A power battery SOH prediction method based on OCV-SOC calibration

Abstract

The present invention discloses a power battery SOH prediction method based on OCV-SOC calibration, which belongs to the field of battery health monitoring technology. The method first constructs the SOC and temperature gradient vectors under different charging modes; secondly, calculates the gap between these vectors and the comparison vector; then evaluates the charging instability and charging time deviation; further calculates the slow charging and fast charging compensation coefficients; and finally uses the SOH prediction model, combined with the compensation coefficient, to accurately estimate the health state of the power battery. The method effectively overcomes the technical problem of low accuracy of traditional battery health state prediction through multi-dimensional data fusion and compensation mechanism.

Description

Technical Field

[0001] The present invention relates to the technical field of battery health monitoring, and in particular to a power battery SOH prediction method based on OCV-SOC calibration. Background Art

[0002] In the field of electric vehicles and new energy storage, the state of health (SOH) assessment of power batteries is a key technology to ensure the safety, reliability and performance of battery systems. Traditional SOH assessment methods mainly rely on a single parameter or a simple electrochemical model, which is difficult to fully and accurately reflect the actual degradation of the battery. Existing technologies usually use static charge and discharge tests or capacity decay analysis under a single operating condition, which cannot effectively capture and quantify the microscopic change characteristics of battery performance under different charging modes (slow charging and fast charging), resulting in a large deviation between the SOH prediction results and the actual battery health status. Summary of the invention

[0003] In view of the above-mentioned deficiencies in the prior art, the present invention provides a power battery SOH prediction method based on OCV-SOC calibration to solve the problem of low accuracy in health status prediction of power batteries in the prior art.

[0004] In order to achieve the above-mentioned invention object, the technical solution adopted by the present invention is: a power battery SOH prediction method based on OCV-SOC calibration, comprising the following steps:

[0005] S1. In the slow charging mode and the fast charging mode, the power battery to be tested is charged multiple times to construct a slow charging SOC vector, a fast charging SOC vector, a slow charging temperature gradient vector and a fast charging temperature gradient vector;

[0006] S2, subtracting the slow charging SOC vector, the fast charging SOC vector, the slow charging temperature gradient vector, and the fast charging temperature gradient vector from the comparison vector to obtain a slow charging SOC gap vector, a fast charging SOC gap vector, a slow charging temperature gradient gap vector, and a fast charging temperature gradient gap vector;

[0007] S3, calculating the slow charging instability and the fast charging instability for the slow charging SOC gap vector and the fast charging SOC gap vector respectively;

[0008] S4. Record the time it takes for the power battery to be tested to be fully charged in the slow charging mode and the fast charging mode, respectively, to obtain the slow charging time and the fast charging time, and calculate the slow charging time deviation and the fast charging time deviation;

[0009] S5. Add the slow charging instability and the slow charging time deviation to obtain a slow charging compensation coefficient, and add the fast charging instability and the fast charging time deviation to obtain a fast charging compensation coefficient;

[0010] S6. Use the power battery SOH prediction model to process the slow charging SOC gap vector, the fast charging SOC gap vector, the slow charging temperature gradient gap vector and the fast charging temperature gradient gap vector, and obtain the power battery health status value based on the compensation of the slow charging compensation coefficient and the fast charging compensation coefficient.

[0011] Furthermore, S1 comprises the following sub-steps:

[0012] S11. In the slow charging mode, the power battery to be tested is charged multiple times. After each charge, the power battery to be tested is left to stand until the battery voltage is stable, and then an open-circuit SOC calibration value is collected to construct a slow charging SOC vector. In the slow charging mode, each charging time is the same.

[0013] S12. In the slow charging mode, collect the temperature value of each charging process of the power battery to be tested, calculate the temperature gradient value, and construct a slow charging temperature gradient vector;

[0014] S13. In the fast charging mode, the power battery to be tested is charged multiple times. After each charge, the power battery to be tested is left to stand until the battery voltage is stable, and then an open-circuit SOC calibration value is collected to construct a fast charging SOC vector. In the fast charging mode, each charging time is the same.

[0015] S14. In the fast charging mode, the temperature value of each charging process of the power battery to be tested is collected, the temperature gradient value is calculated, and the fast charging temperature gradient vector is constructed.

