Department News

Professor Byung-Dong Yoon’s research team, Jong-Min Park, Hyun-Hee Choi, Joo-Hyun Lim doctoral students won the Best Paper Awar

Author
jinjookim01
Date
2023-06-27
Views
302


 

On May 18th, Ph.D. students Jong-min Park and Hyun-hee Choi from the System Integrity and Risk Management Lab (Advisor Byung-dong Yoon) of the Department of Mechanical Engineering at Seoul National University won the Best Paper Award at the 2022 Spring Conference of the Korean Society of Mechanical Engineers in CAE and Applied Mechanics. The thesis was honored with the award as it was considered to contribute greatly to the development of science and technology with original research contents.

The paper title by Park Jong-min, Ph.D. student, is "Development of speed-independent bearing failure feature extraction method using autoencoder under various speed conditions", and introduces a method for robust bearing failure diagnosis under various speed conditions. In this study, failure diagnosis was attempted usingly failure-related information from which speed-related information was separated. In this process, multi-scale and multi-task methods were used. Therefore, the proposed method drew attention as it showed high accuracy under various noise and speed conditions.

The title of the dissertation of Hyun-hee Choi, Ph.D. student, is “Development of an equivalent dynamics model for a pouch-type battery considering swelling”, which is equivalent dynamics of a battery cell and its internal structure to predict surface pressure caused by swelling inside a battery pack used inside an electric vehicle. The paper was recognized for its originality and excellence in that it made it possible to predict the surface pressure of battery modules under various conditions in design situations by conducting research to develop a “model”.

Ph.D. student Ju-Hyun Lim's paper titled “Deep learning-based physical property estimation for FDM-type 3D printers” selected channels and features that have a high influence manufacturing quality through correlation analysis, and estimated material properties through convolutional neural networks. As a result, it drew attention in that it showed that estimation performance improved whenly highly correlated channels were used.