BK21 Special Seminar Series  - Spring 2023중 네번째 세미나가 아래와 같이 개최됩니다.

BK21 사업의 참여교수님, 신진연구인력, 참여대학원생, 학사과정 학생들의 많은 참여 부탁드립니다.

각 연구실의 BK과제 담당자께서는 참여학생들이 참석할 수 있도록 안내 꼭 부탁드립니다.

세미나는 영어로 진행됩니다.

- 일시 : 2023년 6월 8일 (목) 오후 4시 

- 연사 : 김태용 교수님(아주대학교) 

- 장소 : 오프라인: 110동 1007호, 온라인(ZOOM) ID:  436 397 4287 

- 주제 : Deep learning-based prediction of nonlinear hysteretic responses: A new paradigm for seismic assessment of structures

​ 

- 초록  

 : This seminar presents new deep learning-based methods for response prediction of nonlinear hysteretic systems, aiming to improve computational efficiency while maintaining high accuracy. Ideas are presented in the context of earthquake engineering, specifically in predicting the peak seismic responses of structural systems, which are crucial in seismic design and assessment. Unlike conventional methods that rely on an idealized monotonic pushover curve, the proposed method utilizes the hysteresis loops of structures as input to deep neural network (DNN) models. By integrating this input with a Convolutional Neural Network, the DNN models can capture the effects of complex hysteretic behaviors, such as stiffness or strength degradation, pinching effects, and smooth transition from elastic to inelastic range, on the responses. Three different DNN models are presented, including details on the constructed datasets, loss functions, and architecture. The excellent performance and advantages of the proposed framework are demonstrated through various structural engineering problems and seismic loss assessments of a community.

​

​