Embedding damage detection algorithms in a wireless sensing unit for operational powerefficiency

Abstract

A low-cost wireless sensing unit is designed and fabricated for deployment as the buildingblock of wireless structural health monitoring systems. Finite operational lives ofportable power supplies, such as batteries, necessitate optimization of the wirelesssensing unit design to attain overall energy efficiency. This is in conflict with theneed for wireless radios that have far-reaching communication ranges that requiresignificant amounts of power. As a result, a penalty is incurred by transmitting rawtime-history records using scarce system resources such as battery power and bandwidth.Alternatively, a computational core that can accommodate local processing ofdata is designed and implemented in the wireless sensing unit. The role of thecomputational core is to perform interrogation tasks of collected raw time-historydata and to transmit via the wireless channel the analysis results rather thantime-history records. To illustrate the ability of the computational core to execute suchembedded engineering analyses, a two-tiered time-series damage detection algorithm isimplemented as an example. Using a lumped-mass laboratory structure, localexecution of the embedded damage detection method is shown to save energy byavoiding utilization of the wireless channel to transmit raw time-history data.