[0016] Furthermore, the formulas for the temperature gradient values ​​in S12 and S14 are both: , where r is the temperature gradient value, T max is the maximum temperature during a charging process, T min is the minimum temperature value during a charging process, and △t is the charging time.

[0017] Further, S2 includes the following sub-steps:

[0018] S21, subtracting the slow charging SOC vector from the slow charging SOC comparison vector to obtain a slow charging SOC gap vector, wherein the slow charging SOC comparison vector is a slow charging SOC vector obtained by a healthy power battery;

[0019] S22, subtracting the fast charge SOC vector from the fast charge SOC comparison vector to obtain a fast charge SOC gap vector, wherein the fast charge SOC comparison vector is a fast charge SOC vector obtained by a healthy power battery;

[0020] S23, subtracting the slow charging temperature gradient vector from the slow charging temperature gradient comparison vector to obtain a slow charging temperature gradient difference vector, wherein the slow charging temperature gradient comparison vector is a slow charging temperature gradient vector obtained by a healthy power battery;

[0021] S24. Subtract the fast charging temperature gradient vector from the fast charging temperature gradient comparison vector to obtain a fast charging temperature gradient difference vector, wherein the fast charging temperature gradient comparison vector is a fast charging temperature gradient vector obtained for a healthy power battery.

[0022] Furthermore, the formula for calculating the slow charging instability in S3 is: , where θ s is the slow charging instability, e is a natural constant, x s,min is the minimum value in the slow charging SOC gap vector, x s,max is the maximum value of the slow-charging SOC gap vector;

[0023] The formula for calculating fast charging instability in S3 is: , where θ F For fast charging instability, x F,min is the minimum value in the fast charge SOC gap vector, x F,max It is the maximum value in the fast charging SOC gap vector.

[0024] Furthermore, the formula for calculating the slow charging time deviation in S4 is: , where μ s is the slow charging time deviation, e is a natural constant, | | is an absolute value operation, T s T is the time it takes to fully charge the power battery under test in slow charging mode. s,E Time to fully charge a healthy battery in slow charge mode;

[0025] The formula for calculating the fast charge time deviation in S4 is: , where μ F is the fast charge time deviation, T F is the time it takes to fully charge the power battery under test in fast charging mode, T F,E Time to fully charge a healthy battery in fast charge mode.

[0026] Further, the power battery SOH prediction model in S6 includes: a first fully connected layer, a second fully connected layer, a third fully connected layer, a fourth fully connected layer, a slow charging abnormal value calculation unit, a fast charging abnormal value calculation unit, a first compensation unit, a second compensation unit and an adder;

[0027] The input end of the first fully connected layer is used to input the slow charge SOC gap vector; the input end of the second fully connected layer is used to input the slow charge temperature gradient gap vector; the input end of the third fully connected layer is used to input the fast charge SOC gap vector; the input end of the fourth fully connected layer is used to input the fast charge temperature gradient gap vector;

[0028] The input end of the slow charging abnormal value calculation unit is connected to the output end of the first fully connected layer and the output end of the second fully connected layer respectively; the input end of the fast charging abnormal value calculation unit is connected to the output end of the third fully connected layer and the output end of the fourth fully connected layer respectively; the first input end of the first compensation unit is connected to the output end of the slow charging abnormal value calculation unit, and its second input end is used to input the slow charging compensation coefficient; the first input end of the second compensation unit is connected to the output end of the fast charging abnormal value calculation unit, and its second input end is used to input the fast charging compensation coefficient; the input end of the adder is connected to the output end of the first compensation unit and the output end of the second compensation unit respectively, and its output end serves as the output end of the power battery SOH prediction model.

[0029] Furthermore, the expression of the slow charging abnormal value calculation unit is: , where y s is the slow charging outlier, X1 is the output of the first fully connected layer, X2 is the output of the second fully connected layer, W1 is the weight of X1, and W2 is the weight of X2;

[0030] The expression of the fast charging abnormal value calculation unit is: , where y F is the fast charging outlier, X3 is the output of the third fully connected layer, X4 is the output of the fourth fully connected layer, W3 is the weight of X3, and W4 is the weight of X4.

[0031] Furthermore, the expression of the first compensation unit is: , where y s,c is the output of the first compensation unit, y s is the slow charging outlier, γ s is the slow charging compensation coefficient;

[0032] The expression of the second compensation unit is: , where y F,c is the output of the second compensation unit, y F is the fast charging outlier, γ F is the fast charge compensation coefficient.

