Application of compressed sensing in full-field structural health monitoring

Embedded sensors are used in layers of composite structures to provide local damage detection. The presence of these sensors causes material and geometric discontinuities which in turn causes unwanted peaks of stress and strain with consequences on stiffness reduction. Often several of these sensors are embedded in structures aggregating the adverse effects of discontinuities to degrade the structural integrity. Structural damage is a sparse phenomenon and the mechanical metrics are smooth functions with few spikes near the location of damage. This sparsity and spikiness can be exploited to reduce the number of embedded sensors in composite structures. The goal of this paper is to adapt the compressed sensing theory and detect damage using far fewer sensors than conventionally possible. To demonstrate the efficacy of our approach, we performed a numerical experiment on a rectangular plate with a center hole, and have shown that the 2D strain-field can be recovered from few samples of discrete strain measurements acquired by embedded sensors.