Benchmarking of optimal sensor placement criteria through exhaustive combinatorial search

The present study investigates optimal sensor placement (OSP) strategies for plate-type structures through a comprehensive comparison of three criteria-based methods against a global benchmark. All methods are implemented through exhaustive combinatorial search. The benchmark determines globally optimal sensor locations by minimizing the difference between the true mode shapes and the radial basis function (RBF) regression surface constructed from sensor data. In contrast to the majority of studies that use the Modal Assurance Criterion (MAC), this study includes two alternative criteria: Pearson correlation (Corr) and Guyan-based Orthogonality (GbO) for criteria-based optimization. Each criterion is formulated as a similarity matrix and optimized under various (Formula presented.) norms. The performance of the criteria-based methods is evaluated against the benchmark using the effective mass weighted mode shape interpolation error across three boundary condition cases. Our results show that the criteria-based methods demonstrate comparable accuracy to the global optimum while significantly reducing computational cost. Furthermore, both Corr and GbO criteria exhibit competitive performance compared to MAC. The proposed framework provides a physically interpretable performance measure and offers practical guidance for selecting the optimal criterion-norm combinations.