[0033] The beneficial effects of the present invention are:

[0034] 1. The present invention considers the SOC vector and temperature gradient vector under the two charging modes of slow charging and fast charging, and subtracts the SOC vector and temperature gradient vector from the comparison vector to obtain the difference in SOC and temperature gradient between the power battery to be tested and the healthy battery, that is, the SOC gap vector and temperature gradient gap vector under the two charging modes of slow charging and fast charging, so as to fully capture the microscopic change characteristics of battery performance and significantly improve the accuracy of health status prediction. Compared with the traditional single working condition test method, this method can more comprehensively reflect the actual degradation of the battery.

[0035] 2. The present invention calculates the slow charging instability and the fast charging instability for the slow charging SOC gap vector and the fast charging SOC gap vector, reflects the electrochemical performance degradation characteristics of the power battery to be tested under different charging modes, and quantifies the instability of the electrochemical reaction during the charging process.

[0036] 3. The present invention counts the time it takes for the power battery to be tested to be fully charged, and calculates the slow charging time deviation and the fast charging time deviation. The slow charging time deviation reflects the performance stability under low current charging conditions, and the fast charging time deviation reflects the performance stability under high current charging conditions, and quantifies the attenuation of the power battery to be tested.

[0037] 4. Add the slow charging instability and the slow charging time deviation to get the slow charging compensation coefficient, which reflects the deterioration of slow charging in the slow charging mode. Add the fast charging instability and the fast charging time deviation to get the fast charging compensation coefficient, which reflects the deterioration of fast charging in the fast charging mode.

[0038] 5. The present invention processes the SOC gap vector and temperature gradient gap vector under slow charging and fast charging respectively through the power battery SOH prediction model, and then obtains the power battery health state value based on the compensation of the slow charging compensation coefficient and the fast charging compensation coefficient, thereby improving the health state prediction accuracy of the power battery. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a flow chart of a power battery SOH prediction method based on OCV-SOC calibration;

[0040] Figure 2 Schematic diagram of the structure of the power battery SOH prediction model. DETAILED DESCRIPTION

[0041] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0042] like Figure 1 As shown, a power battery SOH prediction method based on OCV-SOC calibration includes the following steps:

[0043] S1. In the slow charging mode and the fast charging mode, the power battery to be tested is charged multiple times to construct a slow charging SOC vector, a fast charging SOC vector, a slow charging temperature gradient vector and a fast charging temperature gradient vector;

[0044] S2, subtracting the slow charging SOC vector, the fast charging SOC vector, the slow charging temperature gradient vector, and the fast charging temperature gradient vector from the comparison vector to obtain a slow charging SOC gap vector, a fast charging SOC gap vector, a slow charging temperature gradient gap vector, and a fast charging temperature gradient gap vector;

[0045] S3, calculating the slow charging instability and the fast charging instability for the slow charging SOC gap vector and the fast charging SOC gap vector respectively;

[0046] S4. Record the time it takes for the power battery to be tested to be fully charged in the slow charging mode and the fast charging mode, respectively, to obtain the slow charging time and the fast charging time, and calculate the slow charging time deviation and the fast charging time deviation;

[0047] S5. Add the slow charging instability and the slow charging time deviation to obtain a slow charging compensation coefficient, and add the fast charging instability and the fast charging time deviation to obtain a fast charging compensation coefficient;

[0048] S6. Use the power battery SOH prediction model to process the slow charging SOC gap vector, the fast charging SOC gap vector, the slow charging temperature gradient gap vector and the fast charging temperature gradient gap vector, and obtain the power battery health status value based on the compensation of the slow charging compensation coefficient and the fast charging compensation coefficient.

[0049] OCV (Open Circuit Voltage) is the open circuit voltage, which is the voltage of the battery when there is no external load. SOC (State of Charge) is the battery state of charge.

[0050] In this embodiment, S1 includes the following sub-steps:

[0051] S11. In the slow charging mode, the power battery to be tested is charged multiple times. After each charge, the power battery to be tested is left to stand until the battery voltage is stable, and then the open circuit SOC calibration value is collected to construct a slow charging SOC vector. In the slow charging mode, each charging time is the same, and the elements of the slow charging SOC vector are multiple SOC calibration values.

