【Background】
Lithium evolution can lead to a serious loss of lithium-ion battery capacity and cycle life, as well as internal short circuits and thermal runaway. The development of an in-situ real-time lithium evolution detection method is essential to ensure the safety and reliability of batteries. In addition, the early warning method of lithium evolution can bring new opportunities for the development of battery life extension strategies and optimization of charging strategies. The negative electrode potential is undoubtedly the most relevant feature for lithium evolution, however, the introduction of a reference electrode into the battery will affect the internal stability and cycle life of the battery. The ideal lithium evolution detection method should be in-situ and non-destructive, without altering the existing battery structure and increasing costs. In addition, the method should also implement lithium separation early warning to make trade-offs between fast charging rate and battery life in advance.
【Job Introduction】
Recently, Wang Jiajun's research group at Harbin Institute of Technology and others have developed a lithium evolution detection and early warning method based on negative electrode potential reconstruction and voltage prediction. The negative potential reconstruction model constructs a high-precision mapping between the battery voltage and the negative potential, which can reconstruct the negative potential without introducing a reference electrode. In addition, the negative electrode potential for the next 2 minutes is predicted by combining voltage prediction models. Based on the negative electrode potential in the next 2 minutes, the early warning of lithium evolution can be realized for the first time. In this study, only the battery voltage is used as input for lithium evolution detection and early warning, and in future work, the most appropriate charging strategy can be matched in real time before lithium evolution occurs, so as to achieve the fastest charging rate while maximizing battery life. The article was published in the top international journal Energy Storage Materials. Wang Han is the first author of this article.
【Content Description】
Firstly, a high-precision mapping between the battery voltage and the negative potential was constructed through the negative potential reconstruction model, and the negative potential reconstruction was realized without introducing a reference electrode. In the case of 0 mV and 50 mV negative electrode potentials as the detection standard for lithium evolution, lithium precipitation can be accurately detected, and the detection rates are 84% and 99.9%, respectively. Then, based on the current 2-minute battery voltage sequence, a voltage prediction model was proposed to predict the battery voltage in the next 2-minute. In addition, the negative potential for the next 2 minutes can be predicted by using a combination of previously established negative potential reconstruction models. Based on the negative electrode potential in the next 2 minutes, the early warning of lithium evolution can be realized for the first time with the negative electrode potential of 0 mV and 50 mV as the early warning criteria for lithium evolution, and the early warning rates are 74.7% and 98.7%, respectively. By analyzing the results of lithium evolution detection and early warning, the best online fast charging strategy and life extension strategy can be formulated. For example, if lithium evolution does not occur in the next 2 minutes, the charging rate can be increased to reduce the charging time. Otherwise, the ambient temperature can be increased and the charging rate can be reduced to avoid lithium segregation and prolong battery life.
Fig.1 Schematic diagram of lithium evolution detection and early warning method
Figure 2 Battery cycle data
Fig.3 Lithium evolution test results
Fig.4 Early warning results of lithium separation
Figure 5 Characterization validation
【Conclusion】
In-situ, real-time, non-destructive, sensorless, and accurate lithium evolution detection and early warning methods can ensure the safety of batteries and bring new opportunities for developing battery life extension strategies and optimizing charging strategies. The researchers developed a combination of a negative battery reconstruction model and a battery voltage prediction model to accurately detect and warn lithium evolution under different operating conditions and temperatures. The negative potential reconstruction model is used to reconstruct the negative potential based on the voltage, current, and temperature of the battery without introducing a reference electrode. Through the combination of the negative electrode potential reconstruction model and the battery voltage prediction model, the proposed method can accurately realize the early warning of lithium separation when the 50mV negative electrode potential is used as the standard for lithium separation, and the accuracy rate reaches 98.7%. The proposed method can be easily applied to BMS, which provides new possibilities for the development of adaptive intelligent batteries with safer, longer cycle life and faster charging speed.
Han Wang, Yajie Song, Xue Sun, Shengkai Mo, Cong Chen, Jiajun Wang*, Onboard In-situ Warning and Detection of Li Plating for Fast-charging Batteries with Deep Learning, Energy Storage Materials, 2024.
https://doi.org/10.1016/j.ensm.2024.103585
Source: Energy Scholar
Note: Most of the articles reproduced on this site are collected on the Internet, and the copyright of the article belongs to the original author and the original source. The views in the article are only for sharing and exchange, if it involves copyright and other issues, please let us know, and I will deal with it in a timely manner.