DSpace Community: KAIST Dept. of Mechanical EngineeringKAIST Dept. of Mechanical Engineeringhttp://hdl.handle.net/10203/132024-03-05T19:43:58Z2024-03-05T19:43:58ZSingle domain generalizable and physically interpretable bearing fault diagnosis for unseen working conditionsKim, IljeokKim, Sung WookKim, JeongsanHuh, HyunsukJeong, IljooChoi, TaegyuKim, JeongchanLee, Seungchulhttp://hdl.handle.net/10203/3171742024-01-02T06:00:12Z2024-05-01T00:00:00ZTitle: Single domain generalizable and physically interpretable bearing fault diagnosis for unseen working conditions
Authors: Kim, Iljeok; Kim, Sung Wook; Kim, Jeongsan; Huh, Hyunsuk; Jeong, Iljoo; Choi, Taegyu; Kim, Jeongchan; Lee, Seungchul
Abstract: State-of-the-art deep learning methods have demonstrated impressive performance in the intelligent fault diagnosis of rolling element bearings. However, in previous studies, critical issues such as domain discrepancy and the inability to interpret a classification decision made it difficult to apply deep learning in real industrial scenarios. Although domain adaptation and domain generalization-based methods have been investigated to solve domain discrepancy, collecting labeled data for various domains (especially continuous and non-stationary working conditions) is extremely difficult in an engineering application. Furthermore, since the classification decision cannot be physically explained, serious reliability problems arise for unseen working conditions (i.e., target domain with domain discrepancy). This study proposes the single domain generalizable and physically interpretable (SDGPI) framework. The proposed model embeds prior knowledge into the neural network combined with signal-preprocessing, which simultaneously enables single source domain generalization and domain interpretation with physical guarantees. Comprehensive case studies demonstrate that domain generalizable representation leads to 1) superior performance and robustness compared with existing methods for various untrained working conditions, as well as 2) efficient data inference even with small data size. Finally, the diagnosis results could be physically understood by displaying the classification decision in terms of the theoretical characteristic fault frequency (i.e., the characteristic fault order), indicating that SDGPI is a versatile and reliable diagnostic tool for unseen working conditions.2024-05-01T00:00:00ZReal-time estimation of longitudinal tire stiffness considering dynamic characteristics of tireDo, JongyongHyun, DongyoonHan, KyoungseokChoi, Seibum Bhttp://hdl.handle.net/10203/3178932024-01-17T08:00:19Z2024-04-01T00:00:00ZTitle: Real-time estimation of longitudinal tire stiffness considering dynamic characteristics of tire
Authors: Do, Jongyong; Hyun, Dongyoon; Han, Kyoungseok; Choi, Seibum B
Abstract: To enhance the effectiveness of active safety control, the tire–road friction coefficient (TRFC) must be precisely estimated, and the longitudinal tire stiffness coefficient is an important vehicle dynamic parameter to estimate TRFC. In this research, we present an observer that improves the performance of longitudinal tire stiffness coefficient estimation by applying tire dynamics that were previously applied in the lateral direction to the longitudinal direction. To begin, we model longitudinal tire dynamics using the relaxation length concept and validate the model using vehicle braking tests. We develop an observer that estimates the longitudinal tire stiffness coefficient by integrating the proposed tire dynamics and vehicle dynamics. The observer, which is based on an extended Kalman filter, can be applied to nonlinear systems and successfully removes noise from wheel speed measurement. The observer's estimation performance is verified using CarSim simulation and vehicle tests, and the results are compared to existing approaches that do not account for longitudinal tire dynamics. Even in the transient section when the vehicle begins accelerating, the difference between the estimate and the reference value is about 0.3% using the proposed method, but if tire dynamics are not taken into account, the estimate is 6.5% lower than the reference value.2024-04-01T00:00:00ZElastic size effect of single crystal copper beams under combined loading of torsion and bendingChoi, Jae HoonRyu, HyeminSim, Gi-Donghttp://hdl.handle.net/10203/3179202024-01-23T02:00:49Z2024-04-01T00:00:00ZTitle: Elastic size effect of single crystal copper beams under combined loading of torsion and bending
Authors: Choi, Jae Hoon; Ryu, Hyemin; Sim, Gi-Dong
Abstract: Among various strain gradient theories, the modified couple stress theory, which introduces a single length scale parameter as an additional material property, has garnered significant interest owing to its simplified portrayal of the material behavior. In this study, we investigated whether a single length scale parameter is sufficient to predict the mechanical behavior under two different type of strain gradients: torsion and bending. L-shaped beams made of single crystal copper with thicknesses ranging from 2.4 μm to 9.1 μm were fabricated, and loads were applied using an indenter. The contributions of bending and torsion were controlled by adjusting the loading position. Through these experiments, we demonstrated the existence of elastic size effect of single crystalline materials under strain gradients. Specifically, size effect was observed in both bending and torsion, with a larger effect observed in cases closer to pure bending. Moreover, we report that the modified couple stress theory and the modified strain gradient theory are not applicable for simulating size effect under combined loading. This discovery highlights the necessity for the development of a new theory capable of adequately simulating size effect under the intricate loading scenarios encountered in practical applications.2024-04-01T00:00:00ZExperimental and numerical investigation of cryogenic no-vent fill (NVF) process using adsorption on activated carbonKim, JinwookKim, Kyoung JoongBae, Jun HyukJeong, Sangkwonhttp://hdl.handle.net/10203/3182842024-02-27T05:00:29Z2024-04-01T00:00:00ZTitle: Experimental and numerical investigation of cryogenic no-vent fill (NVF) process using adsorption on activated carbon
Authors: Kim, Jinwook; Kim, Kyoung Joong; Bae, Jun Hyuk; Jeong, Sangkwon2024-04-01T00:00:00Z