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

581

7

of the left side of the wall, between the vertical bar and the outer edge of the wall, where deformation is expected to be highest (Point 1); on the frontal face of the vertical bar (Point 2), and on the panel, near the vertical bar (Point 3). These points and their respective measurements are shown in Figure 6.

Fig. 6: Simulation results.

Point 1 shows a measurement of approximately 4.204 mm, Point 2, 1.850 mm and Point 3 reports 2.302 mm. Table 1 shows the final measurement of each method, per DIC, TLS and the simulation, for the areas of the three points extracted from the simulation. DIC measurements were obtained using the magnet speckle pattern for Point 1 and Point 2, while pattern AP2.5 was used for Point 3.

Table 1: Point 1 measurement comparison.

DIC[mm]

TLS [mm]

Simulation [mm]

Point 1 Point 2 Point 3

4.784 1.855 1.727

3.518

4.204 1.850 2.302

— —

The highest absolute di ff erence is between DIC and TLS measurements in Point 1, reaching 1.266 mm. However, excluding this value, the remaining absolute di ff erences are all sub-millimetre, in Point 2, the value measured by DIC was strikingly similar to the simulation, with an absolute di ff erence of 0.005 mm. Some factors that could have contributed to the discrepancies between measurements include: lower contrast between the magnet speckle pattern (black) and the background, which was grey; continuous vibrations from heavy machinery in the factory may have slightly shifted the optical setups during testing; limited space required personnel to pass through the area where the sensors were installed, potentially disturbing their alignment; since painting the surface was not possible, relative motion between the adhesive or magnet targets and the wall may have introduced additional measurement errors; due to uncertainty in the available space between the DIC setup and the tank wall prior to testing, several pattern sizes were used, resulting in speckle diameters that were not optimal; and potential errors in the simulated values. Despite these challenges, the highest absolute measurement di ff erence has an order of magnitude of 10 − 3 m (between DIC and TLS in Point 1), the lowest, 10 − 5 m (between DIC and the simulation in Point 2), and the remaining absolute di ff erences, an order of magnitude of 10 − 4 m.

4. Conclusions and future works

This paper presents a study of two optical methods applied during pressure testing of a transformer tank wall, all measurements were subsequently compared, including simulated data (Table 1). A stereo optical setup using two