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

582

8

Alvium U-508c Allied Vision cameras with 25 mm MeVis-C Linos lenses captured images for DIC analysis. Since it was not possible to paint the speckle pattern on the surface of the wall, two alternative targets were tested: speckle patterns printed on adhesive paper and composed of magnets. In addition, a phase-shifting TLS (FARO FOCUS S) was used to obtain point clouds of the wall, from which deformation was measured. The simulated data considered the base of the tank to be simply supported, and the pressure of the test’s operating conditions. Three points were extracted from the simulation to be compared to the optical measurements, however, due to the TLS sensor position, it was only possible to compare this sensor’s readings to the DIC and simulation in Point 1. Regarding the DIC, the magnet speckle pattern was used to measure deformation in Point 1 and Point 2’s area, while AP2.5 was used for Point 3. Both optical methods were capable of producing measurements in a factory environment. Despite the variation in speckle types and sizes, DIC results appear to be consistent and reflected the expected wall behaviour under near vacuum conditions. Excluding the absolute measurement di ff erence regarding DIC and TLS in the area of Point 1, which is 1.266 mm, all remaining absolute di ff erences are sub-millimetre. The highest discrepancy has an order of magnitude of 10 − 3 m, the smallest di ff erence, has an order of magnitude of 10 − 5 m (between the DIC and simulation, in Point 1) and the remaining di ff erences have an order of 10 − 4 m. Several factors may have contributed to the discrepancies between measurements, including: lower contrast be tween the black magnet speckle pattern and the grey wall; environmental vibrations in the factory; necessary personnel access through the area where the optical setups were installed, which could have altered sensor alignment; potential relative motion between the wall and DIC targets (adhesive paper and magnets) could also contribute to the relative errors; uncertainty in available space between the DIC optical setup and the wall, required the use of several pattern sizes which resulted in speckle diameters that were not optimal; and possible errors in the simulated data, however, considering the factory setting and the small absolute di ff erences observed, these approaches may represent a practical alternative to conventional speckle painting, although further testing is needed for full validation. Subsequent research on the implementation of these optical methods could benefit from calibrating the speckle pattern to a single diameter, selected according to the optical setup configuration. A systematic comparison between speckle patterns using magnets and small adhesive cutouts would help validate these alternatives. Additional tests in a controlled environment could validate the results without contributing to the relative errors; alternatively, a focused study on mitigating factory environment vibrations would be valuable. The DIC setup could be expanded to monitor all tank walls simultaneously. Similarly, the TLS (FARO FOCUS S) could be used to stitch point clouds of the entire transformer, enabling a full-body deformation analysis.

Acknowledgements

This work is a result of Agenda “ATE – Alianc¸a para a Transic¸a˜o Energe´tica”, nr C644914747-00000023, invest ment project nr 56, financed by the Recovery and Resilience Plan (PRR) and by European Union – NextGeneration EU.

References

[1] Nuruzzaman, M., Limon, G.Q., Chowdhury, A.R., Khan, M.A.M., 2025. Predictive maintenance in power transformers: A systematic review of AI and IoT applications. ASRC Procedia: Global Perspectives in Science and Scholarship 1, 34–47. [2] de Jesus Ribeiro, C., Marques, A.P., Azevedo, C.H.B., Souza, D.C.P., Alvarenga, B.P., Nogueira, R.G., 2009. Faults and defects in power trans formers — a case study. In: 2009 IEEE Electrical Insulation Conference , 142–145. [3] Dias, L., Ribeiro, M., Leita˜o, A., Guimara˜es, L., Carvalho, L., Matos, M.A., Bessa, R.J., 2021. An unsupervised approach for fault diagnosis of power transformers. Quality and Reliability Engineering International 37, 2834–2852. [4] Abbas, M.A., 2022. Time saving strategy for finite element analysis of transformer tank. International Journal of Mechanical And Production Engineering 4, 40–44. [5] Hasan, M.M., Rahman, A., Islam, A., Khondoker, M.A.H., 2024. Strengthening transformer tank structural integrity through economic sti ff ener design configurations using computational analysis. Applied Mechanics 5, 717–730. [6] Ranade, A., Rajkumar, E., Gaikwad, S., 2016. Transformer tank optimization using design of experiments. Indian Journal of Science and Technol ogy 9. [7] Hasan, M.M., Islam, A., Khondoker, M.A.H., 2024. Computational analysis of di ff erent sti ff ener designs for power transformer tank walls. Engi neering Proceedings 76, 57.