A grid-based stress-field imaging method for concrete structures using directional acoustoelastic measurements

This study presents a stress-field imaging method for concrete structures. A two-stage inversion framework is developed by integrating acoustoelastic theory with a grid-based time-variation decomposition strategy to recover local stress information from long-path measurements. In the first stage, the principal stress direction field is determined using directional acoustoelastic measurements collected along multiple non-collinear sensing paths, and radial basis function (RBF) interpolation is adopted to reconstruct a continuous first principal stress direction vector field. In the second stage, the measured ultrasonic time variations along each transmitter-receiver path are decomposed into cell-wise contributions, and the in-plane stresses are reconstructed using the previously determined principal stress direction field and the angle-dependent acoustoelastic formulation. Experimental validation on a reinforced concrete corbel column subjected to eccentric compression confirms the feasibility of the proposed method and demonstrates its ability to capture both the spatial distribution and directional evolution of internal stresses. The reconstructed principal stress direction field shows good agreement with finite element results, with a curl RMSE of 0.01337. The reconstructed y-direction stress also agrees well with the FE results, with an absolute error of 0.05 MPa at 40 kN. The inversion accuracy improves as stress-induced time variations become dominant over concrete heterogeneity. The proposed framework provides a reliable basis for qualitative and quantitative characterization of internal stress fields in concrete structures.