DEVELOPMENT AND VERIFICATION OF REAL-TIME DAMAGE ASSESSMENT BASED ON STATISTICAL PATTERN RECOGNITION TECHNOLOGY FOR STRUCTURE MAINTENANCE GwangHee Heo 1 , ChungGil Kim 1 , ChinOk Lee 2 , ByeongChan Ko 3 , ChaeRin Park 3 1 Civil Engineering Dept. Professor, Konyang University e-mail: {heo, cg-kim}@kyu.ac.kr 2 Civil Engineering Dept. Professor, Chungnam University e-mail: colee@cnu.ac.kr 3 Department of Disaster Safety Engineering Dept. Master’s Candidate, Konyang University e-mail: {yunil-0, cherish_7169}@naver.com Abstract This study focuses on the development of a real-time damage detection technology based on statistical pattern recognition methods to detect damages in a complex structure, such as a cable-stayed bridge. Further, the performance of the proposed method is evaluated experi- mentally. The real-time damage detection method is devised on a statistical pattern recogni- tion technology from a damage data assessment section and a simulation based on a non- damage data generation section. The damage data assessment section applies the improved Mahalanobis distance theory from among the available statistical pattern recognition tech- nologies. Further, the non-damage data generation section comprises logic based on the state-space equation. To verify the performance of the technology developed, a damage detec- tion test was conducted on the model structure of the cable-stayed bridge. The experiment confirms that the real-time damage detection technology determines the location of damages in the cable-stayed bridge in real time. Keywords: Statistical pattern recognition, Real-time damage assessment, Mahalanobis dis- tance, State-space equation, Maintenance, Cable-stayed bridge 3861 COMPDYN 2019 7 th ECCOMAS Thematic Conference on Computational Methods in Structural Dynamics and Earthquake Engineering M. Papadrakakis, M. Fragiadakis (eds.) Crete, Greece, 24–26 June 2019 Available online at www.eccomasproceedia.org Eccomas Proceedia COMPDYN (2019) 3861-3869 ISSN:2623-3347 © 2019 The Authors. Published by Eccomas Proceedia. Peer-review under responsibility of the organizing committee of COMPDYN 2019. doi: 10.7712/120119.7191.18957