[0052] S12. In the slow charging mode, collect the temperature value of each charging process of the power battery to be tested, calculate the temperature gradient value, and construct a slow charging temperature gradient vector, where the elements of the slow charging temperature gradient vector are multiple temperature gradient values;

[0053] S13. In the fast charging mode, the power battery to be tested is charged multiple times. After each charge, the power battery to be tested is left to stand until the battery voltage is stable, and then the open circuit SOC calibration value is collected to construct a fast charging SOC vector. In the fast charging mode, the charging time is the same for each time, and the elements of the fast charging SOC vector are multiple SOC calibration values.

[0054] S14. In the fast charging mode, the temperature value of each charging process of the power battery to be tested is collected, the temperature gradient value is calculated, and a fast charging temperature gradient vector is constructed. The elements of the fast charging temperature gradient vector are multiple temperature gradient values.

[0055] The open circuit SOC value after standing still is designated as an open circuit SOC calibration value.

[0056] Taking the electric vehicle ternary lithium battery (capacity: 50Ah) as an example, the initial SOC is 20%, slow charging mode: charging current: 0.2C (10A), charging times: 5 times, each charging time: fixed at 2 hours, the battery standing time is 0.5 hours; fast charging mode: charging current: 1C (50A), charging times: 5 times, each charging time: fixed at 0.5 hours, the battery standing time is 0.5 hours.

[0057] The present invention allows the power battery to be tested to stand still after each charging so that the battery reaches an electrochemical equilibrium state. In the equilibrium state, the SOC value corresponding to the open circuit voltage of the battery can be measured more accurately.

[0058] When the open circuit voltage fluctuation amplitude (or battery voltage change rate) of the power battery to be tested is less than a threshold value, the state of the power battery to be tested is a stable state. The threshold value is a threshold set for the voltage fluctuation amplitude or the battery voltage change rate.

[0059] In this embodiment, the formulas for the temperature gradient values ​​in S12 and S14 are both: , where r is the temperature gradient value, T max is the maximum temperature during a charging process, T min is the minimum temperature value during a charging process, and △t is the charging time.

[0060] The temperature gradient value reflects the internal resistance of the battery. Over a period of time, the larger the temperature gradient value, the larger the internal resistance of the battery and the worse the battery health status.

[0061] In this embodiment, S2 includes the following sub-steps:

[0062] S21, subtracting the slow charging SOC vector from the slow charging SOC comparison vector to obtain a slow charging SOC gap vector, wherein the slow charging SOC comparison vector is a slow charging SOC vector obtained by a healthy power battery;

[0063] S22, subtracting the fast charge SOC vector from the fast charge SOC comparison vector to obtain a fast charge SOC gap vector, wherein the fast charge SOC comparison vector is a fast charge SOC vector obtained by a healthy power battery;

[0064] S23, subtracting the slow charging temperature gradient vector from the slow charging temperature gradient comparison vector to obtain a slow charging temperature gradient difference vector, wherein the slow charging temperature gradient comparison vector is a slow charging temperature gradient vector obtained by a healthy power battery;

[0065] S24. Subtract the fast charging temperature gradient vector from the fast charging temperature gradient comparison vector to obtain a fast charging temperature gradient difference vector, wherein the fast charging temperature gradient comparison vector is a fast charging temperature gradient vector obtained for a healthy power battery.

[0066] The healthy power battery should be of the same model as the power battery to be tested and should not have been used since leaving the factory.

[0067] In this embodiment, the formula for calculating the slow charging instability in S3 is: , where θ s is the slow charging instability, e is a natural constant, x s,min is the minimum value in the slow charging SOC gap vector, x s,max It is the maximum value in the slow charging SOC gap vector.

[0068] In this embodiment, the formula for calculating the fast charge instability in S3 is: , where θ F For fast charging instability, x F,min is the minimum value in the fast charge SOC gap vector, x F,max It is the maximum value in the fast charging SOC gap vector.

[0069] The present invention extracts the maximum and minimum values ​​in the SOC gap vector under slow charging and fast charging modes, calculates the gap between the maximum and minimum values, and uses an exponential function to amplify the gap to make the instability more significant. By calculating the instability of slow charging and fast charging, the stability of the battery under different charging modes can be effectively evaluated, helping to identify potential charging problems.

