time”

EikoTwin DICThe displacement results obtained from the software are surrounded by the results obtained from the two physical sensors. This confirms the robustness of the algorithms used in the software for this test.EikoTwin DICThe maximum deviation measured under normal displacement between the software and the reference software wasEikoTwin DIC15μm, the non-contact displacement sensor was18μm, and the dial gauge was19μm.The next step of this study may be to compare with simulations. Since the results are measured directly on the

FEEikoTwin DICmesh usingsoftware, this comparison will be faster and easier to achieve compared to using standard stereo correlation software. In this case, a cumbersome post-processing step is required, including the transformation of the frame coordinate system for measuring displacement. Furthermore, by combiningwith physical sensors, the repositioning of the sensors relative to the mesh will be easier than using traditional image correlation software (see FigureEikoTwin DIC). Finally, the measurements obtained using4EikoTwin Digital Twinsoftware can update the simulation model.To further validate the robustness of the software, more complex additional tests (multi-axial loading, large displacements, dynamic loading, etc.) can be conducted. Similarly, by comparing the values measured by digital image correlation with those measured by strain gauges, the comparison can be extended to the strain field to compare the performance of the second metric software.

为了进一步验证软件的稳健性，可以进行更复杂的其他测试（多轴加载、大位移、动态加载等）。类似地，通过将数字图像相关测量的值与应变计测量的值进行比较，可以将比较扩展到应变场，以便比较第二个度量的软件性能。