RECOVERY PROCEDURE FOR DEEPLY DISCHARGED LiFePO4 TRACTION BATTERY

Presented paper deals with possibility of recovery of damaged traction batteries by deep discharge. For the test of proposed charging algorithm, new, unused cell was selected, the deeply discharged condition of which was caused by the self-discharge during the improper storage. The cell had significant damage of package in the central part. For the battery testing, experimental set-up was realized for automated recovering procedure and other tests of cells. For verification of the proposed algorithm, a recovered cell was compared to a reference/new cell by testing the delivered ampere-hours for various discharging current. The final evaluation shows that the proposed algorithm for recovery of the deeply discharged cell can recover up to 70% of the nominal cell capacity.

reuse in producing new batteries. However, manufacturing batteries from the recycled materials is five times more expensive than manufacturing from the new ones, so for many manufacturers it is more cost-effective to produce batteries from the new materials, [7][8][9][10].
After the battery's state of health falls to 80% and battery is not suitable for original application, it is possible to continue using it in other application where demands on battery performance are not high, for example in storage for photovoltaics, [9][10][11]. Recovery of damaged batteries as well as recycling is therefore a topic, which must be accepted if sustainability related to environment and costs are considered, [12][13][14]. Therefore, within the presented paper, the experimental methodology for deeply discharged battery recovery is being introduced.

Selected cell and test-stand
For testing of the battery recovery algorithm, WINA 3.2V 60Ah LiFePO4 cell was selected. Parameters of the cell are listed in Table 1. This cell has significant damage of package in central part of the cell. The width of the cell reached 43.8 mm while the width of the new cell, specified by manufacturer is 36 mm, (Figure 1). The deep discharge condition was confirmed by measurement of the open circuit voltage, which was only 2.04V. The minimal voltage of the selected cell in datasheet is 2.5V. Selected cell was new, unused and its deep discharge was caused by improper storage and cell voltage decreased below the minimal cell voltage because of the self-discharge. The permissible selfdischarge, specified by the cell manufacturer is a loss of 3% of capacity over a month period, [15].

Introduction
Lifetime of lithium batteries can be divided into four phases. The first phase starts at production, where batteries are made and tested. If the test failed, batteries are directly recycled. Batteries are produced as uncharged, whereupon batteries must be charged before the first use, which is the next and most important phase of the battery lifetime. Initial charging and discharging are called formatting and has significant impact on the battery performance. The most important parameters during the formation are number of charging/discharging cycles, battery current and battery voltage. This process is executed by the manufacturer during the manufactured batteries testing. Those tests are performed in order to detect bad batteries with lower capacity. Most often, batteries are discharged to 50% of capacity and stored several weeks for testing the self-discharge and open circuit voltage (OCV) is measured. Batteries, capacity of which falls below 47 %, fails the test and are recycled. The next phase is using batteries in application, which has significant impact on the battery lifetime, based on charging/discharging current, charging voltage and other parameters, [1][2][3]. The next conditions, which can impact the battery lifetime are overcharging or over-discharging. End of the battery lifetime is the last phase and usually occurs at 80% of state of health of a battery. Batteries can be recycled by various methods, based on the battery chemistry. For the lead acid batteries, about 99% of lead from used batteries is reclaimed. From the lithium-ion batteries, lithium-iron-phosphate (LiFePO4) is possible to recycle for up to 80% of the batteries' material, [4][5][6]. Depending on the battery active materials, it is possible to recover cobalt, manganese and nickel for D A N K O e t a l .
electronic load KIKOSUI PLZ 100W and NI PXI-1031 for control, temperature measurement and logging (current/ voltage/temperature). To prevent the hazardous situation, the power line of a battery is protected by a mechanical switch. By using the LabVIEW application of NI PXI it is possible to create a sequence of charging or discharging of the cell.

The test-stand for battery recovery testing
To secure safety, the recovery procedure and other tests of traction batteries were executed on a designed test stand, which can be seen in Figures 2 and 3. The test stand consists of a metal box where the battery is located, programmable power supply EA PSI 8080-60, programmable    recommendation from manufacturer, to 20 A, which is 1/3 of capacity of a cell. Maximum of charging voltage was selected to 3.65 V. For the given battery, application of six sequences was realized in order to achieve required OCV on the device, while 16 hours of resting period was applied between individual sequences. Figures 5-7 show time waveform of the cell voltage during each sequence application. It is seen that voltage increased from 2.04 V up to 3.19 V at the end of the first sequence, while during the last sequence the voltage level

Charging algorithm for recovery of the deeply discharged cells
Algorithm for recovery of the deeply discharged cells consists of 30 steps of charging and 30 steps of pause, mutually alternating, while each steps lasts 1 second (Figure 4). After that, battery is resting for 5 minutes for a cell to regenerate. The whole algorithm lasts 126 minutes; it consists of 21 sequences, as can be seen in Table 2.
Amplitude of the charging current was selected, following

Verification of the recovery algorithm
Recovery was tested trough delivered ampere-hours. Before the testing, it is required to fully charge the recovered cell. Selected cell was charged by the CC&CV (Constant   (Figure 8, right), which is for 16.932 Ah less and surface temperature reached was 35.703 °C.

Figure 7 Voltages of cells during the fifth sequence (left) and sixth sequence (right) charging
Delivered ampere-hours of the new cell, discharging with current 20 A, was slightly higher than the nominal capacity of a cell, 60.869 Ah and the surface temperature reached 35.397 °C. Delivered ampere-hours of the recovered cell was only 43.608 Ah, which is for 17.261 Ah less than the nominal capacity (Figure 8, left)  The third discharging current for the test was 60 A (Figure 9, left). At this discharge current, the new cell was able to deliver 57.862 Ah, while the recovered cell was able delivered only 39.709 Ah. The surface temperature of the new cell during the test was 41.502 °C and temperature of the recovered cell was 38.403 °C. Difference of delivered ampere-hours in this case was 18.117 Ah. The last test was realized with discharging current 80 A (Figure 9, right). In this case, the new cell delivered 55.965 Ah and the recovered cell delivered only 35.886 Ah. Temperature of the new cell during the test was 44.168 °C and temperature of the recovered cell was 40.881 °C. Results of the tests are listed in Table 4.

Conclusions
This paper deals with the recovery algorithm of the lithium-iron-phosphate traction battery damaged by the deep discharge. Deeply discharged condition was caused by improper storage and it was confirmed by measurement of the open circuit voltage (OCV), which was only 2.04 V, while the minimal voltage, specified by the manufacturer, should not drop below 2.5 V. The traction cell also had visible deformation in the central part, width in this part was 43.8 mm, while width of the new cell, specified by the manufacturer, is 36 mm. The recovery procedure is based on the charging process, whereby the