Research on three-dimensional cavity shape identification in concrete structures based on machine learning using hammering response data

Abstract

The aging of concrete structures such as infrastructures is becoming a problem in Japan. To automate the maintenance of concrete structures, there is a demand for a system to estimate the location of defects inside concrete structures. In this paper, we propose a method for identifying the location and shape of 3D cavities in concrete structures using machine learning. Acceleration time history data obtained from hammering tests are used as training data, and a convolutional neural network, which is generally utilized for image recognition, is applied to estimate the cavity position and shape. Since the convolutional neural network can retain relative positional relationships within the input information, it can input acceleration response data while maintaining information on the positional relationships of multiple sensors and is considered to be suitable for estimating the location of cavities in concrete structures. For the hammering test data, preprocessing by dividing the acceleration response waveform by the maximum impact force was performed to eliminate the effect of impact force, and it was confirmed that the preprocessing significantly improved the accuracy of cavity location estimation. In this paper, we challenged to estimate the 3D position of internal cavities in concrete structures and presented numerical results and discussion.