Dalian Institute of Chemical Physics, Chinese Academy of Sciences and others developed a deep learning model for battery life prediction

This website reported on September 3 that accurate prediction of lithium battery life is crucial for the normal operation of electrical equipment. However, accurate prediction of battery life faces challenges due to the nonlinearity of the battery capacity degradation process and the uncertainty of operating conditions. The Chinese Academy of Sciences stated that the team of researcher Chen Zhongwei and associate researcher Mao Zhiyu from the Power Battery and System Research Department of the National Key Laboratory of Energy Catalytic Conversion of the Dalian Institute of Chemical Physics, together with Professor Feng Jiangtao of Xi'an Jiaotong University, have made progress in battery health management research. Relevant research results have been published in the Journal of Transportation Electrochemistry of the Institute of Electrical and Electronics Engineers (DOI: 10.1109/TTE.2024.3434553 attached to this site).

1. According to reports, the research team has developed a new deep learning model, which overcomes the traditional method's reliance on a large amount of charging test data, provides new ideas for real-time battery life estimation, and achieves End-to-end evaluation of lithium battery life.

2. This model serves as a component of the first-generation battery digital brain PBSRD Digit core model developed by the team, providing a solution for intelligent battery management.

   

   1. Battery life prediction model based on deep learning

This study proposes a deep learning model based on a small amount of charging cycle data. This model captures and fuses multi-time scale hidden features through the Vision Transformer structure and efficient self-attention mechanism to achieve accurate prediction of the battery's current cycle life and remaining service life.

2. Prediction accuracy and generalization ability

At the same time, the model combines the remaining service life and The current cycle life prediction errors are controlled within 5.40% and 4.64% respectively. Furthermore, the model is still able to maintain low prediction errors when faced with charging strategies that do not appear in the training data set, demonstrating its zero-short generalization ability.

3. Integration with Battery Digital Brain

This battery life prediction model is the first generation of Battery Digital BrainPBSRD Digit component. The researchers further improved the accuracy of the system by integrating the above model into the system.

4. Deployment and Application

Currently, the battery digital brain system serves as the energy management core for large-scale industrial and commercial energy storage and electric vehicles and can be deployed on cloud servers and client embedded devices.

5. Model optimization

This model balances prediction accuracy and computing cost, improving the battery digital brain for life estimation application value. In the future, the team will further optimize the model through model distillation, pruning and other methods to improve the system's robustness and resource utilization.

The above is the detailed content of Dalian Institute of Chemical Physics, Chinese Academy of Sciences and others developed a deep learning model for battery life prediction. For more information, please follow other related articles on the PHP Chinese website!