Detection of Internal Defects in Concrete Structures Based on Machine Learning Using Hammering Response Data Containing Artificial Noise (Evaluation of Machine Learning Models with High Generalization Performance)

Abstract

Aging of concrete structures has become a serious problem in Japan, and periodic maintenance is essential for preventing accidents caused by structural aging. In this study, a method for estimating defects in concrete from hammering test data on a concrete plate using machine learning was developed. A neural network based on Convolutional Neural Network, Residual Network and Self-Attention Network to estimate the three-dimensional position and size of the defects was constructed. Moreover, a dataset was created from the topology of the internal defects and the acceleration response waveform when a concrete plate was struck. The entire plate was represented as a nondimensional density matrix. The scalograms generated from the acceleration response waveform was used as the input. Furthermore, estimation was performed using acceleration response waveforms containing artificial noise. In this study, as a preliminary step to detecting internal defects in concrete structures based on machine learning using hammering response data, we conduct a study based on machine learning using hammering response data containing artificial noise, and evaluate machine learning models with high generalization performance.