No. | Video | Title・Author (Affiliation) |
---|---|---|
1 | ◯ |
A Study on the Development, CAE Analysis, and Test Validation of a Cell Frame Assembly Module for Advanced Battery Systems GeonHee Cheon・Gun In・NamJin Kim・DongHoon Kim・JungSub Kim・Hyun Sung (SeoJin Industrial)・GyuHo Shim (ECOPLASTIC) This study presents the Cell Frame Assembly Module, an integrated structure combining the chassis frame and battery case to enhance EV performance. Replacing aluminum with steel and optimizing the layout improves structural integrity, reduces cost, and expands battery space for longer range. Virtual analysis (vibration, fatigue, impact, cooling) and prototype testing confirmed performance. A form gasket and new bolt-sealing structure ensured IP67 level sealing. This integration simplifies manufacturing, maximizes space, and enhances competitiveness. The paper details CFAM's design, CAE methodologies, and validation tests, offering practical insights for next generation EV platforms. |
2 | ◯ |
Experimental research on fire prevention involving electric vehicles (1st report) Yoshihiro Sukagawa・Koji Yamazaki (JARI) While the global adoption of electric vehicles (EVs) has been accelerating rapidly in recent years, addressing the risk of fires caused by thermal runaway in the high-voltage batteries installed in these vehicles has become a critical issue. Although EVs are generally considered to have a lower risk of fire compared to gasoline-powered vehicles, when a fire does occur, it tends to cause significant damage. It has also been reported that extinguishing fires involving high-voltage batteries requires a large volume of water and a prolonged period of time. This risk is particularly heightened following vehicle collisions, as battery damage increases the likelihood of fire. Therefore, special precautions must be taken when storing EVs after a collision. |
3 | ◯ |
A Study on the Development of Predictive Method for Structural Weakness of Bus Body in Concept Stage Using 1D Beam Model and Machine Learning gyuhee kim (Hyundai Motor) This study proposes a machine learning-based model that estimates joint strength levels using 1D simulation data, eliminating the need for 3D modeling. By leveraging the correlation between 1D and 3D analysis data, the model enables early-stage strength evaluation without detailed CAD or 3D construction, significantly reducing time and cost. The model was validated using domain knowledge, focusing on vulnerable joints. The process, which previously took weeks, was shortened to under an hour with over 80% accuracy. Future work includes expanding training datasets and developing features to enhance accuracy by incorporating diverse joint characteristics. |
4 | ✕ |
A Study on Suspension Input Load Prediction using RNN-based Virtual Sensor for Durability Application SEUNGWAN SON SON・DAEJIN KIM (Hyundai Motor) In this study, a method was developed to predict wheel input forces and suspension component input loads without wheel force transducers and suspension load cells. For the development and verification of the process, a set of training data was acquired by measuring the wheel forces and the suspension forces in one vehicle. Subsequently, a training model was created using the Recurrent Neural Network model. Finally, the wheel force and suspension load were predicted using a vehicle equipped with a low-cost single Inertial Measurement Unit, and the model performance was verified by comparing the predicted values with the measured data. |
5 | ◯ |
Efficient Optimization of Component Placement Using Replica Exchange Method Koichi Seki・Masaya Michishita・Hideaki Bunazawa (Toyota Motor) Efficient placement of components in vehicle development is crucial for performance enhancement and cost reduction. This study applies the replica exchange method to the placement problem of dummy components, comparing it with traditional hill-climbing and simulated annealing methods. The results demonstrate that the replica exchange method enables more efficient optimization, achieving improved component placement. |
6 | ◯ |
Optimization of Frame Cross-Sectional Shapes Using Kernel QA Wataru Shimoda・Toshiki Kondo・Takehisa Kohira (Mazda) In recent automotive development, the rapid increase in design parameters necessitates the practical application of efficient optimization techniques such as Quantum Annealing (QA). In this study, we optimized the cross-sectional shapes of automobile frames using QA, and applied kernel regression to reduce computation time. Additionally, by refining the binarization process, we achieved improved the performance even with finer discretization. |
7 | ◯ |
Prediction Technology Development for Millimeter Wave Radar Transmittance with Paint Composition in New Color Development. Naoya Osaki・CHEETUCK HO・Hironori Tsutsui・Keiko Ukishima・Natsuko Kaji・Tomoyuki Okamoto (Nissan Motor) With the proliferation of ADAS vehicles, radar transmittance of coatings has become a critical issue. Traditionally, coating evaluation was performed through measurements after paint development, leading to development rework if composition changes were necessary. To address this, we developed a technology that simulates radar transmittance from paint composition. This innovation allows for early evaluation of coatings before the completion of paint development. |