Revolutionizing Battery Management: AI-Powered Digital Twin for Predictive Maintenance and Enhanced Performance

Nikolaus Keuth･Gerhard Schagerl (AVL List)

This presentation introduces an AI-powered Digital Twin technology designed to enhance battery safety and reduce warranty costs. Leveraging data from development, telematics, and in-vehicle usage, it offers accurate state-of-health monitoring, anomaly detection, and range optimization. Key benefits include a 97% identification rate of battery issues a month before occurrence, a 92% reduction in recall volume, and thrice more accurate range predictions. The AI-powered Digital Twin uses machine learning for continuous BMS improvement and a scalable analytics backend, enhancing battery performance and safety significantly.

Development of Slip Rate Estimation Method Based on Torque Control of Four-Wheel In-Wheel Motor

Toshiyuki Ajima･Wataru Hatsuse･Masaru Yamasaki (Hitachi)

In order to control the four-wheel in-wheel motors and improve vehicle characteristics, a slip rate estimation method using motor torque has been developed. In this presentation, we report a basic study of slip rate estimation method based on motor torque control of each wheel during driving, and the effectiveness of the method is verified on a test vehicle.

High fidelity IPMSM model used for sensor-less algorithm comparative study

Marek Lazor･Petr Kuchar･Xiang Wang (Ricardo PLC)

Mathematical models of Interior Permanent Magnet Synchronous Motors (IPMSM) are well known in industry, however they are normally low fidelity or computationally intensive. Recent interest has focused on higher fidelity models, known as Digital Twins. These models complement the development of advanced control techniques, such as sensor-less algorithms, where sensitivity of position estimation accuracy to minor parameters like cross-coupling terms can be understood. These algorithms enable reduced cost and increased robustness in automotive applications. This paper presents a high-fidelity digital twin model, a comparative study of published sensor-less algorithms, and an innovative sensor-less algorithm allowing low-speed operation without high-frequency injection.

Development of Next-Generation e: HEV Control System for Synchronization with Driver's Emotion

Ryosuke Narimoto･Sadaharu Maeda･Yohei Ukai･Masatoshi Saito･Shinobu Kurachi･Naoya Murata･Hiroki Gunji･Kazuki Shiki･Akari Nagakura (Honda)

Due to global trends towards environmental protection, the electrification of automobiles is accelerating. However, pursuing emotional value in electric vehicles is also important in automobile development. In this study, a new control system has been developed for the e: HEV system that synchronizes with the driver's emotions to enhance the emotional value of electric vehicles, and confirmed performance that sets a new benchmark for electric vehicles.

Changes in dQ/dV depending on battery abnormalities

Hyunjun Jo･Hyunjun Jang･Sijoong Kim･Taekyu Kang (Hyundai Motor)

As the adoption of electric vehicle(EV) expands, the need for diagnostic technology for the safety of battery is gaining attention. A variety of abnormal conditions may occur in battery, each leading to distinct impacts. The Abnormalities are dependent on each other and affect each other, which makes difficult to understand the exact occurrence mechanism of the event. This research suggests the relationship with the change in dQ/dV depending on diverse battery abnormalities. It shows that dQ/dV changes occur under abnormal conditions, enabling the diagnosis of abnormal cells and also the classificaiton of the abnormalities. We expect the abnormal cells can be identified through monitoring the changes in dQ/dV.