[0070] In this embodiment, the formula for calculating the slow charging time deviation in S4 is: , where μ s is the slow charging time deviation, e is a natural constant, | | is an absolute value operation, T s T is the time it takes to fully charge the power battery under test in slow charging mode. s,E Time to fully charge a healthy battery in slow charge mode.

[0071] In this embodiment, the formula for calculating the fast charge time deviation in S4 is: , where μ F is the fast charge time deviation, T F is the time it takes to fully charge the power battery under test in fast charging mode, T F,E Time to fully charge a healthy battery in fast charge mode.

[0072] The present invention can evaluate the charging capacity of the power battery to be tested and the healthy battery when they are fully charged according to the difference in time between the power battery to be tested and the healthy battery in the slow charging and fast charging modes. For the battery, as the years of use increase, the battery capacity decreases and the charging time becomes shorter. Therefore, the charging time of an unhealthy battery is shorter.

[0073] like Figure 2 As shown, the power battery SOH prediction model in S6 includes: a first fully connected layer, a second fully connected layer, a third fully connected layer, a fourth fully connected layer, a slow charging abnormal value calculation unit, a fast charging abnormal value calculation unit, a first compensation unit, a second compensation unit and an adder;

[0074] The input end of the first fully connected layer is used to input the slow charge SOC gap vector; the input end of the second fully connected layer is used to input the slow charge temperature gradient gap vector; the input end of the third fully connected layer is used to input the fast charge SOC gap vector; the input end of the fourth fully connected layer is used to input the fast charge temperature gradient gap vector;

[0075] The input end of the slow charging abnormal value calculation unit is connected to the output end of the first fully connected layer and the output end of the second fully connected layer respectively; the input end of the fast charging abnormal value calculation unit is connected to the output end of the third fully connected layer and the output end of the fourth fully connected layer respectively; the first input end of the first compensation unit is connected to the output end of the slow charging abnormal value calculation unit, and its second input end is used to input the slow charging compensation coefficient; the first input end of the second compensation unit is connected to the output end of the fast charging abnormal value calculation unit, and its second input end is used to input the fast charging compensation coefficient; the input end of the adder is connected to the output end of the first compensation unit and the output end of the second compensation unit respectively, and its output end serves as the output end of the power battery SOH prediction model.

[0076] The present invention calculates abnormal values ​​in slow charging and fast charging modes through the SOC gap vector and temperature gradient gap vector in slow charging and fast charging modes, and then compensates them through the slow charging compensation coefficient and the fast charging supplement coefficient respectively, so as to further improve the prediction accuracy of the health status value of the power battery.

[0077] In this embodiment, the expression of the fully connected layer is: , where X is the output of the fully connected layer, x i is the i-th element in the vector, ω i For x i The weight of b i For x i The bias of , i is a positive integer, and N is the number of elements in the vector.

[0078] In this embodiment, the expression of the slow charging abnormal value calculation unit is: , where y sis the slow charging outlier, X1 is the output of the first fully connected layer, X2 is the output of the second fully connected layer, W1 is the weight of X1, and W2 is the weight of X2;

[0079] The expression of the fast charging abnormal value calculation unit is: , where y F is the fast charging outlier, X3 is the output of the third fully connected layer, X4 is the output of the fourth fully connected layer, W3 is the weight of X3, and W4 is the weight of X4.

[0080] In this embodiment, the expression of the first compensation unit is: , where y s,c is the output of the first compensation unit, y s is the slow charging outlier, γ s is the slow charging compensation coefficient;

[0081] The expression of the second compensation unit is: , where y F,c is the output of the second compensation unit, y F is the fast charging outlier, γ F is the fast charge compensation coefficient.

[0082] In this embodiment, the weights and biases in the power battery SOH prediction model can be obtained by training using a gradient descent method.

[0083] The range of the slow charge compensation coefficient and the fast charge compensation coefficient is between 0 and 1. The larger the slow charge compensation coefficient and the fast charge compensation coefficient are, the worse the battery health is.

[0084] In the present invention, the larger the power battery health status value is, the worse the battery health condition is.

[0085] The present invention considers the SOC vector and temperature gradient vector in the two charging modes of slow charging and fast charging, and subtracts the SOC vector and temperature gradient vector from the comparison vector to obtain the difference in SOC and temperature gradient between the power battery to be tested and the healthy battery, that is, the SOC gap vector and temperature gradient gap vector in the two charging modes of slow charging and fast charging, so as to fully capture the microscopic change characteristics of battery performance and significantly improve the accuracy of health status prediction. Compared with the traditional single working condition test method, this method can more comprehensively reflect the actual degradation of the battery.

