This team consists of 6 students from the School of Electrical Engineering and the School of Science at Xi’an Jiaotong University (Wang Ziqiao, Li Dong, Zhao Hongfei, Wang Ning, Zhang Ranjie, Huang Ye), who have conducted a series of studies aimed at improving the calculation accuracy of the battery state of charge (SOC) in battery management systems (BMS).
Due to the depletion of fossil fuels and the impact of global warming, “zero emissions” pure electric vehicles have developed rapidly under the joint promotion of governments and automobile manufacturers. In recent years, the production and sales of electric vehicles in China have continued to grow, maintaining the world’s top position. In the first three quarters of 2018, the national production and sales of electric vehicles reached 735,000 units and 541,000 units, respectively, representing year-on-year growth of 73% and 81.1% respectively.
18650 battery pack
Model X battery pack
The battery system is the power source of electric vehicles and is the most core component of the entire industrial chain. Taking Tesla Model S as an example, the cost proportion of its battery system (lithium battery+ battery management system) is 56%, while the cost proportion of traditional car engines is only about 15%~25%. By 2016, the cost proportion of the battery system had decreased, and the cost structure had changed, with the cost of individual batteries accounting for 83%. At the same time, ensuring the safety and lifespan of battery packs is also crucial for the development of electric vehicles. To effectively solve these issues, research on battery management systems (BMS) is indispensable.This project uses Tesla’s main battery——PanasonicNCR18650B battery as the test battery, selecting Long Short-Term Memory networks (LSTM) to enhance the SOC calculation accuracy of the BMS, and based on this method, a prototype of theBMS was developed.
We compared various electrode materials for 18650 batteries and selected lithium iron phosphate batteries. We conducted mixed pulse power characteristic experiments (HPPC), testing the dynamic response capabilities of the batteries at different depths of discharge according to the battery testing manual published by the U.S. Department of Energy, and performed a large number of charge and discharge experiments.
HPPC pulse sequence
HPPC testing process 1
HPPC testing process 2
Next, we designed aBMS scheme (as shown below)—— a battery analog quantity acquisition circuit based onBQ40Z50 chip andLSTM-basedSOC calculation software, and created a measurement and control module. The circuit collects real-time voltage, current, temperature, and other analog quantities from the battery and transmits them to theMCU. TheMCU, after determining the working state, calculates theSOC of the battery. The working state andSOC are displayed in real-time on a mobile APP via Bluetooth, allowing users to control the battery’s charge and discharge state with their phones, and further achieve balance between individual cells in the battery pack. The control process is implemented by theBQ40Z50 acquisition circuit andBQ2961 protection circuit. ForSOC calculation, we utilizeLSTM to train the experimental data, feeding back the corresponding parameters to theBMS, enabling theBMS to derive real-timeSOC values based on the collected analog quantities.LSTM is a type of time recursive neural network (RNN), suitable for handling and predicting events with longer intervals and delays in time series, and has advantages in predicting the battery’sSOC.
BMS overall design concept
Hardware system architecture diagram
LSTM model
Finally, we conducted joint debugging of the 18650 battery with the designed software and hardware system to obtain the corresponding data. The results show that thisBMS can achieve: collection of battery current, voltage, and temperature; judgment of battery status andSOC calculation; battery balancing; display of remaining battery power on the screen; fault alarm and automatic protection; communication with mobile phones via Bluetooth; and control of battery charge and discharge states through a mobile mini-program. In terms of testing accuracy, the SOC calculation accuracy for the same type of battery under test conditions is within 4%. After 300 charge and discharge cycles, its accuracy remains within 4% (the error mainly comes from the differences in individual battery parameters), indicating that this system has high accuracy in calculating theSOC of the battery.
Neural network prediction results 1
New battery data
300 charge and discharge cycles battery data
Results display:
Project demonstration video
Currently, the project has applied for: one invention patent (application number: 201910261973.3), one utility model patent (application number: 201920450205.8), and has reached a cooperation intention with Dongguan Paishida Electronic Technology Co., Ltd.
Moreover, the project has been recommended by several experts and professors:
Academician of the Chinese Academy of Sciences, Foreign Academician of the National Academy of Engineering, 863 Program Expert Professor Yao Xi
Academician of the Chinese Academy of Sciences Professor Tao Wenquan
Deputy Director of the Department of Electronic Engineering at Tsinghua University Professor Deng Beixing
Director of the Institute of Power Electronics and Motor Systems at Tsinghua University Researcher Sun Kaili
Huazhong University of Science and Technology, National Excellent Young Scientist Fund Recipient Professor Wang Kangli
Advisor:
Ding Shujiang:
Ding Shujiang, born in 1978 in Harbin, Heilongjiang Province, is a professor and doctoral supervisor at the School of Science, Xi’an Jiaotong University. He is a “New Century Excellent Talent” of the Ministry of Education and a “Young Science and Technology Star” in Shaanxi Province. He is a specially-appointed professor at Xi’an Jiaotong University and a selected candidate for the Young Top Talent A category. His research work involves the design, preparation, and application of polymer/inorganic nano-structured composite materials in electrochemical energy storage (lithium/sodium-ion batteries, lithium-sulfur batteries, solid-state batteries, fuel cells), sensors, electric drive, and electrocatalysis. He has published over 140 papers as the first author or corresponding author in journals such as Nat. Commun., J. Am. Chem. Soc., Angew. Chem. Int. Ed., Energy Environ. Sci., Adv. Energy Mater., Adv. Funct. Mater., Nano Energy, Chem. Mater., J. Mater. Chem A, among which 14 papers have been selected as highly cited papers in the “Essential Science Indicators (ESI)”. He has also served as a reviewer for several prestigious international academic journals. His ongoing projects include the National Natural Science Foundation’s general and youth projects, doctoral fund projects, and Shaanxi provincial fund projects. His awards include: the Shaanxi Youth Science and Technology Award in 2016, the First Prize of the Shaanxi Provincial Higher Education Science and Technology Award in 2017 (as the first contributor). In November 2018, he was selected as a highly cited scientist in the interdisciplinary field by Clarivate Analytics, and in January 2019, he was selected as a highly cited scholar in China by Elsevier.
XU Jun:
Research direction: Electric vehicles, new energy, robotics, and related technologies
He has long been committed to research in energy management technology for electric vehicles and control of complex mechatronic systems, achieving fruitful results in the fields of new energy vehicles, electric vehicle (electric vehicle) battery management, energy management, and system control. He has received one China Industry-Academia-Research Cooperation Innovation Award and one First Prize of the Shaanxi Provincial Higher Education Science and Technology Award (as the first contributor). He has led more than 20 national, provincial, and corporate cooperation projects. He has co-authored one English monograph and published over 50 high-level academic papers, including 2 ESI highly cited papers. He has applied for more than 30 invention patents and one software copyright. He has served as a reviewer for more than 20 international high-level journals and multiple international conference papers, and is an outstanding reviewer for JPS and Energy.
Zhang Hong:
Zhang Hong, Professor, Vice Dean of the School of Microelectronics, Xi’an Jiaotong University
Research direction
1 High-speed communication system integrated circuit research and design (including ADC design, RF transceiver analog front-end, high-speed interface circuits)
2 New energy system battery management integrated circuit research and design
3 Implantable biomedical chip research and design
WeChat mini-program (needs to cooperate with my team’s official hardware for practical use)