A method for early prediction of battery cycle dive

Abstract

The application discloses a method for predicting battery cycle diving in advance, which comprises the following steps: S1, regularly placing a battery; S2, charging the battery to the upper limit of a voltage design in a constant-current charging mode at a rated current, and then charging the battery to 0.02-0.05C of a current in a constant-voltage charging mode at the upper limit of the voltage; S3, regularly placing the battery; S4, discharging the battery to the lower limit of the voltage design of the battery in a constant-current discharging mode at the rated current; S5, regularly placing the battery, and counting a voltage change trend after rebounding and a voltage change rate dV5; S6, repeating S2-S5 until the end of a designed cycle life of the battery; and S7-S8, adding the following steps during the cycles S2-S5: S7, discharging the battery to the lower limit of the voltage design in a constant-current discharging mode at 0.05C-0.3C of a current; and S8, regularly placing the battery and counting a voltage change trend after rebounding. The method can reduce the dependence on high-precision testing equipment as much as possible, the errors among the parameters can be offset, and the purpose of improving the prediction of battery diving is achieved.

Description

Technical Field

[0001] This invention relates to the field of new energy batteries, and particularly to cycle life testing of lithium-ion batteries. Background Technology

[0002] A battery cell is a rechargeable secondary battery cell that can be repeatedly charged and discharged. It consists of main components such as anode and cathode plates, separator, electrolyte, and mechanical parts. According to national standards, each batch of battery cells must undergo cycle life testing. However, cycle life testing is time-consuming and costly. For conventional battery cells, after a certain number of cycles, the final cycle life of the cell can be roughly estimated through data prediction. However, for battery cells that "plummet" in life during the test, there is currently no good way to predict their lifespan, and this method interferes with data prediction, seriously affecting the accuracy of the prediction. Current reports on cycle lifetime predictions, whether in literature or patents, are based on conventional decay patterns, such as the loss of active lithium, characteristic parameters like SEI films, or data-driven methods like AR / ANN / RVM. All of these methods have drawbacks. Actual testing of active lithium is very difficult, requiring sophisticated testing equipment; general methods cannot measure the loss rate of active lithium. SEI films are even more difficult to characterize, potentially requiring additional high-precision tools. Data-driven methods demand extremely high accuracy from existing data and require minimizing external interference; otherwise, the predicted model will differ significantly from reality.

[0003] In existing technologies, the testing process is basically fixed. We can obtain the charge and discharge capacity, energy, and end voltage data of the battery cell for each cycle. Usually, we pay more attention to the capacity and energy decay, and less attention to the voltage. Generally, when looking at the capacity decay, we can only see the general trend of decay from the initial capacity, but it is difficult to judge whether the battery cell will suddenly fail. The energy data we use as an auxiliary indicator is also difficult to predict the sudden failure of the battery cell. Based on the above reasons, this application designs a method to predict the battery cell cycle failure in advance. Summary of the Invention

[0004] The purpose of this invention is to solve the problems in the prior art by proposing a method for predicting the cycle drop of battery cells in advance.

[0005] A method for predicting battery cell cycle drops in advance includes the following steps: S1: Time out the battery cells; S2: The battery cell is charged to the upper limit of the voltage design using the rated current constant current charging method, and then charged to the upper limit voltage constant voltage to the current of 0.02~0.05C to end; S3: Time out the battery cells; S4: Discharge the battery cell to the lower limit of the battery cell's voltage design using a constant current discharge method with the rated current. S5: Time the battery cell to rest, and statistically analyze the voltage change trend after resting, and calculate the voltage change rate, denoted as dV5. Count the number of points where the voltage change rate is continuously greater than 0, and denote it as condition one. S6: Repeat S2 to S5 until the designed end number of cycles for cell life. Every 10 to 100 cycles of S2 to S5, add the following steps S7 to S8: S7: Discharge at a constant current of 0.05C to 0.3C to the lower limit of the voltage design; S8: Set aside the battery cell, statistically analyze the trend of the rebound voltage after setting aside, and calculate the trend of the change of the voltage at the end of S5 minus the voltage at the end of S7 in the same cycle. Statistically analyze the rate of change of the voltage difference, denoted as d(V5-V7). The number of consecutive decreases of d(V5-V7) is denoted as condition two. S9: Count the number of conditions 1 and 2, and filter out the portions of condition 1 greater than 10 and condition 2 greater than 3 as abnormal states.

[0006] Preferably, the resting time of S1 is set to 5 min to 1 h.

[0007] Preferably, the S3 resting time is set to 5 min to 1 h.

[0008] Preferably, the resting time of S5 is set to 5 min to 1 h.

[0009] Preferably, the resting time of S7 is set to 5 min to 1 h.

