High-resolution mode shape identification using mobile sensors

Abstract

Structural health monitoring (SHM) procedures such as system identification (SID) or damage detection were originally developed to process data from fixed sensor networks. Interest in mobile sensors has recently grown with the popularity of mobile devices and smartphones. As documented in the literature, a mobile sensor network is capable of collecting spatially rich data using only a few sensors. Thus, mobile sensor networks can provide comparable or superior information at a lower monitoring cost when compared to a dense fixed sensor network. Until a recent breakthrough, a framework for SID using mobile sensor data sets was unavailable, since the data class is fundamentally different from fixed sensor data sets. When mobile sensor data is classified as dynamic sensor network (DSN) data, changes in sensors’ positions can be properly incorporated into a mathematical model, enabling SID via the structural identification using expectation maximization (STRIDE) algorithm. In this study, a wireless mobile sensor platform, comprised of a motor-drive pulley system, is implemented to record ambient vibrations of a simple beam structure. Using only two mobile sensors, mode shapes with twenty points are computed, further exhibiting the high spatial capacity of mobile senor data.