Battery remaining life prediction method based on historical degradation path

Abstract

The invention provides a historical degradation path-based battery remaining life prediction method, which belongs to the technical field of battery management and prediction maintenance, and comprises the following steps: acquiring historical degradation paths of a battery, and allocating a weight variable and a translation variable for each path, minimizing a mean square error between an online degradation path and an initial prediction path calculated through a weight variable and a translation variable through an optimization method, and determining an optimal weight vector and a translation variable; calculating a final prediction path according to the optimal weight vector and the translation variable; compensating the deviation between the final predicted path and the online degradation path in the latest time period by setting a compensation variable, and applying the compensation variable to the final predicted path to obtain a final predicted degradation path; and according to the final predicted degradation path and a set failure threshold value, calculating to obtain the remaining life of the battery. The method solves the problem of high requirements on mathematical knowledge and calculation cost of a battery remaining life prediction method in the prior art.

Description

Technical Field

[0001] The present invention relates to the technical field of battery management and predictive maintenance, and in particular to a method for predicting remaining battery life based on historical degradation paths. Background Art

[0002] Batteries are an indispensable part of modern energy systems. They can efficiently store and release energy to power a variety of portable devices, electric vehicles, and large-scale energy storage systems. Accurate prediction of remaining battery life can identify battery performance degradation in advance, so that preventive measures can be taken to avoid equipment downtime and production interruptions caused by battery failure. This will enable more efficient planning of battery replacement and maintenance cycles, reduce unnecessary waste, and improve overall operational efficiency.

[0003] However, the current battery remaining life prediction method based on deep learning requires a large amount of data and complex neural network models for training and reasoning, involving high-performance computing resources and long data processing. The battery remaining life prediction method based on statistical processes involves complex data analysis, probability models and mathematical derivation, and requires high mathematical knowledge.

[0004] Therefore, a battery remaining life prediction method that requires low mathematical knowledge and computational cost is necessary. Summary of the invention

[0005] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a battery remaining life prediction method based on historical degradation paths, which realizes the battery remaining life prediction while solving the problems of high mathematical knowledge and computational cost requirements of the current battery remaining life prediction methods, and realizes accurate battery remaining life prediction, which is beneficial to improving the safety and reliability of recyclable batteries.

[0006] To achieve the above object, the present invention provides the following solutions:

[0007] A method for predicting remaining battery life based on a historical degradation path comprises the following steps:

[0008] S1. Obtain the historical degradation path of the battery, assign weight variables and translation variables to each path, minimize the mean square error between the online degradation path and the initial prediction path calculated by the weight variables and translation variables through the optimization method, and determine the optimal weight vector and translation variable;

[0009] S2, calculate the final prediction path based on the optimal weight vector and translation variable;

[0010] S3, compensating for the deviation between the final predicted path and the online degradation path in the latest time period by setting a compensation variable, and applying the compensation variable to the final predicted path to obtain a final predicted degradation path;

[0011] S4. Calculate the remaining life of the battery according to the final predicted degradation path and the set failure threshold.

[0012] Preferably, in S1, the initial prediction path is calculated by weight variables and translation variables, including:

[0013] A weight variable is assigned to each historical degradation path, and the weight vector composed of weight variables is expressed as in, is the weight variable of the rth historical degradation path;

[0014] A translation variable is set for the vertical translation of the historical degradation path. The translation variable is a scalar, represented by x t ;

[0015] Calculate the initial prediction path based on the weight variables and translation variables The formula is:

[0016]

[0017] In the formula, is the weight variable of the i-th historical degradation path, N h is the number of historical degenerate paths, is the i-th historical degradation path at t k The performance degradation variable value at time.

[0018] Preferably, in S1, the mean square error between the online degradation path and the initial prediction path is minimized by an optimization method to determine the optimal weight vector and translation variable, including:

[0019] The BFGS optimizer is used for optimization, and the optimal weight vector and translation variable are obtained by repeated iterative calculation. The formula is:

[0020]

[0021] In the formula, is the optimal weight vector, is the optimal weight variable of the rth historical degradation path; is the optimal translation variable; P o is the online degradation path, where t1=0,t k is the online degradation path P o The time scale of P o The length of the vector.

[0022] Preferably, in S2, the final prediction path is calculated according to the optimal weight vector and translation variable, including:

[0023]

[0024] In the formula, is the final predicted path, is the optimal weight variable of the i-th historical degradation path, is the i-th historical degradation path in The performance degradation variable value at time , For the smallest N h is the number of historical degenerate paths.

