기계공학부 구성원들의 많은 관심과 참여 부탁드립니다.

▣ Title: Recent efforts towards enhanced finite elementanalysis with machine learning

▣ Speaker: Gunwoo Noh(Professor)

▣ Affiliation: Dept. of Mechanical Engineering, Korea University

▣ Date: 2021. 6. 11.(Fri) 13:00

▣ Venue: Online(Zoom)

▣ Host: Prof. Jongsup Hong

▣ Abstract

Finite Element Method (FEM) is now an essential tool inalmost all fields of engineering, including the analysis of structure,fluid, heat transfer and electromagnetic fields. Recently, thechallenge of enhancing FEM is largely addressed with themachine learning techniques. In this talk, we will first discuss howa supervised-learning enhances the numerical integrations whencalculating the matrices in modern finite elements, such asMethod of Finite Spheres, Enriched FE, and Enriched BeamElement. Then, the deep neural networks as surrogate models ofFEM will be discussed using the cases of ultrasound propagationsin the human and of the designing patterned surgical tubes

https://yonsei.zoom.us/j/81398414478?pwd=OGExenRZRUI1elJmZDJ4dHhUZW16UT09