PSI - Issue 77

Francisco Afonso et al. / Procedia Structural Integrity 77 (2026) 575–583 F. Afonso et al. / Structural Integrity Procedia 00 (2026) 000–000

578

4

A FARO FOCUS S phase-shifting TLS was also used to obtain point clouds of the transformer wall, by sampling points in the sensor’s environment.

3. Results and discussion

An important step after manufacturing transformer tanks is the verification of its structural behaviour. Depending on the transformer type, this can be assessed with a vacuum test, in which the tank is subjected to very low (near-vacuum) pressure for several hours. An FEA simulation of the tank structure is typically performed beforehand to anticipate the wall deformation values. A vacuum test was conducted and the deflection on one tank wall was monitored using the aforementioned optical setups, until the tank reached near-vacuum pressure. The results are then compared to three points extracted from simulated data. Several speckle pattern targets were placed on the tank wall, enabling DIC analysis. The targets used include two types of black speckle patterns printed on white adhesives, one with a diameter of 2 mm and another with 5 mm; additionally, black 12 mm and 8 mm diameter cylindrical-shaped magnets were attached to the wall in a speckle pattern configuration. These targets were positioned at multiple locations within the ROI. The Allied Vision cameras captured images at the beginning of the test and when the tank pressure was near-vacuum, these images were analysed using the VIC-3D 9 Software. Figure 2 shows the out-of-plane deformation results obtained from the speckle pattern composed of 8 mm and 12 mm magnets, this analysis was conducted with a subset size of 55, and a step size of 3. 3.1. Digital Image Correlation

Fig. 2: Digital image correlation with magnets, speckle pattern results.

As expected, the highest deformation values were recorded in the wall section between its outer edge and the vertical bar, which corresponds to the farthest area from structural supports, with approximately 4.784 mm measured. The lowest values appear close to the main structural support near the edge of the wall (approximately 0.524 mm) and near the vertical bar (approximately 2.171 mm), which also provides structural support. On the frontal face of the vertical bar, approximately 1.855 mm were measured, this section is reinforced by the bar, increasing sti ff ness and reducing the deformation. This analysis was then repeated for the adhesive targets; DIC analysis was unable to be performed on the adhesive target placed on the side of the vertical bar, likely due to this position resulting in lower quality speckle images. For clarity, each speckle pattern was named, Figure 3 shows the results for the 2 mm speckle pattern printed on adhesive paper, this analysis was performed with a subset size of 31, and a step size of 3.