PSI - Issue 25

Romanin Luca et al. / Procedia Structural Integrity 25 (2020) 149–158 Author name / Structural Integrity Procedia 00 (2019) 000–000

156

8

Table 1. Equivalent loads for the various type of joints. Flange Left

Flange Right

Web

ϵ

ΔT [°C/mm]

ϵ

ΔT [°C/mm]

ϵ

ΔT [°C/mm]

Flange 120 mm, web 80 mm, PJP Flange 80 mm, web 90 mm, PJP Flange 120 mm, web 90 mm, PJP

4.3E‐03 4.1E‐03 3.0E‐03

‐47.39 1.0E‐04 ‐75.14 2.8E‐03 ‐34.52 1.6E‐03

‐9.31 7.3E‐03 ‐60.78 1.2E‐03 ‐25.45 1.0E‐04

52.17

227.77 3.0E‐03

An open question is how to treat cruciform joints, the problem is intrinsic in the use of middle plane geometry with shell elements. In fact, the virtual weld bead region of the flange of the two joints corresponds to the same elements and two equivalent loads are to be applied on the same region. The better solution would be to create a reference cruciform joint, where a total of 8 parameters are to be calibrated to include transverse shrinkage and bending deformation modes. Instead, in order to simplify the model, the loads are superimposed on the same region. 3.2 Results We are interested in the out of plane deformation of the wings. Accordingly, the probes for the laser tracker have been positioned in the most external points of them as shown in Fig. 8.

Fig. 8. Markers positioned on the wings

From numerical results it was found that the particular welding sequence and parameters used do not require corrections because the overall maximum deformation was around 2 mm, as shown in Fig. 9. The base plate is not affected by welding deformation for its large thickness. From Fig. 9, it is observed that the results have not a symmetry plane, as expected, because welding was not carried out on both sides at the same time.