[0010] Preferably, the temperature of the test environment is maintained at 25±3℃.

[0011] Compared with existing technologies, the advantages of this invention are: 1. The method is simple, only requiring an additional step in the process design.

[0012] 2. The effect is obvious, with no additional cost and no negative impact.

[0013] 3. This invention can minimize the dependence on high-precision testing equipment and improve the prediction of battery cell failure by canceling out the errors between various parameters.

[0014] 4. It can accurately select cells that have shown abnormalities in the early stages, shorten the testing cycle, and greatly save testing resources. Attached Figure Description

[0015] Figure 1 This is a flowchart from the present invention.

[0016] Figure 2 This is a data diagram of the capacity in Example 1 of this invention.

[0017] Figure 3This is a data diagram of dV5 for Example 1 in this invention.

[0018] Figure 4 This is a data graph of d (V5-V7) in Example 1 of this invention.

[0019] Figure 5 This is the tactile impedance curve of Example 1 in this invention.

[0020] Figure 6 This is a data diagram of the capacity in Example 2 of the present invention. Figure 7 This is a data diagram of dV5 for Example 2 in this invention.

[0021] Figure 8 This is a data graph of d (V5-V7) in Example 2 of this invention.

[0022] Figure 9 This is the impedance curve of the coining resistor in Example 2 of this invention.

[0023] Figure 10 This is a data graph of the volume in control group 1 in this invention.

[0024] Figure 11 This is a graph showing the dV5 data in control group 1 of this invention.

[0025] Figure 12 This is the impedance curve of the control group 1 ohm in this invention.

[0026] Figure 13 This is a data graph of the volume in control group 2 in this invention.

[0027] Figure 14 This is a graph showing the dV5 data of control group 2 in this invention.

[0028] Figure 15 This is the tactile impedance curve of control group 2 in this invention. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] In this invention, the charge / discharge capacity, energy, and end voltage data of each cell cycle are acquired. The rebound voltage cycle trend after discharge is also added. However, simply adding the rebound voltage trend is insufficient to accurately determine the cell's health status, as it is significantly affected by ambient temperature changes, circuit aging, and wiring methods. Therefore, this invention adds a small-current discharge process after the initial discharge cycle. This small current is generally controlled below 0.3C to mitigate polarization caused by high-current discharge and reduce the impact of temperature rise, but it should not be lower than 0.05C to save testing time. This small-current discharge is not added after every cycle, but rather, based on the cell's lifespan requirements, it is generally added every 10 to 100 cycles. The monitoring frequency typically accounts for 5% to 10% of the total number of cycles. High frequency of small current discharge increases the cell's depth of discharge, accelerating cell degradation and hindering lifespan assessment. Conversely, high frequency may miss the cell's degradation inflection point, making it difficult to predict cell health status. Therefore, the following steps are designed based on the above: like Figure 1 As shown, the first step is to let the battery cell rest for 5 minutes to 1 hour. In actual testing, the time may fluctuate depending on the size of the battery cell. The second step is to charge the battery cell to the upper limit of the voltage design using the rated current constant current charging method, and then charge it to a current of 0.02 to 0.05C using the upper limit voltage constant voltage charging method. The upper limit of the battery cell voltage is determined by the potential difference between the positive and negative electrode materials of the battery cell. The third step is to let the battery cell rest for another 5 minutes to 1 hour. The time may fluctuate depending on the size of the battery cell. The fourth step is to discharge the battery cell to the lower limit of the voltage design using the rated current constant current discharging method. The fifth step is to let the battery cell rest for 5 minutes to 1 hour. The time may fluctuate depending on the size of the battery cell. Then, steps 2 to 5 are repeated until the design lifespan ends. This is the cyclic testing method in the prior art. In this application, the additional difference is that the voltage change trend after the rebound in step 5 is statistically analyzed, and the voltage change rate is calculated and denoted as dV5. Points where the voltage change rate is continuously greater than 0 are recorded as condition one. When performing 10 to 100 cycles, a step is added after step 5, including constant current discharge at a current of 0.05C to 0.3C to the lower limit of the voltage design and resting for 5 minutes to 1 hour. Then, the change trend of the S5 end voltage minus the S7 end voltage value in the same cycle is calculated, and the change rate of the voltage difference is statistically analyzed and denoted as d(V5-V7). The number of consecutive decreases of d(V5-V7) is recorded as condition two.

