PSI - Issue 58

Mirjana Opačić et al. / Procedia Structural Integrity 58 (2024) 87 – 94 M. Opa č i ć et al. / Structural Integrity Procedia 00 (2019) 000–000

93 7

Fig. 9. Stress distribution results for the new and improved model.

5. Discussion and conclusions Results presented in the previous chapter clearly indicate that the second approach to modelling of pressure vessel 970 was the more relevant one. In this case, the crack was not constrained by the lower surface of the vessel, since the fixed boundary condition was removed, and was allowed to freely deform, although only in the transversal direction. From our point of view, this was favourable, since crack growth along its length would suggest that this defect is actually compromising the integrity of the vessel. It is interesting to observe that the maximum stresses in the second model were almost equal to the yield stress of the weld metal (560 MPa) - thus causing minimum plastic strain in the model, which was around 0.8%. This strain however, was not sufficient enough to cause the crack to propagate. As for the strain in the initial model, there was no plasticity (since maximum stresses were considerably below the yield stress of either materials). Furthermore, elastic strain in this case was much lower than the already small plastic strain in the improved model, which also confirmed that the initial selection of boundary conditions was not sufficiently realistic. It should also be noted that this improved model is also quite conservative, as it considers a scenario where the crack is still free on one end, despite being located within the plate and welded joint - since the lower segment of the pressure vessel mantle was removed. This implies that a model with both segments would have more constraints on the crack, which would be “surrounded” by solid material from all sides. As a direct consequence, its deformation would be limited compared to the model presented here and the stresses would be even lower. For this reason, it can be concluded that the real pressure vessel is still safe for operation, as the stresses within the defect are not high enough to cause crack propagation in any case. One should also be aware that the defect 5.6 which was found in the welded joint is not necessarily a crack - this assumption was made following common engineering logic to always assume a worst case scenario. Of course, what was said above raises the question of why a more complex model with both segments of the mantle was not made. The answer however is simple - there was no time, or available resources to complete the calculation at this point. Attempts were made, but failed due to lack of memory on the computers which were used, and this approach was abandoned after realising that the improved model which was made provided sufficient compliance with the real conditions of the pressure vessel 970. It can be concluded that the goal of this research - to confirm the integrity of a pressure vessel by combining NDT methods, namely ultrasonic testing, with numerical simulations. The presence, location and dimensions of the defect was determined using PAUT and ToFD technique, and was used as the basis for numerical simulations using finite element method. Eventually, numerical models confirmed that the integrity of the vessel is not compromised by the defect in question, even in the most extreme cases considered. Since the pressure vessel 970 itself has been operating