Nondestructive evaluation of elastic properties of concrete using simulation of surface waves

Abstract

Nondestructive evaluation using the propagation of an impact-induced surface wave can be effectively applied in estimating in situ material properties. In this study, to evaluate information of a surface waveform beyond the simple wave velocity, artificial intelligence engines are employed to estimate simulation parameters, that is, the properties of elastic materials. The developed artificial neural networks are trained with a numerical database having secured its stability. In the process, the appropriate shape of the force-time function for an impact load is assumed so as to avoid Gibbs phenomenon, and the proposed principal wavelet-component analysis accomplishes a feature extraction with a wavelet transformed signal. The results of estimation are validated with experiments focused on concrete materials.