[0025] Preferably, in S3, the deviation between the final predicted path and the online degraded path in the latest time period is compensated by setting a compensation variable, including:

[0026] Setting a compensation variable Used to compensate the final predicted path With the online degradation path P o The deviation between the latest time stage, where the online degradation path P o Expressed as Final predicted path From t1 to t k The degradation path of t k represents the online degenerate path P o Length;

[0027] Among them, the compensation variable Obtained by the following formula:

[0028] len=min(20,int(0.2*t k ));

[0029]

[0030] Where len is the time window length of the latest time stage, int(·) is the function of rounding down, is the jth performance degradation variable value of the online degradation path, is the jth performance degradation variable value of the final predicted path.

[0031] Preferably, in S3, applying the compensation variable to the final predicted path to obtain the final predicted degradation path includes:

[0032] The compensation variable will be obtained To compensate the final predicted path The error is obtained to obtain the nearest time stage of the final predicted path:

[0033]

[0034] In the formula, is the final predicted path.

[0035] Preferably, in S4, the remaining life of the battery is calculated according to the final predicted degradation path and the set failure threshold, including:

[0036] The failure threshold is set and the remaining life of the battery is calculated based on the final predicted degradation path:

[0037]

[0038] Where ξ is the failure threshold, The final predicted path In t k The performance degradation variable value at time +l, inf is the lower bound, The remaining battery life.

[0039] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0040] The present invention provides a method for predicting the remaining life of a battery based on a historical degradation path, including: obtaining the historical degradation path of the battery, and assigning weight variables and translation variables to each path, minimizing the mean square error between the online degradation path and the initial prediction path calculated by the weight variables and translation variables through an optimization method, and determining the optimal weight vector and translation variable; calculating the final prediction path according to the optimal weight vector and translation variable; compensating the deviation between the final prediction path and the online degradation path in the latest time period by setting a compensation variable, applying the compensation variable to the final prediction path to obtain the final predicted degradation path; and calculating the remaining life of the battery according to the final predicted degradation path and the set failure threshold. The present invention uses weights, translations, and compensation variables to achieve high-precision remaining life prediction under low mathematical knowledge requirements and low computational costs, while solving the problem of high mathematical knowledge and computational cost requirements of the current battery remaining life prediction method, and achieves accurate battery remaining life prediction, which is beneficial to improving the safety and reliability of recyclable batteries. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0042] Figure 1A flowchart of a method for predicting remaining battery life based on a historical degradation path according to the present invention;

[0043] Figure 2 A comparison chart of the predicted values ​​of the remaining life of a CALCE lithium-ion battery provided in Example 1 of the present invention;

[0044] Figure 3 This is a comparison chart of the absolute prediction errors on the CALCE lithium-ion battery provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0047] Example 1

[0048] like Figure 1 As shown, the present invention provides a method for predicting the remaining life of a battery based on a historical degradation path, which is characterized by comprising the following steps:

[0049] S1. Obtain the historical degradation path of the battery, assign weight variables and translation variables to each path, minimize the mean square error between the online degradation path and the initial prediction path calculated by the weight variables and translation variables through the optimization method, and determine the optimal weight vector and translation variable;

[0050] S2, calculate the final prediction path based on the optimal weight vector and translation variable;

[0051] S3, compensating for the deviation between the final predicted path and the online degradation path in the latest time period by setting a compensation variable, and applying the compensation variable to the final predicted path to obtain a final predicted degradation path;

[0052] S4. Calculate the remaining life of the battery according to the final predicted degradation path and the set failure threshold.

[0053] Specifically, in S1, the initial prediction path is calculated through weight variables and translation variables, including:

[0054] A weight variable is assigned to each historical degradation path, and the weight vector composed of weight variables is expressed as in, is the weight variable of the rth historical degradation path;

[0055] A translation variable is set for the translation of the historical degradation path in the vertical direction. The translation variable is a scalar, represented by x t ;

[0056] Calculate the initial prediction path based on the weight variables and translation variables The formula is:

[0057]

[0058] In the formula, is the weight variable of the i-th historical degradation path, N h is the number of historical degenerate paths, is the i-th historical degradation path at t k The performance degradation variable value at time.

[0059] More specifically, in S1, the mean square error between the online degradation path and the initial prediction path is minimized by an optimization method to determine the optimal weight vector and translation variable, including:

[0060] The BFGS optimizer is used for optimization, and the optimal weight vector and translation variable are obtained by repeated iterative calculation. The formula is:

[0061]

[0062] In the formula, is the optimal weight vector, is the optimal weight variable of the rth historical degradation path; is the optimal translation variable; P o is the online degradation path, where t1=0,t k is the online degradation path P o The time scale of P o The length of the vector.

