Smart structural health monitoring system for damage identification in bridges using relative wavelet entropy

A smart structural health monitoring system for damage identification in bridges is proposed. The system is based on the fact that structural damage induces abrupt changes and singularities in measured acceleration signals of bridges and that discrete wavelet transform is a powerful tool to decompose the measured signals in order to detect these singularities. In addition, relative wavelet entropy is used to quantify the degree of damage-induced disorder in the signals. The main advantages of the system are that (1) there is no need to obtain the dynamic data of the undamaged state of bridges (reference-free); (2) there is no need to control or even measure the input excitations (response-only); (3) it is capable of evaluating both the global dynamic properties of bridges and the local structural condition of their elements; and (4) it can be utilised for identification of different types of damage in different types of structure. To illustrate robustness of the system, experiments were conducted on two structural systems: a reinforced concrete beam strengthened with steel-reinforced polymer (SRP) sheets, and a precast hybrid truss girder made of glass fibre-reinforced polymer (GFRP) tubes filled with concrete reinforced and connected to pretensioned top and bottom concrete chords by double-headed GFRP bars. Different damage scenarios including concrete cracking and debonding of the SRP sheets in the strengthened beam, and rupture of the GFRP tubes in the precast truss elements were investigated. The results demonstrate ability of the system to detect, localise, and estimate severity of the damage in the elements tested.