A Study on Improving Correlation between Analysis and Test Results for Vibration Fatigue Durability and Computational Fluid Dynamics

GeonHee Cheon･NamJin Kim･JinKyu Song･WonSeok Choi･JungSub Kim (Seojin Industrial)･GyuHo Shim (Eco-Plastic)

The purpose of CAE analysis is to anticipate the performance of vehicle components in advance of making the prototype. It is an important tool for effectively cutting the costs and time during the product development process. Thus, maintaining the accuracy of analysis results is critical and further study is necessary to improve the correlation with test outcomes. Consequently, this paper aimed to describe the correlation between analysis and test outcomes during the battery case and chassis frame development process under three topics including : vibration fatigue durability, pressure drop and electrodeposition coating (E-Coat).

A Study on the Development of High Strength Rear Trailing Arm for Electric Vehicles

HeeSang Gong･JinKyu Song･JinSung Kim･MoonSub Song･HoSang Park･DongHoon Kim (Seojin Industrial)

Due to global environmental regulations, the automotive industry is focusing on developing electric vehicles to reduce greenhouse gas emissions. The main parts of an electric vehicle is a battery pack, and the weight of the car is gradually increasing due to the increase in battery capacity. As the weight of the vehicle increases, the development of high strength chassis parts is required. This paper describes the process of developing a high strength rear trailing arm for EV vehicles. The main topics include formability verification for high strength steel and performance verification through test and CAE Analysis.

Development of Accuracy Improvement Technology for Surrogate Models using Shape Generation AI

Hiroaki Onodera･Noriko Ohtsuka･Mashio Taniguchi･Masatake Kimura (Toyota Motor)

To improve the performance prediction accuracy of surrogate models for exterior panels, a cycle that explores and detect a lack of training data in design space, creates new training data using 3D shape generation AI and CAE for the detected area and iteratively update the surrogate model is established. The process was applied to the surrogate model of hood outer rigidity performance and the prediction accuracy is improved.

Study of Surrogate Model Construction for Evaluation of Crash Box Performance

Yoshitaka Wada (Kindai University)･Asuka Endo (Marelli)･Yuki Okumoto (Mazda)･Kai Morimoto (DENSO)･Takasue Kikuchi (DAIHATSU MOTOR)

Expectations for the construction of surrogate models of physical problems using machine learning have been growing year by year. Understanding how to use and build surrogate models is an essential key to their practical application. In this study, we attempted to construct a surrogate model to predict the maximum reaction force and energy absorption for a rectangular cylinder shape that resembles a crash box. The machine learning methods used were neural networks and decision trees to construct the predictor. Since the introduction of parameters that represent the phenomenon in a more linear form is necessary to improve accuracy, input data design was conducted. An analysis of data in which the improvement of accuracy is hampered by differences in buckling modes is also reported. In this presentation, the results obtained through the activities of the AI-ML Technology Application WG in the Structural Strength Section Committee will be presented.

Study on the Body Torsional Rigidity of Personal Mobility Vehicles (PMVs) with Active Tilting Mechanisms

Tetsunori Haraguchi (Nagoya University / Nihon University)･Tetsuya Kaneko (Osaka Sangyo University / Nihon University)

From the perspective of vehicle dynamics, there is little research on the body torsional rigidity of PMVs with active tilting mechanisms. In this paper, we used a body torsional load balance model and conducted a comparative analysis with four-wheeled vehicles to examine the necessity of body torsional rigidity under various conditions. Additionally, we investigated the impact of body torsional rigidity on obstacle avoidance ability using a dynamic model, validating the effectiveness of this balance model.