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

580

6

Pattern AP5.1 is in agreement with the values measured in the magnet speckle pattern analysis. Patterns AP5.2, AP5.3 and AP5.4 show slightly lower values than nearby measurements from other speckle patterns, in accordance with the sti ff ness provided by the wall’s structural supports and vertical bar. Considering the deformation measure ments from these three adhesives, it is possible to extrapolate a point of low deformation towards the intersection of the base of the wall and the vertical bar, a region expected to have high sti ff ness. Pattern A5.5 measured the frontal face of the vertical bar, which shows a lower deformation than other speckles in this area, approximately 1.568 mm, possibly due to being closer to the base of the wall. Pattern AP5.6 was placed on the panel, near the bar and AP2.5 of the previous DIC analysis, both showing similar values.

3.2. Terrestrial Laser Scanner

The FARO FOCUS S TLS was also used to obtain point clouds of the transformer wall. The data was subsequently processed using FARO’s software, SCENE to export and register the point clouds on the same referential system. The CloudCompare software was then used for comparing the point clouds. Figure 5 shows a point cloud of the tank wall, with the ROI highlighted in red.

Fig. 5: Tank wall point cloud with ROI highlighted in red.

Due to factory environment constraints, the wall was scanned at an oblique angle, which resulted in partial occlu sion of the right side of the wall by the vertical bar. The frontal face of the vertical bar was also not correctly captured, likely due to suboptimal laser reflection on that surface at that angle, relative to the TLS. As explained in the following subsection, the simulation provided values for certain select points, some of which lie in areas that were occluded or poorly sampled by the TLS. Consequently, the TLS measurement point for comparison with simulated data was taken in the central section of the left side of the wall, where deformation is expected to be highest. The deformation measured by the TLS at this point was 3.518 mm, which is lower than the corresponding DIC measurement. This discrepancy could be the result of environmental vibrations and personnel access through the area where the sensors were acquiring data. In addition, due to the alternative speckle pattern implementations, any potential relative motion between the speckles and the tank wall could aggravate the measurement di ff erences.

3.3. Measurement comparison

The simulation was conducted by EFACEC, where the base of the tank was considered to be simply supported and the pressure set at the test’s operating conditions. From the simulation, three points were selected: the middle section