Safety qualification methodology for assessing used lithium-ion batteries from vehicles for second-life applications

The growing number of retired lithium-ion batteries from electric vehicles (EVs) presents significant environmental and economic challenges. Repurposing these batteries for second-life applications offers a promising solution. However, as batteries experience several degradation mechanisms during use that affect their behavior, ensuring safety throughout their lifecycle remains a complex challenge. Additionally, the lack of information about the aging history of the battery at End of First Life (EOFL) makes safety assessments even more difficult.

This thesis investigates the potential of second-life batteries, identifying over 65 applications: 41 mobile, 7 semi-stationary, and 17 stationary. Among these, Automated Guided Vehicles (AGVs) and industrial Energy Storage Systems (ESSs) for renewable energy firming emerge as most promising based on technical, legal, and economic criteria. The diverse range of applications is characterized by distinct electrical, thermal, and mechanical loads, which can differ significantly from those experienced during the batteries' first life. This underscores the need to study the effects of various load cases on aged batteries to ensure safety across all conditions.

To address safety concerns at EOFL, this thesis introduces a safety qualification strategy for second-life batteries through an empirical mathematical model. The model evaluates safety using several qualification parameters, including State of Health (SOH), Coulombic Efficiency, Area VdQ, parameters derived from the Voltage Relaxation Profile (VRP) and parameters derived from the hysteresis Open Circuit Voltage (OCV) curves. These parameters provide a comprehensive description of battery aging mechanisms. In addition, the qualification parameters are chosen to be application oriented: all can be measured with sensors commonly found in battery packs and are extracted from a simple charging-waiting-discharging cycle. This makes the protocol easy to implement and highly relevant for real-world applications.

Differently aged pouch cells were subjected to various load cases, specifically cycling with high C-rate, mechanical indentation, and mechanical shock. This investigation aimed to evaluate the impact of these different loads on the evolution of the measured electrical qualification parameters. The results showed that each load case affected the qualification parameters in distinct ways. High C-rate significantly influenced SOH and the OCV curve but did not affect the VRP voltage drop. In contrast, mechanical loads such as indentation and shock significantly impacted Coulombic Efficiency and Area VdQ while not affecting the OCV parameters. These findings highlight that different loads impact the battery in distinct ways, underscoring the importance of load-specific safety considerations in second-life battery applications.

Additionally, differently aged cells were subjected to prolonged electrochemical cycling to determine the safety margins for each qualification parameter, as the previous load cases did not lead the battery to a safety-critical condition. These safety margins enable the definition of thresholds for unsafe battery conditions. Specifically, the defined safety margins are as follows: SOH Discharging at 51.46%, Coulombic Efficiency at 79.38%, VRP Voltage Drop at 89 mV, VRP Slope at -7.09 µV/s, and an OCV increase at 25 Ah at 1039 mV. The safety margins are integrated into the empirical mathematical model and are used to detect safety-critical battery conditions. Furthermore, the safety margins can be adjusted to align with the expected second-life operational conditions. For example, in high C rate second-life applications, where high charging and discharging currents accelerate battery degradation, or in mobile second-life applications, where mechanical loads such as shock or indentation further degrade the battery, the margins can be set more conservatively. More accurate knowledge of the expected second-life load cases can lead to better calibration of the safety margins.

In this work, a safety qualification approach has been developed to assess safety at EOFL, enabling second-life use. This demonstrates that repurposing batteries for second-life applications is a viable strategy. The approach not only facilitates the transition from first to second life but, by being application oriented, also lays the foundation for continuous monitoring of battery safety and the definition of an online State of Safety (SOS). In the future, further research can refine the SOS framework and integrate it into Battery Management Systems (BMS). Additionally, given the high variety of potential second-life load cases, a more comprehensive study of their effects is recommended to enhance the transferability of the safety qualification model. This approach has the potential to enable predictive maintenance and prevent unnecessary replacements, ultimately extending battery lifespan, reducing waste, and supporting significant environmental and economic benefits by ensuring that batteries are used safely and efficiently in real-world settings.