Information entropy-based algorithm of sensor placement optimization for structural damage detection

The structural health monitoring (SHM) benchmark study on optimal sensor placement for the Canton Tower, sponsored by the Asia-Pacific Network of Centers for Research in Smart Structures Technology (ANCRiSST), is first detailed in this paper. Then a two-step information entropy-based algorithm for sensor placement optimization is performed on the benchmark model. In the first step, mode shape change due to damage is extracted through modal analysis on the benchmark structure. In the second step, information entropy index is introduced to measure the uncertainties over the damage detection for each mode shape change, and then a multi-objective minimization problem in terms of different sensor placements is constructed. The optimal solution is determined as the one that provides almost equally informative data for all objectives. To validate the effectiveness of the determined optimal sensor placement, damage detection is performed on damage scenarios of the benchmark model using pre- and post-damage mode shapes obtained from the selected sensor placement. The results show that the obtained sensor configuration is able to provide information sensitive to damage.