The Classification for Corroded Reinforced Concrete Beams using Linear Discriminant Analysis
Ahmad Zaki (a,*), Zainah Ibrahim, Yessi Jusman

(a) Department of Civil Engineering, Faculty of Engineering, Universitas Muhammadiyah Yogyakarta
*ahmad.zaki[at]umy.ac.id
(b) Department of Civil Engineering, Faculty of Engineering, University of Malaya
(c) Department of Electrical Engineering, Faculty of Engineering, Universitas Muhammadiyah Yogyakarta


Abstract

Corrosion of the steel reinforcements in RC structures is a worldwide problem. The corrosion has been recognized as the major deterioration mechanism which affects RC degradation due to the environmental actions. The costs of repair and maintenance of corroded structures worldwide exceed billions of dollars per year. It becomes necessary that the effects of steel reinforcement corrosion to the functionality of RC structures be detected at an early stage and studied in detail using acoustic emission (AE) technique in order to provide effective means of remedial. The purpose of the study is damage classification method for corroded reinforced concrete (RC) beams using AE data. This study proposes a damage classification method for corroded reinforced concrete (RC) beams subjected to flexural loading by linear discriminant analysis (LDA) of acoustic emission (AE) data. In this study, damage classification of the corroded beam specimens using AE parameters and LDA has been successfully conducted. RA value of AE parameter experienced a significant drop at Stage 2. Furthermore, the drop of RA value is used for the classification using LDA. The classification data give information in terms of statistical features based on the correlation of the distribution data. The effectiveness of LDA has been demonstrated empirically to classify the corroded beam to the classification with the high accuracy. The promising results obtained in the analysis are proposed to classify the fracture type of the corroded specimens.

Keywords: Corrosion, Concrete, Acoustic Emission, AE parameters, LDA

Topic: Engineering

BISMAS 2020 Conference | Conference Management System