2022 Volume 68A Pages 661-670
This paper deals with a corrosion crack judgement procedure based on neural network. The primary focus of this study is put in the judgement of cracks occurred by re-bar corrosion using analyses of hitting sound spectra. For this purpose, neural network which is one kind of machine learning method is employed to determine the presence of the corrosion cracks. The validity of this procedure is evaluated from an experiment with reinforce concrete specimens, and, effects of parameters of hitting conditions and process of sound spectra are discussed with various indices.