Even though two of the welding procedures used had welding parameters significantly outside those recommended in any standards, of the short-term
tests used, neither the guided side bend test (ASTM F3183) nor the tensile impact test (ASTM F2634) generated failures at the weld. The waisted tensile test (EN 12814-7), however, could discriminate between the six different welding procedures investigated.
FEA modelling using macrographs of sections through the weld beads suggested that the maximum stress concentration was always at the notch between the
outer weld bead and the outer pipe surface. This was confirmed by the WPTCR tests, which showed that the failures always initiated from the notch between the weld bead and the pipe surface. These tests also suggested that welds made using a low welding pressure (0.15MPa) had longer times to failure than those made using higher pressures (>0.5MPa).
The results of the NI measurements revealed not only the boundary of the melt zone, which correlated well with the melt zone thickness measurements from transmission light microscopy (TLM) and DSC, but also the heat affected zone (HAZ) boundary, which was not visible using TLM (Figure 1).
The amount of melt generated during the heating stage of the welding cycle for each welding procedure, as predicted using an FEA thermal model, correlated reasonably well with the size of the weld beads measured from macrographs of the weld, added to the size of the melt zone measured from TLM.