Ultrasonic inspection of multi-wire steel strands with the aid of the wavelet transform

The non-destructive detection of structural defects in multi-wire strands used as post-tensioned tendons and cable stays is a challenging, yet critical task. A promising method under investigation is based on the use of ultrasonic stress waves that propagate within the strand and interact with structural discontinuities. The waveguide-like geometry of the strands lends itself to the monitoring of long lengths at a time (long range). The topic of this paper is the enhancement of ultrasonic monitoring of strands by a joint time–frequency analysis based on the discrete wavelet transform (DWT). The test set-up uses magnetostrictive sensors for the excitation and the detection of ultrasonic guided waves in the strands. The main advantage of the DWT is an unmatched de-noising performance. Effective de-noising becomes necessary for the detection of small defects located far away from the inspection probes as desirable in the field. When compared to the traditional signal averaging, the DWT can be used in real time owing to its computational efficiency. The theory of the DWT filter bank decomposition is first revised. The effectiveness of the wavelet processing is then demonstrated for the detection of small notches of varying depths located in the free portion of the strands as well as in the critical anchored areas. The study also shows the importance of selecting the proper mother wavelet function for best performance. The DWT proves effective in eliminating the need for signal averaging and in reducing the power supply required by the monitoring system. Both outcomes make the guided wave inspection method for strands more suitable for field use.