[0086] The present invention calculates the slow charging instability and the fast charging instability for the slow charging SOC gap vector and the fast charging SOC gap vector, reflects the electrochemical performance degradation characteristics of the power battery to be tested under different charging modes, and quantifies the instability of the electrochemical reaction during the charging process.

[0087] The present invention counts the time it takes for a power battery to be fully charged, calculates a slow charging time deviation and a fast charging time deviation, wherein the slow charging time deviation reflects the performance stability under low current charging conditions, and the fast charging time deviation reflects the performance stability under high current charging conditions, and quantifies the attenuation of the power battery to be tested.

[0088] The slow charging instability and the slow charging time deviation are added together to obtain the slow charging compensation coefficient, which reflects the deterioration of slow charging in the slow charging mode. The fast charging instability and the fast charging time deviation are added together to obtain the fast charging compensation coefficient, which reflects the deterioration of fast charging in the fast charging mode.

[0089] The present invention processes the SOC gap vector and the temperature gradient gap vector under slow charging and fast charging respectively through the power battery SOH prediction model, and then obtains the power battery health state value based on the compensation of the slow charging compensation coefficient and the fast charging compensation coefficient, thereby improving the health state prediction accuracy of the power battery.

[0090] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A power battery SOH prediction method based on OCV-SOC calibration, characterized in that: The following steps are involved: S1. In the slow charging mode and the fast charging mode, the power battery to be tested is charged multiple times to construct a slow charging SOC vector, a fast charging SOC vector, a slow charging temperature gradient vector and a fast charging temperature gradient vector; S2, subtracting the slow charging SOC vector, the fast charging SOC vector, the slow charging temperature gradient vector, and the fast charging temperature gradient vector from the comparison vector to obtain a slow charging SOC gap vector, a fast charging SOC gap vector, a slow charging temperature gradient gap vector, and a fast charging temperature gradient gap vector; S3, calculating the slow charging instability and the fast charging instability for the slow charging SOC gap vector and the fast charging SOC gap vector respectively; S4. Record the time it takes for the power battery to be tested to be fully charged in the slow charging mode and the fast charging mode, respectively, to obtain the slow charging time and the fast charging time, and calculate the slow charging time deviation and the fast charging time deviation; S5. Add the slow charging instability and the slow charging time deviation to obtain a slow charging compensation coefficient, and add the fast charging instability and the fast charging time deviation to obtain a fast charging compensation coefficient; S6. Using the power battery SOH prediction model to process the slow charge SOC gap vector, the fast charge SOC gap vector, the slow charge temperature gradient gap vector, and the fast charge temperature gradient gap vector, and based on the compensation of the slow charge compensation coefficient and the fast charge compensation coefficient, obtain the power battery health status value; The formula for calculating the slow charging instability in S3 is: , where θ s is the slow charging instability, e is a natural constant, x s,min is the minimum value in the slow charging SOC gap vector, x s,max is the maximum value of the slow-charging SOC gap vector; The formula for calculating the fast charge instability in S3 is: , where θ F For fast charging instability, x F,min is the minimum value in the fast charge SOC gap vector, x F,max is the maximum value of the fast charge SOC gap vector; The formula for calculating the slow charging time deviation in S4 is: , where μ s is the slow charging time deviation, e is a natural constant, | | is an absolute value operation, T s T is the time it takes to fully charge the power battery under test in slow charging mode. s,E Time to fully charge a healthy battery in slow charge mode; The formula for calculating the fast charge time deviation in S4 is: , where μ F is the fast charge time deviation, T F is the time it takes to fully charge the power battery under test in fast charging mode, T F,E Time to fully charge a healthy battery in fast charge mode.

