Acoustic emission analysis of crack type identification of corroded concrete columns under eccentric loading: A comparative analysis of RA-AF method and Gaussian mixture model

Abstract

Crack location and type identification of reinforced concrete columns in service is of great interest to researchers, since it influences the decision-making of maintenance and retrofit. Due to the existing of load and corrosion, the complex stress state leads to complicated crack development which has been a challenge for researchers to determine the crack type. Acoustic emission (AE) method gives a promising solution to non-destructively locate and identify the crack position and type. In this work, corroded reinforced concrete columns were subjected to compression test with different eccentric degree. AE signals during the loading process were recorded. The traditional RA-AF method and a newly proposed Gaussian mixture model (GMM) method were applied to process the data for crack type identification. It was found that due to the large data set, RA-AF cannot identify the crack type in between the shear crack and the tensile crack types. The probability density distribution of GMM result gives an attractive offer to determine the crack type. A simple 4-step method based on GMM plot is proposed to identify the crack type which can calculate the ratio of shear crack and tensile crack in collected signals.