STUDY ON APPLICATION OF DEEP LEARNING TO FILLING EVALUATION OF PC GROUT

Abstract

The impact echo method using elastic waves is a method for evaluatingthe fillingstatus of PC grout.

Since the reflected elastic wave obtained by the impact echo method is complicated, it is difficult to determine the grout filling onlyfrom the frequencyspectrum. Therefore, the SIBIE method hasbeen proposed to visualize the results of frequency response analysis as a two-dimensional contour diagram. The SIBIE method makes it easier to determine the filling of grout, but it requires experience and skill. Therefore, in this study, we investigated amethod to supportthe filling judgment of engineersby applyingdeep learning to thecontour diagram created by the SIBIE method.