A new approach for detecting the damage of underground bridge foundations using transfer resistance

Abstract

Natural disasters such as earthquakes and tsunamis have shown that bridge structures are an important part of the arterial roadways used for emergency transportation and evacuation. Accordingly, continuous research and retrofitting have been conducted on vulnerable bridge structures to mitigate potential damage. A common characteristic of these efforts is the lack of strengthening of the foundations, perhaps because their modelling assumptions led to a conservative structural design. Nevertheless, when the retrofitting specifies a large column capacity over the foundation, earthquake damage is likely to occur. To improve the decision-making, a methodology was proposed to identify bending damage based on the transfer resistance that reveals the error structure due to voltage sensing. Unstructured mesh discretisation modelling, non-deterministic external effects, and statistical error analysis were combined to minimise the error due to mesh gridding, consider temporal effects that increase the dispersion of the dataset, and quantify the correction parameters. Then, the residual power spectrum of the normalised error was used to observe the changes in the spatial component within the foundation and associate them with cracking patterns. By combining the residual power spectrum and reconstructed image, the earthquake bending damage could be identified as physical pathways controlled by reinforcement responses.