Hardware-accelerated Wireless Sensor Network for Distributed Structural Health Monitoring

Civil infrastructure objects are subject to safety-related issues such as increasing loads and extended service live. Costly manual inspections of these structures should therefore be supplemented by automated continuous monitoring. In this work, a hardware-accelerated wireless sensor network built entirely with energy-efficient embedded components is proposed as the basis for a distributed structural health monitoring (SHM) implementation. In addition to detection and localization of structural damage, the energy-efficiency of the wireless data acquisition system is the major topic of this work. By utilizing the Random-Decrement (RD) technique, the structure's modal parameters are acquired based on ambient excitation such as wind or traffic. The RD functions are calculated by a Field-Programmable Gate Arrays (FPGA) designed for mobile applications. To demonstrate the benefits of the proposed monitoring network, the model of a truss bridge is excited by a train set to simulate realistic operational excitations. Dominant mode shapes of the bridge model are extracted from the RD functions using frequency domain Operational Modal Analysis and compared to previously determined reference measurements. The loosening of a single bolted joint simulates damage and is found to be reflected in significant deviations of the first vertical bending mode, located at 68 Hz.