3.4.3. Comparison of NDT with destructive peel test results
Following initial NDT analysis, the whole floor assembly was torn
down to reveal weld plugs, which were then measured using digital
Vernier. The differences between NDT and measured plug diameters
are shown as residuals in Fig. 20. For all 36 welds evaluated, the
majority (32 welds) had a residual less than 0.5 mm indicating a
good agreement between the two sets of independently made
measurements. These results are promising and demonstrated a
production viable method for surface preparation prior to NDT
measurements. It is therefore suggested that the technique offers a
robust NDT method for production components.
Out of the 52 welds in the floor assembly, 16 welds were not
ultrasonically tested; whilst 4 welds exhibited a residual greater
than 0.5 mm suggesting some inaccuracy. This was because some
difficulties were encountered during NDT of the floor assembly;
for example, non-flat contact surface due to forming; no space for
NDT access. Fig. 21 shows an example of C-scan image obtained
from a surface where the curvature was too large to obtain good
coupling with the scanner. No weld was revealed in the image
due to loss of signals. It is indicated that surface flatness is a
limitation for NDT of production components. However, in typical
assemblies there are usually many flange areas that are suffi-
ciently flat and could be effectively measured.