A virtual lab for damage identification in concrete using coda wave interferometry

Early identification and prevention of damage in concrete structures can significantly reduce maintenance and repair costs. Weak material degradation, such as load-induced microcracking, generally is a precursor of localized damage in concrete structures can be detected by means of ultrasonic signals. To reliably identify and quantify damage, a systematic method that translates ultrasonic coda signals into the damage state is required. To this end, the effect of material degradation on the coda variations at the specimen level is systematically investigated using a combination of multiscale computational modeling, wave propagation simulations, and Coda Wave Interferometry. The study reveals a strong correlation between relative velocity variation and stiffness variations under stress, confirming the method’s sensitivity to microstructural changes. Simulations of mesoscale concrete models in a virtual lab reveal that relative velocity change increases linearly with stress during initial deformation (up to 1.23%) and decreases significantly during the microcracking stage (−3%), correlating reasonably with experimental data. Additionally, the computational framework enables testing across a robust sample set to estimate the probability of failure, supporting more informed decision-making in structural health monitoring. Finally, a strategy for using specimen scale information to predict the state of damage at the structural scale is presented.