In this study, to develop the basis of damage prediction system for abutment type rigid-frame bridge, measurement data is generated by artificially expressing damage by Abaqus, a commercial structural analysis program, and applied to machine-learning. The rigid-rame bridge structural analysis model is expressed as closely as possible to the actual bridge condition considering the specification, damage expression, analysis method, boundary condition, and load. CNN(Convolutional Neural Network), one of the neural network algorithm, is used for machine-learning and accuracy is confirmed when there was no measurement error as a result of machine learning.

ABSTRACT
 1. 서 론
 2. 유한요소해석
  2.1 해석 개요
  2.2 손상 표현
  2.3 하중 및 경계조건
 3. 머신러닝
 4. 소결
 References

• 변남주(고려대학교 건축사회환경시스템공학과) | Byun Nam Joo
• 한휘석(고려대학교 건축사회환경시스템공학과) | Han Whi Seok
• 권영웅(고려대학교 건축사회환경시스템공학과) | Kyon Young Woong
• 김평화(고려대학교 건축사회환경시스템공학과) | Kim Pyoung Wha
• 한상윤(고려대학교 공과대학 초대형건설기술연구소) | Han Sang Yun 교신저자