Task Outline
The Auto Non destructive Testing department of Nuclear electric is involved in
developing mathematical models to predict the ultrasonic responses from
defects in welds. The models for defects with smooth faces are well established
but further work is in hand to extend the models to cope with responses from
defects with rough faces. As part of this work it is necessary to obtain detailed
information of the surface profile of real rough crack faces.
Some years ago a roughness measuring machine was developed using a BBC
computer which drives stepper motors to scan the surface and compressed air to
force an LVDT detector onto the surface while the motors are stationary. This
method is used rather than dragging the detector along the surface because the
surfaces being tested are too jagged for the normal method. The system
requires updating, the most significant change being to convert it to use a PC
rather than a BBC computer.
With a budget of zero there was no alternative other than to update, keeping the
existing method. Therefore only a brief look at the alternatives will be given here.
The Talysurf [6] , [16] and Protrace [17] work by dragging a stylus along the
surface of the sample, but the dragging process would cause the stylus to snag
on the particularly rough surfaces with near perpendicular gradients which
needed to be tested. Optical fibres can be used [24] but the height range is only
about 0.5 mm whereas 5 mm is required, it is also likely that changes in surface
reflectivity could not be completely compensated for. Colour fringe projection
[23] requires a neutral coloured surface, and so is again unlikely to be able to
cope with the variations in surface texture reflectivity.
The LVDT (linear variable displacement transformer) [11] , [12] can be lowered and
raised by connecting an air supply, it has been around since 1946 but still gives
a good linear output over a range of 5 mm.