Development of a Damage Detecting Method for RC Slabs by Means of Machine Learning

Abstract

It is beneficial to understand damage conditions of RC structural members by non-destructive methods. In this study, K-means clustering method was applied to the AE wave data in order to develop a damage detection method. From the result of this study, the damaged area had the relatively large number of AE hits and AE energy. In addition, from the result of the K-means clustering of the power spectral density of AE waves, the data could be separated into the obviously damaged area and other areas.

Full document