Supersonic inspection of concrete using radial basis function neural network

A neural network approach to model the damage of concrete subjected to triaxial compressive loading history is presented in this paper. The damage trial of concrete cube was experimentally studied. The damage of concrete due to loading history is defined as the reduction of compressive strength and tensile strength. A radial basis function neural network (RBFNN) is used for training and testing the experimental data in order to acquire the relationship between the damage and the descent of ultrasonic velocity. A good agreement between the measured data and predicted results demonstrates that the model is able to capture significant variability inherent in the concrete samples. It is concluded that the application of RBFNN in supersonic inspection is a new method to evaluate damage of concrete, and has promising applications in structural engineering problems.