Keywords: crack detection, strain measurement, digital image correlation,Stereo grid DIC

Background

Some structures have a long service life from the start of service to replacement. These structures must be able to maintain their physical integrity throughout their entire service life while performing their basic functions. To achieve this, their reliability must be ensured. For structures like metal water pipes, different types of damage may occur during operation, but the most harmful is the opening and propagation of cracks in the pipes (Figure1). To prevent this irreversible damage and to intervene in a timely manner, numerical simulations have been conducted to make it possible to predict damage caused by pipe aging.

In this case, the experiments discussed in this paper focus on tracking the opening and propagation of cracks on bent specimens to register and verify the numerical behavior patterns of cracking in pipe materials.

Figure1–Different types of crack opening on cylindrical structures

Crack propagation tracking based on stereo correlation measurements

To demonstrate, pre-crack tests were conducted in the test pipe. A pair of cameras was arranged to observe the specified area (research area) and the propagation of cracks in the direction predicted by numerical simulations. The research area (green in Figure2) and pre-calibration points (red) were defined directly from the finite element mesh provided by the client (in our digital image correlation software) before loading.

The measurement results will be directly represented on the finite element mesh, allowing us to immediately compare them with the displacement field predicted by the simulation.

Figure2–Definition of speckle pattern and research area

Determination of crack tip

To track the evolution of cracks over time, we pay special attention to the calculation of residuals. In fact,EikoTwin DICcalculated residuals make it possible to quantify the overall correlation error. The residual corresponds to the difference between the reference image (first image) and the distorted image, and is corrected for the measured displacement field. It is calculated for each evaluation point, allowing us to"real-time"tracking of crack emergence.

Figure3Two consecutive images are shown, in which we can see the beginning of the crack (yellow at the pre-crack level) by measuring the display of residuals.

Figure3–Determination of the appearance of the crack tip in two consecutive images

The position of the crack tip is manually collected frame by frame, as shown in Figure4. This is the length of extension starting from the initial defect tip, with the position of that defect being prioritized. The operator uncertainty for this measurement is estimated to be about0.2 mm.

Figure4–Operator frame-by-frame measurement of crack tip

Crack propagation over time

With the help of maps, the evolution of cracks can be tracked image by image, and the coordinates of the crack tip over time can be recovered. By extracting the position image for each image, the length of the crack (relative to the length from the initial defect tip) can be measured across all images, thus tracking its evolution. Figure5shows a series of maps that allow visualization of the appearance and evolution of the crack from the primer to the end of the test. Figure6depicts the evolution of crack length over time.

Figure5–Residual maps at the start and end of the test

Figure6–Evolution of crack length over time

In summary, we can derive a large amount of data from this test.Stereo correlationcan provide three-dimensional displacement fields and deformation fields, and also allows for the detection, monitoring, and measurement of the appearance and propagation of cracks during specialized tests conducted on instrumented bending specimens through the study of maps. The obtained mappings enable clear visualization of the emergence and propagation of cracks during the test, and quantitatively measure their evolution.