Reliability of wavelet analysis of mode shapes for the early detection of damage in beams

Abstract

Numerous procedures have been developed for the detection and the localization of damage in structures based on changes in the dynamic or static response of structures. Among these, procedures based on wavelet analysis of mode shapes appear to offer a superior performance especially for low levels of damage. In order to evaluate the relative merit of these approaches, criteria based on statistical and probabilistic performance are evaluated as a function of damage level for an experimental beam. These measures include the probability of detection, the probability of false alarms, and the safety index. The safety index used in this application is for the beam under pure bending, which provides a uniform criteria for damage at any location along the beam. The experimental data is obtained on a steel beam where the level of damage is controlled at two locations along its length. A total of 16 equally spaced accelerometers are deployed along the length of the beam. The dynamic properties used to illustrate the proposed procedure are changes in the frequency of the first mode of vibration of the beam and changes in the wavelet coefficients for the first mode of vibration. The data obtained for 5 increasing level of damage at two locations is used to derive prediction equations for the dynamic properties and to estimate the probability of detection and of false alarms as a function of damage level. The results indicate that the procedure based on wavelets is more efficient than the one based on natural frequencies in detecting and localizing low levels of damage. The results also indicate that a monitoring strategy based on wavelets can detect damage before structural safety is significantly compromised while maintaining low probabilities of false alarms.