2. The power battery SOH prediction method based on OCV-SOC calibration according to claim 1, characterized in that: The S1 comprises the following sub-steps: S11. In the slow charging mode, the power battery to be tested is charged multiple times. After each charge, the power battery to be tested is left to stand until the battery voltage is stable, and then an open-circuit SOC calibration value is collected to construct a slow charging SOC vector. In the slow charging mode, each charging time is the same. S12. In the slow charging mode, collect the temperature value of each charging process of the power battery to be tested, calculate the temperature gradient value, and construct a slow charging temperature gradient vector; S13. In the fast charging mode, the power battery to be tested is charged multiple times. After each charge, the power battery to be tested is left to stand until the battery voltage is stable, and then an open-circuit SOC calibration value is collected to construct a fast charging SOC vector. In the fast charging mode, each charging time is the same. S14. In the fast charging mode, the temperature value of each charging process of the power battery to be tested is collected, the temperature gradient value is calculated, and the fast charging temperature gradient vector is constructed.

3. The power battery SOH prediction method based on OCV-SOC calibration according to claim 2, characterized in that: The formulas for the temperature gradient values ​​in S12 and S14 are: , where r is the temperature gradient value, T max is the maximum temperature during a charging process, T min is the minimum temperature value during a charging process, and △t is the charging time.

4. The power battery SOH prediction method based on OCV-SOC calibration according to claim 1, characterized in that: The S2 comprises the following sub-steps: S21, subtracting the slow charging SOC vector from the slow charging SOC comparison vector to obtain a slow charging SOC gap vector, wherein the slow charging SOC comparison vector is a slow charging SOC vector obtained by a healthy power battery; S22, subtracting the fast charge SOC vector from the fast charge SOC comparison vector to obtain a fast charge SOC gap vector, wherein the fast charge SOC comparison vector is a fast charge SOC vector obtained by a healthy power battery; S23, subtracting the slow charging temperature gradient vector from the slow charging temperature gradient comparison vector to obtain a slow charging temperature gradient difference vector, wherein the slow charging temperature gradient comparison vector is a slow charging temperature gradient vector obtained by a healthy power battery; S24. Subtract the fast charging temperature gradient vector from the fast charging temperature gradient comparison vector to obtain a fast charging temperature gradient difference vector, wherein the fast charging temperature gradient comparison vector is a fast charging temperature gradient vector obtained for a healthy power battery.

5. The power battery SOH prediction method based on OCV-SOC calibration according to claim 1, characterized in that: The power battery SOH prediction model in S6 includes: a first fully connected layer, a second fully connected layer, a third fully connected layer, a fourth fully connected layer, a slow charging abnormal value calculation unit, a fast charging abnormal value calculation unit, a first compensation unit, a second compensation unit and an adder; The input end of the first fully connected layer is used to input a slow charge SOC gap vector; the input end of the second fully connected layer is used to input a slow charge temperature gradient gap vector; the input end of the third fully connected layer is used to input a fast charge SOC gap vector; the input end of the fourth fully connected layer is used to input a fast charge temperature gradient gap vector; The input end of the slow charging abnormal value calculation unit is respectively connected to the output end of the first fully connected layer and the output end of the second fully connected layer; the input end of the fast charging abnormal value calculation unit is respectively connected to the output end of the third fully connected layer and the output end of the fourth fully connected layer; the first input end of the first compensation unit is connected to the output end of the slow charging abnormal value calculation unit, and its second input end is used to input the slow charging compensation coefficient; the first input end of the second compensation unit is connected to the output end of the fast charging abnormal value calculation unit, and its second input end is used to input the fast charging compensation coefficient; the input end of the adder is respectively connected to the output end of the first compensation unit and the output end of the second compensation unit, and its output end serves as the output end of the power battery SOH prediction model.

6. The power battery SOH prediction method based on OCV-SOC calibration according to claim 5, characterized in that: The expression of the slow charging abnormal value calculation unit is: , where y s is the slow charging outlier, X1 is the output of the first fully connected layer, X2 is the output of the second fully connected layer, W1 is the weight of X1, and W2 is the weight of X2; The expression of the fast charge abnormal value calculation unit is: , where y F is the fast charging outlier, X3 is the output of the third fully connected layer, X4 is the output of the fourth fully connected layer, W3 is the weight of X3, and W4 is the weight of X4.

7. The power battery SOH prediction method based on OCV-SOC calibration according to claim 5, characterized in that: The expression of the first compensation unit is: , where y s,c is the output of the first compensation unit, y s is the slow charging outlier, γ s is the slow charging compensation coefficient; The expression of the second compensation unit is: , where y F,c is the output of the second compensation unit, y F is the fast charging outlier, γ F is the fast charge compensation coefficient.

Images