With the increasing adoption of lithium-ion batteries (LIBs) as the battery of choice in electromobility, personal electronic devices, and other applications comes the challenge of ageing, which prevents the batteries from performing optimally and meeting their design intent. This is observed as a decline in power capability due to increased resistance and reduced capacity. Unfortunately, the cost of assessing batteries after the first use remains a daunting challenge. In our work, we propose an approach that performs fast preliminary grading based on resistance and capacity by first connecting old cells of the same chemistry and model in series with resistors to limit the branch current, and then connecting the branches in parallel to equalise the voltages. A Simulink model of NCR18650PF Panasonic cells with adaptive-series resistance is compared with a fixed-series resistance and found to improve the balancing time from over 24 h to just 8 h.

Nine Panasonic CGR18650E LCO-G (2550 mAh) cells from two different first-life applications were connected in parallel with a fixed resistance of 1.5 Ω until their terminal voltages were equal. Electrochemical impedance spectroscopy (EIS) was performed on the individual balanced cells over 0.1 Hz to 5 kHz, and the real, imaginary, absolute, and phase impedances were compared with the cells’ SOH at each frequency. Results show that the imaginary impedance in the 6.6 Hz frequency range exhibits a strong correlation coefficient (>0.98) with SOH, especially at a state of charge (SOC) of about 75–85% for the LCO cells. By selecting a sample from all cells that covers a wide range of SOH and performing a full-capacity check-up on them, a simple correlation between SOH and EIS measurements at different frequencies can be used to estimate the SOH of the other cells connected in the same parallel connection. This represents a considerable time savings in charge and discharge times for other cells in facilities that lack the capacity to cycle all cells simultaneously. There are also huge energy savings in not having to cycle all the cells. Therefore, it offers a more efficient approach to grading spent cells than performing full-capacity tests.