Stress and damage estimation is essential to ensure the safety and performance of concrete structures. The capsule-like smart aggregate (CSA) technique has demonstrated its potential for detecting early-stage internal damage. In this study, a 2 dimensional convolutional neural network (2D CNN) model that learned the EMI responses of a CSA sensor to integrally estimate stress and damage in concrete structures is proposed. Firstly, the overall scheme of this study is described. The CSA-based EMI damage technique method is theoretically presented by describing the behaviors of a CSA sensor embedded in a concrete structure under compressive loadings. The 2D CNN model is designed to learn and extract damage-sensitive features from a CSA's EMI responses to estimate stress and identify damage levels in a concrete structure. Secondly, a compression experiment on a CSA-embedded concrete cylinder is carried out, and the stress-damage EMI responses of a cylinder are recorded under different applied stress levels. Finally, the feasibility of the developed model is further investigated under the effect of noises and untrained data cases. The obtained results indicate that the developed 2D CNN model can simultaneously estimate stress and damage status in the concrete structure.
Keywords: PZT sensor; concrete structure; electromechanical impedance method; smart aggregate; stress and damage monitoring; two-dimensional convolutional neural network.