The electric vehicle (EV) market has been rapidly expanding due to supportive policies, increased public awareness of climate change, and declining battery prices. By the end of 2021, the global EV stock has surpassed 16 million, as reported by the International Energy Agency (IEA). However, as EVs become more prevalent, managing the significant quantity of retired batteries poses a great challenge to the industry and the government. These batteries are reported to retain approximately 80% of their designed capacity and can be given a second life in less demanding scenarios. In this context, the second-life battery (SLB) concept has been introduced as a solution, offering the potential to extend battery life and reduce the costs associated with manufacturing new batteries.
Moreover, the increasing development of renewable energy generation has led to the popularity of photovoltaic (PV) plus energy storage systems. SLBs have emerged as a promising option for energy storage in these systems. However, before SLBs can be widely adopted, their reliability, cost-effectiveness, and sustainability must be thoroughly evaluated. To address this need, a research group led by Assistant Professor Xuan Zhang, Associate Professor Guodan Wei, and Associate Professor Guangmin Zhou at Tsinghua Shenzhen International Graduate School (Tsinghua SIGS) has undertaken research on SLB evaluation, application and recycling.
Fig. 1. Recycling and application process of SLB cells.
The research team has proposed a fast screening and capacity estimation framework for Li-ion batteries based on rapid pulse tests and machine learning methods. Compared with the traditional long-time constant current-constant voltage (CC−CV) charging and constant current (CC) discharging test, the proposed framework can reduce the evaluation time by more than 80% and achieve an average accuracy of 95%. The team’s study shows that the rapid pulse test can reflect the changes in internal resistance and voltage stage of the battery as aging deepens, thus reflecting the degree of battery aging. When a batch of retired batteries with unknown capacity is given, it is only necessary to perform a rapid pulse test on them and input the obtained pulse voltage curves into the proposed framework to output their estimated capacity. This new strategy, combining the rapid pulse test and machine learning method, provides efficient assessment of the remaining capacity of Li-ion batteries and online diagnosis of State of Health (SoH).
The team has also developed an online SLB SoH estimation method, which leverages the Kalman filter's estimation power for combining short-term and long-term prediction results to achieve higher accuracy. Such SoH estimation is a determinant factor in deciding whether the batteries should be granted a second life or put into recycling. Furthermore, the team has proposed an SLB optimal dispatch model that considers degradation with the online SLB SoH estimation method. The proposed approach was compared to an alternative dispatch approach that considers degradation with a State of Charge (SoC) based model and a dispatch approach with no degradation consideration. The results show that the proposed approach leads to less battery degradation and costs and demonstrates the batteries' complementary behavior in providing energy balancing and energy arbitrage.
Fig.2. Illustration of estimating battery capacity using short-time pulse tests and data-driven methods (different colors indicate varied battery capacities).
Fig. 3. The optimal dispatch approach considering degradation with online SoH estimation.
Additionally, the research team has proposed an SLB pricing method based on the batteries' SoH by combining the market average price associated with the cost of recycling and repurposing retired batteries. The model takes into account the quantity of EVs, the battery replacement rate, remaining capacity, and degradation curves. The pricing information enables economic evaluations for SLB applications, facilitating stakeholders to make informed decisions on the investment and adoption of SLBs. The team has further proposed sustainability indexes for evaluating the sustainability of SLB applications. The performance of SLBs in grid-connected PV-battery systems under three typical load scenarios (including stationary load, single peak load, and double peak load scenarios) is examined. The results show that the sustainability of the SLBs depends mainly on their degradation curves, including the SoH and degradation rate of the batteries. Higher initial SoH and lower degradation rates are favorable for improving the sustainability performance of SLBs. However, when considering the purchasing costs of SLBs, cheaper options may remain competitive, even with a fast degradation rate.
Fig. 4. Structure of the sustainability evaluation of second-life battery applications in grid-connected PV-battery systems.
The research results on the fast clustering of SLBs were published in ACS Energy Letters in an article entitled "Fast clustering of retired Lithium-ion batteries for secondary life with a two-step learning method." Aihua Ran and Zheng Liang, both from Tsinghua SIGS, are the co-first authors of the paper. The paper's corresponding authors are Associate Professor Guangmin Zhou, Assistant Professor Xuan Zhang, and Associate Professor Guodan Wei from Tsinghua SIGS.
The research results on SLB optimal dispatch with online SoH estimation, and SLB pricing and sustainability evaluation have been published in Renewable and Sustainable Energy Reviews and Journal of Power Sources, respectively. The titles of the papers are "Optimal dispatch approach for second-life batteries considering degradation with online SoH estimation" and "Sustainability evaluation of second-life battery applications in grid-connected PV-battery systems." The first author of both papers is Ming Cheng from Tsinghua SIGS and the corresponding author is Tsinghua SIGS Assistant Professor Xuan Zhang.
Link to full articles:
1. https://pubs.acs.org/doi/10.1021/acsenergylett.2c01898
2. https://doi.org/10.1016/j.rser.2022.113053
3. https://doi.org/10.1016/j.jpowsour.2022.232132
Written by Ming Cheng & Zheng Liang
Edited by Alena Shish & Yuan Yang
