Visio-Acoustic Data Fusion for Structural Health Monitoring Applications

Structural health monitoring has been an expanding discipline due to its potential to decrease maintenance and downtime costs, detect failure early, extend life spans, and fulfill the increased need of safety and security. Acoustic source identification techniques’ have been used for remote structural health monitoring, but the applicability of each technique has been limited by factors ranging from achievable spatial resolution to hardware costs. This paper aims to mitigate current acoustic techniques’ limitations by exploring the possibility of fusing acoustic and video data. This paper focuses on combining microphone acoustic measurements with vibrational information recovered from video-based measurements. Among acoustic methods, acoustic arrays have been used for remotely detecting, localizing and characterizing acoustic sources. Acoustic-array based techniques are limited in their ability to discriminate multiple closely-spaced acoustic sources from far-field acoustic pressure signals. On the contrary, video-based techniques have shown the ability to recover full-field, high resolution mode shapes, and the associated frequencies and damping ratios with virtually no dependence with the distance from the target. The challenge with video methods, applied to acoustic source identification, is that acoustic sources may occur in the kilohertz range requiring a higher frame per second sampling rate than most low cost cameras. Acoustic measurements provide additional information content that is used to recover the correct frequency content of an acoustically radiating structure from temporally-aliased (sub-Nyquist) video measurements. Experiments are conducted to show how combining acoustic and video data relaxes the hardware requirements for acoustic source detection and localization applications.