PSI - Issue 82

A.T. Andreasen et al. / Procedia Structural Integrity 82 (2026) 146–152 A.T. Andreasen et al. / Structural Integrity Procedia 00 (2026) 000–000

150

5

will provide the most robust results due to the local gauge being very sensitive to the high stress gradient near the weld. The framework is developed in MATLAB and in Fig. 4 a flowchart of the framework is given.

Fig. 4. Flowchart of the framework.

5. Results The framework has been used with the experimentally obtained strain gauge data to estimate the loading occurring for the 10 different test series. To validate the framework, the determined loading has been used to predict the strains at the strain gauge positions, based on the FE model. These strains are predicted using Eq. (2) in combination with the determined loading { /0 } and the influence matrix [ /0 ] , thus returning the FE-predicted strains { /0 } . The accuracy of the load estimation framework can then be approximated by comparing the experimentally obtained strains { 1&2 } to the predicted strains { /0 } using the load estimation framework. In Fig. 5, the predicted strains from the FE-mode and experimentally obtained strains are compared for strain gauge 4 and test number 2. In Fig. 5(a), the predicted strain follows the trends of the full signal, while Fig. 5(b) shows a zoomed-in picture of the same graph, where the two signals overlap. As seen from the figure, the predicted and measured signal show the same tendencies and when examining the close-up view, it is obvious that the predicted strains are very close to the measured strains, indicating that the load estimation framework is very accurate. It should be noted that the comparison in Fig. 5 is made for a global gauge, which is also used for determining the loading itself. Thus, it is to be expected that the signals are very close, if the FE model and experiment are relatively comparable.

Fig. 5. (a) Full signal of global strain for test 2; (b) Zoomed in signal of global strain for test 2.