Structural damage detection based on features insensitive to ambient factors

Vibration based damage detection usually faces two fundamental problems, namely the low sensitivity to the relatively small but potentially dangerous damage, and the dependence of vibration characteristics of structures on ambient factors (primarily temperature). Indeed, the effect of ambient factors on the vibration characteristics may be more significant than the influence of damage. In such situations, the reliable detection of damage is practically impossible. To solve this problem, the ratios of the natural frequencies for various mode shapes have been used. As the presented investigation reveals, these ratios are independent of the ambient factors and hence can improve substantially the sensitivity and reliability of damage detection. To illustrate the higher efficiency of the proposed ratios applied to damage detection, the analytical model of the reinforced Z24 concrete bridge with the damage of crack type was developed. The natural frequencies of bending vibrations in vertical and horizontal planes at different crack depths and locations were calculated and compared with those measured on the Z24 bridge, demonstrating that the sensitivity of the natural frequencies change and ratios of frequencies for various mode shapes with respect to the size of damage is of the same order. At the same time, as opposed to the change of natural frequencies, the frequency ratios are dependent only on the parameters of damage, which makes them a much more efficient tool for health monitoring of structures operating in variable ambient conditions.