[0031] If a battery cell meets the following conditions: Condition 1: dV5 is greater than 0 for more than 10 consecutive points, and Condition 2: d(V5-V7) decreases by more than or equal to 3 consecutive points, it can be determined that the battery cell will experience a voltage drop. The cell can be removed early to analyze the cause of failure, thus effectively saving evaluation time and testing resources. This is because while a continuously positive dV5 indicates an increase in internal resistance and a possible abnormal side reaction, the reasons for a continuous voltage rise are not limited to this. Other causes include ambient temperature fluctuations, poor contact between the battery cell and the power harness, or aging power harnesses. In response to interference issues, this design includes a second condition: if the increase in dV5 is not caused by the cell itself, d(V5-V7) will remain stable or increase because they will rise or fall synchronously. After V5-V7, external interference is eliminated. Since the second discharge rebound voltage is more sensitive to changes in the cell's internal resistance, if d(V5-V7) also shows a continuous decrease, it indicates that the cell itself has a problem. Therefore, the added steps in this application effectively eliminate interference caused by the environment, temperature, and workflow, and can more accurately predict the cell's voltage drop in advance, saving evaluation time and testing resources.

[0032] The following specific examples illustrate the workflow and testing results of this application.

[0033] Example 1: 100Ah lithium iron phosphate aluminum-cased cell (using the new testing procedure) Mass-produced 100Ah lithium iron phosphate aluminum-cased cells were selected as test sample cells. Thirty fresh finished cells were selected for cycle testing at one time. The design cycle life of the 100Ah lithium iron phosphate aluminum-cased cells is 5000 cycles, and the capacity retention rate is ≥80%.

[0034] Specific process flow: Connect the experimental battery cells to the test channel, check the connection status, and after confirming there are no abnormalities, issue the unified procedure. Maintain the ambient temperature at 25±3℃. The detailed procedure is as follows: S1: Let the battery cell rest for 60 minutes; S2. Charge at a constant current of 100A to its design voltage limit of 3.65V, then charge at a constant voltage of 3.65V to a current of 0.02C to end; S3. Let the battery cell rest for 15 minutes; S4. The lower limit of the voltage design for constant current discharge to the cell at a rated current of 100A is 2.5V; S5. Let the battery cell rest for 15 minutes; S6. Repeat S2 to S5 50 times; S7. Discharge at a constant current of 20A to the design lower voltage limit of 2.5V; S8. Let the battery cell rest for 15 minutes; S9. Repeat S2 to S8 100 times.

[0035] The data after the experiment were summarized as follows: Figure 2-5 Based on the capacity, the decay rate showed no abnormalities. However, the rebound voltage development trends after two discharge cycles showed significant differences. After 300-400 cycles, several types of cells with significant differences in rebound voltage were disassembled to check the cell health status. Electrode samples were taken to test the coin resistance and specific capacity. It was found that the coin resistance of cells with continuously increasing rebound voltage increased significantly. Based on the data from the example cycle data, cells with abnormal interfaces can be accurately screened according to the conditions, thus meeting the requirements of cycle testing.

[0036] Example 2: 3Ah nickel-cobalt-manganese lithium pouch cell (new testing procedure) Mass-produced 3Ah lithium nickel cobalt manganese oxide pouch cells were selected as experimental sample cells. Thirty fresh finished cells were selected for cycle testing at one time. The 3Ah lithium iron phosphate aluminum shell cell is designed to have a cycle life of 500 cycles and a capacity retention rate of ≥80%.

[0037] Specific process flow: Connect the experimental battery cell to the test channel, check the connection status, and after confirming there are no abnormalities, issue the unified procedure. Maintain the ambient temperature at 25±3℃. The detailed procedure is as follows: S1. Let the battery cells rest for 30 minutes; S2. Charge at a constant current of 1.5A to its design voltage limit of 4.4V, then charge at a constant voltage of 4.4V to a current of 0.02C to end; S3. Let the battery cell rest for 10 minutes; S4. The design lower limit of the voltage of the cell when discharged at a constant current of 1.5A is 3.0V; S5. Let the battery cell rest for 10 minutes; S6. Repeat S2 to S5 10 times; The S7.0.9A constant current discharge is maintained to the design lower voltage limit of 3V; S8. Let the battery cell rest for 10 minutes; S9. Repeat S2 to S8 50 times.

[0038] The data after the experiment were summarized as follows: Figure 6-9Analysis revealed that while the capacity decay rate appeared normal, the rebound voltage trends after two discharge cycles showed significant differences. After more than 100 cycles, several cells with significant differences in rebound voltage were disassembled to examine their health. Electrode samples were taken to test cladding resistance and specific capacity. It was found that cells with continuously increasing rebound voltage had significantly increased cladding resistance and significantly reduced specific capacity. Therefore, based on the conditions, cells with abnormal interfaces can be accurately screened out. Such batteries will inevitably fail in the later stages of cycling.

[0039] Control group 1: 100Ah lithium iron phosphate aluminum-cased cells (using the old testing procedure) Mass-produced 100Ah lithium iron phosphate aluminum-cased cells were selected as experimental sample cells. Thirty fresh finished cells were selected for cycle testing at one time. The design cycle life of the 100Ah lithium iron phosphate aluminum-cased cells is 5000 cycles, and the capacity retention rate is ≥80%.

