A feasibility study has been conducted in order to evaluate the potential of pattern recognition techniques for discriminating between geometrical and crack reflector signals obtained during ultrasonic inspection of the weld zone in 304 austenitic stainless-steel pipes. The analysis did not consider any arrival time, amplitude information, or in fact, any other time-domain features, but was based on various Fourier Transform features. A 100% reliability level was obtained for discriminating between a crack indication and a crown indication using automated pattern recognition algorithms. The overwhelming success of the pattern recognition algorithm employed in this study demonstrates the applicability of this technique for solving such important problems as discrimination between intergranular stress corrosion cracking and geometric reflectors in 304 stainless-steel pipe-welds. 8 refs.
|Original language||English (US)|
|Number of pages||4|
|Journal||British Journal of Non-Destructive Testing|
|State||Published - Nov 1 1979|
All Science Journal Classification (ASJC) codes