TG2: Letter Box Repair Weld (Ferritic Steel)
Task Group 2 originally aimed to assess the impact of stress relieving heat treatments at different temperatures on the residual stress field around an 18-bead letterbox type repair weld in a 2.25CrMo ferritic steel plate of size 200 x 100 x 20 mm3.
Residual stress measurements on these specimens did indeed show the effectiveness of even the lowest heat treatment regime at about 620 oC. However, it soon became apparent that accurate 3-D modelling of such specimens demanded huge computational resources not essential to the achievement of the fundamental scientific and engineering objectives. A number of studies was undertaken nevertheless in order to compare, for example, 2-D and 3-D modelling or to assess the impact of different weld deposition strategies (bead-by-bead, layer-by-layer, or lumping of beads) used for modelling.
TG2: Auxiliary Specimen (Ferritic Steel)
A more simple 'auxiliary specimen' was then developed with just three beads deposited in a slot within a 2.25CrMo ferritic steel plate sized 400 x 200 x 20 mm3. Residual stress measurements were obtained using neutron diffraction, surface X-ray diffraction, the contour method and deep hole drilling, together with measured micro-hardness profiles.
The auxiliary specimen embodied most of the essential features of a real engineering weld in a form that was manageable both experimentally and computationally. The effect of subsequent passes on beads already deposited and the behaviour of the final pass were both represented, although machining of the overfill and post weld heat treatment, that had been included in the 18-bead version, were omitted. Similar three bead slot-weld specimens went on to be adopted a few years later in Task Groups 4 and 6.
A significant amount of modelling work was completed on the auxiliary specimen design. Much of this was still of a preliminary nature involving sensitivity and convergence studies, but a number of later simulations included phase transformation effects, implemented using either proprietary software designed for the purpose or material routines written by the participants. It is therefore probably correct to say that the measurement work was more comprehensive than the numerical efforts, although results from different facilities showed some variation, giving rise to long discussions about lattice planes and d-zero measurements. With hindsight, however, it was realised that, even with just three beads, the use of a transformable steel introduced too many variables into the study. It was difficult, for example, to separate the effects of phase transformation from the effects of cyclic hardening and annealing. As the accurate representation of all relevant physical processes was necessary to provide agreement with experiment, it was difficult to identify the reason when agreement was not achieved. This problem was compounded by the fact that much of the physical data used for modelling was generic rather than measured.
Fig. 3. TG2 auxiliary specimen – 3-dim plot of simulated longitudinal stresses
It is probably true to say that the real contribution of Task Group 2 was it being a learning exercise that lead to the successful implementation of Task Group 4, a three bead slot weld in a non-transformable steel, and Task Group 5, a ferritic autogenous weld. Now that the separate behaviour of both is better understood, there may be some merit in revisiting Task Group 2 in the future in the light of the new knowledge obtained.