Detecting damage events in concrete using diffuse ultrasound structural health monitoring during strong environmental variations

Abstract

Diffuse ultrasonic wave measurements used in structural health monitoring applications can detect damage in concrete. However, the accuracy is very susceptible to environmental variations. In this study, a large concrete floor slab was monitored using diffuse wave fields that were generated by continuous-wave transmissions between ultrasonic transducers. The slab was monitored for several weeks while being subjected to changes in environmental conditions. Subsequently, it was damaged using impact hits, resulting in centimeter-scale cracking. The variations caused by the environment masked the effects of the damage in the measurements. To address this issue, the Mahalanobis distance was used to distinguish between the influence of the damage and the influence of the environmental variations. The Mahalanobis model uses amplitude and phase measurements of continuous waves at a set of different frequencies as inputs. A moving window approach was applied to the baseline data set to account for slow trends. This study shows that this technique greatly suppresses most of the variations caused by environmental conditions. All damage events in our data set have been detected.