A probabilistic optimal sensor design approach for structural health monitoring using risk-weighted f-divergence

•The paper proposes a new form of Bayes risk functional using probability distribution distances.•This Bayes risk functional is used to design an optimal structural health monitoring sensor system.•Efficient Bayesian inference is achieved using surrogate modeling and Gauss-Hermite quadrature.•The framework proposed is demonstrated on a full-scale miter gate structure used in inland waterway navigational lock systems.•The application results show that the optimized SHM sensor placement increased the effectiveness of damage inference at minimal risk when compared to randomized designs.