Detection of damage in multiwire cables based on wavelet entropy evolution

Multiwire cables are widely used in important engineering structures. Since they are exposed to several dynamic and static loads as well as detrimental environmental conditions, their structural health can be compromised. Due to the critical role played by multiwire cables, it is necessary to develop a non-destructive health monitoring method to maintain their structure and proper performance. Ultrasonic inspection using guided waves is a promising non-destructive damage monitoring technique for rods, single and multiwire cables. However, the propagated guided waves are composed of an infinite number of dispersive vibrational modes making their analysis difficult. In this work, an entropy-based method to identify small changes in non-stationary signals is proposed. An experimental system to capture and post-process acoustic signals is implemented. The discrete wavelet transform is computed in order to obtain the reconstructed wavelet coefficients of the signals and to analyze the energy at different scales. The use of the concept of entropy evolution of non-stationary signals to detect damage in multiwire cables is evaluated. The results show that there is a correlation between the entropy value and level of damage of the cable including breaking of single wires and change in the mechanical contact conditions among the wires. It is found that the studied method has low sensitivity to signal noise and can reduce the computational complexity encountered in a typical time-frequency analysis.