Early detection of wire fracture in 7-wire strands through multiband wavelet analysis of acoustic emission signals

In this study, acoustic emission (AE) features to predict and detect wire fracture in seven-wire strands were characterized with multiband wavelet analysis. Two steel strands were tested up to 89 kN with each instrumented with a pair of AE sensors at two ends. The cross section of one wire was locally reduced up to 90% in 10% increment at center and support of the two strands, respectively. For both strands, the AE parameters (hits, energy, and counts) changed little up to 80% reduction in cross section of the partially cut wire, and suddenly jumped at the fracture (under 73 kN) of the notched wire with 90% reduction in cross section. The acoustic signals of inter-wire slippage and fracture initiation are significantly shorter in time duration than the signal of fracture. Their dominant frequencies and frequency bandwidths are increasingly higher and wider. The frequency band of the fracture signal is significantly broader than that of either the fracture-induced echo or artificial tapping noises. The time duration of artificial tapping noises is substantially longer than that of either fracture or fracture-induced echo. These distinct time-frequency characteristics allow an early detection and localization of wire fracture following the proposed procedure.