4.3. Comparison between axisymmetric (2D-SIM) and 3D models
(3D-SIM)
Although 3D models can accurately capture the problem physics,
they are time consuming unlike the axisymmetric models. The
deviations in results are assessed here by comparing two models;
the first is a 3D complete model of the full welded joint (3D-SIM),
whereas the second is a 2D-axisymmetric model (2D-SIM).
Figs. 18 and 19 compare the predicted residual stress distributions
for both 3D model (3D-SIM) and the axisymmetric model
(2D-SIM). Both axial and hoop residual stresses are plotted along
longitudinal paths located at 180 from the weld start on inner and
outer surfaces. Fig. 18 shows relatively close agreement between
predictions of axial and hoop residual stresses along pipe inner
surface for the axisymmetric and 3D models. The predicted
maximum axial residual stress in 2D-SIM model is slightly greater
than predictions by the 3D-SIM model. However, along pipe outer
surface the predicted residual stresses by the axisymmetric model
contradict with those predicted using 3D model. Fig.19a shows that
axisymmetric model develops compressive axial stresses close
to 0.7 Sy near the weld toe, while the 3D model predicts a tensile
stress of approximately 0.3 Sy. Fig.19b shows that the axisymmetric
model develops compressive hoop stresses equal to 0.5 Sy near
the weld toe, while the 3D model predicts a tensile value of 0.4 Sy.
These differences are attributed to the inherent assumption of the
axisymmetric model where the welding speed is assumed to be
infinite leading to uniform heating of the pipe unlike the 3D model
where the weld deposition heat is advancing with the welding arc
and allowed to flow in all directions.
The above comparison shows that, axisymmetric modeling of