Unmanned Aerial System–Based Portable Sensing for Blast-Loaded Cables.

Measuring the dynamics of vibrating cables in situ can be challenging using traditional contact-based sensors (e.g., accelerometer). This work proposes a novel computer vision–based technique using a portable unmanned aerial system (UAS) (also referred to as drones) sensing platform to enable the measurement of the dynamic displacements of a cable in a unique blast-loading experiment where other types of sensing may be impractical. The proposed technique utilizes one optical camera equipped on a UAS to record the dynamic displacement of the blast-loaded cable and measure the drift of the UAS as it hovers. Artificial targets attached to the cable are not required because the technique proves effective when tracking natural features inherent in the cable. This makes the proposed technique versatile and convenient for field implementation. This technique also allows simultaneous displacement measurement at multiple points along the cable across the entire field of view of the camera. The natural frequency, damping, and mode shape are successfully identified from the measured cable displacement time histories. Measuring the vibration of a cable is challenging with traditional techniques. A novel technique is proposed to use only one camera attached to a drone to measure the vibration of a cable directly. However, because the drone hovers in the air, there is inherent drift of the drone. The drone's drift is measured with respect to a stationary target using the camera and compensated for to measure the true motion of the cable. To test the proposed technique, a blast was detonated close to the cable to produce vibration in the cable, and the drone hovered above the cable to record its movements. Because the actual movement of the cable is unknown, finite-element simulations were performed to provide a basis for validating the proposed technique. The natural frequency of the finite-element model is compared with that obtained from the measured vibration of the cable. A good agreement was achieved with a 2.1% difference between the simulated and measured natural frequencies. [ABSTRACT FROM AUTHOR]