[0040] Specific process flow: Connect the experimental battery cells to the test channel, check the connection status, and after confirming that there are no abnormalities, issue the unified procedure. Maintain the ambient temperature at 25±3℃. The detailed procedure is as follows: S1. Let the battery cell rest for 60 minutes; S2. Charge at a constant current of 100A to its design voltage limit of 3.65V, then charge at a constant voltage of 3.65V to a current of 0.02C to end; S3. Let the battery cell rest for 15 minutes; S4. The lower limit of the voltage design for constant current discharge to the cell at a rated current of 100A is 2.5V; S5. Let the battery cell rest for 15 minutes; S6. Repeat S2 to S8 5000 times; The data after the experiment were summarized as follows: Figure 10-12 Analysis revealed that the process data could not be used to determine the quality of the battery cells, based on disassembly of the cells and testing of the charge-to-discharge data.

[0041] Control group 2: 3Ah nickel-cobalt-manganese lithium pouch cell (old testing procedure) Mass-produced 3Ah lithium nickel cobalt manganese oxide pouch cells were selected as experimental sample cells. Thirty fresh finished cells were selected for cycle testing at one time. The 3Ah lithium iron phosphate aluminum shell cell is designed to have a cycle life of 500 cycles and a capacity retention rate of ≥80%.

[0042] Specific process flow: Connect the experimental battery cell to the test channel, check the connection status, and after confirming there are no abnormalities, issue the unified procedure. Maintain the ambient temperature at 25±3℃. The detailed procedure is as follows: S1. Let the battery cell rest for 30 minutes; S2. Charge at a constant current of 1.5A to its design voltage limit of 4.4V, then charge at a constant voltage of 4.4V to a current of 0.02C to end; S3. Let the battery cell rest for 10 minutes; S4. Charge at a constant current of 1.5A to its design voltage limit of 4.4V, then charge at a constant voltage of 4.4V to a current of 0.02C to end; S5. Let the battery cell rest for 10 minutes; S6. Repeat S2 to S8 500 times; The data after the experiment were summarized as follows: Figure 13-15 Furthermore, the battery cells with significant differences in capacity decay and voltage rebound trends were disassembled, and the electrode sheets were taken to test the coin resistance and specific capacity, verifying that the data under this test method is not strongly correlated with the actual health status of the battery cells.

Claims

1. A method for predicting battery cell cycle degradation in advance, comprising the following steps: S1: Time out the battery cells; S2: The battery cell is charged to the upper limit of the voltage design using the rated current constant current charging method, and then charged to the upper limit voltage constant voltage to the current of 0.02~0.05C to end; S3: Time out the battery cells; S4: Discharge the battery cell to the lower limit of the battery cell's voltage design using a constant current discharge method with the rated current. S5: Time the battery cell to rest, and statistically analyze the voltage change trend after resting, and calculate the voltage change rate, denoted as dV5. Count the number of points where the voltage change rate is continuously greater than 0, and denote it as condition one. S6: Repeat S2 to S5 until the final number of cycles required for the cell life design, characterized in that, every 10 to 100 cycles of S2 to S5, the following steps S7 to S8 are added: S7: Discharge at a constant current of 0.05C to 0.3C to the lower limit of the voltage design; S8: Set aside the battery cell, statistically analyze the trend of the rebound voltage after setting aside, and calculate the trend of the change of the voltage at the end of S5 minus the voltage at the end of S7 in the same cycle. Statistically analyze the rate of change of the voltage difference, denoted as d(V5-V7). The number of consecutive decreases of d(V5-V7) is denoted as condition two. S9: Count the number of conditions 1 and 2, and filter the portions of condition 1 greater than a and condition 2 greater than b as abnormal states, where a and b are positive integers.

2. The method for predicting battery cell cycle drops in advance according to claim 1, characterized in that, The value of a is 10, and the value of b is 3.

3. The method for predicting battery cell cycle drops in advance according to claim 2, characterized in that, The S1 resting time is set to 5 minutes to 1 hour.

4. The method for predicting battery cell cycle drops in advance according to claim 2, characterized in that, The S3 resting time is set to 5 minutes to 1 hour.

5. The method for predicting battery cell cycle drops in advance according to claim 2, characterized in that, The S5 resting time is set to 5 minutes to 1 hour.

6. The method for predicting battery cell cycle drops in advance according to claim 2, characterized in that, The S7 resting time is set to 5 minutes to 1 hour.

7. The method for predicting battery cell cycle drops in advance according to claim 3, characterized in that, The temperature of the test environment was maintained at 25±3℃.

Images