[0063] In the above process of using BFGS optimizer to optimize and repeatedly iterate to obtain the optimal weight vector and translation variable, the following is done: Set S h is the historical set degenerate path, P o is the online degradation path, ε is the tolerable gradient norm, iter mis the maximum number of iterations, ξ is the failure threshold; the set parameter values ​​are input into the BFGS optimizer, after the parameters are initialized and the objective function is obtained, the error between the predicted path and the online degradation path under the current parameters is measured, and the direction of parameter update is calculated by the BFGS optimizer, and a line search is performed to determine the step size, the parameter value is updated and the gradient of the next step is calculated, and the quasi-Newton matrix is ​​updated according to the gradient for the parameter update of the next iteration, and the above iterative process is repeated to output the optimal weight vector and translation variable. Here, the BFGS optimizer program is used for optimization and repeated iterative calculation to obtain the optimal solution.

[0064] Specifically, in S2, the final prediction path is calculated according to the optimal weight vector and translation variable, including:

[0065]

[0066] In the formula, is the final predicted path, is the optimal weight variable of the i-th historical degradation path, is the i-th historical degradation path in The performance degradation variable value at time , For the smallest N h is the number of historical degenerate paths.

[0067] Specifically, in S3, the deviation between the final predicted path and the online degraded path in the latest time period is compensated by setting compensation variables, including:

[0068] Setting a compensation variable Used to compensate the final predicted path With the online degradation path P o The deviation between the latest time stage, where the online degradation path P o Expressed as Final predicted path From t1 to t k The degradation path of t k represents the online degenerate path P o Length;

[0069] Among them, the compensation variable Obtained by the following formula:

[0070] len=min(20,int(0.2*t k ));

[0071]

[0072] Where len is the time window length of the latest time stage, int(·) is the function of rounding down, is the jth performance degradation variable value of the online degradation path, is the jth performance degradation variable value of the final predicted path.

[0073] More specifically, in S3, applying the compensation variable to the final predicted path to obtain the final predicted degradation path includes:

[0074] The compensation variable will be obtained To compensate the final predicted path The error is obtained to obtain the nearest time stage of the final predicted path:

[0075]

[0076] In the formula, is the final predicted path.

[0077] Specifically, in S4, the remaining life of the battery is calculated according to the final predicted degradation path and the set failure threshold, including:

[0078] The failure threshold is set and the remaining life of the battery is calculated based on the final predicted degradation path:

[0079]

[0080] Where ξ is the failure threshold, The final predicted path In t k The performance degradation variable value at time +l, inf is the lower bound, The remaining battery life.

[0081] In order to further verify the effectiveness of this embodiment, based on the MATLAB tool, the lithium battery data set of the CALCE Center of the University of Maryland was used to predict the remaining life of the lithium battery. The method provided in this embodiment was compared with not considering the translation variable V(W+C), not considering the compensation variable V(W+T), and not considering the translation variable and the compensation variable V(W). Figure 2 and Figure 3 It can be found that the prediction method provided in this embodiment is effective in predicting the remaining life of lithium batteries and has high prediction accuracy.

[0082] Therefore, the above-mentioned battery remaining life prediction method based on historical degradation path is adopted to realize the battery remaining life prediction while solving the problem of high mathematical knowledge and computational cost requirements of the current battery remaining life prediction method, and realizes accurate battery remaining life prediction, which is beneficial to improve the safety and reliability of recyclable batteries.

[0083] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for predicting remaining battery life based on historical degradation paths, characterized in that: The following steps are involved: S1. Obtain the historical degradation path of the battery, assign weight variables and translation variables to each path, minimize the mean square error between the online degradation path and the initial prediction path calculated by the weight variables and translation variables through the optimization method, and determine the optimal weight vector and translation variable; S2, calculate the final prediction path based on the optimal weight vector and translation variable; S3, compensating for the deviation between the final predicted path and the online degradation path in the latest time period by setting a compensation variable, and applying the compensation variable to the final predicted path to obtain a final predicted degradation path; S4. Calculate the remaining life of the battery according to the final predicted degradation path and the set failure threshold.

2. A method for predicting remaining battery life based on historical degradation paths according to claim 1, characterized in that: In S1, the initial prediction path is calculated through weight variables and translation variables, including: A weight variable is assigned to each historical degradation path, and the weight vector composed of weight variables is expressed as in, is the weight variable of the rth historical degradation path; A translation variable is set for the translation of the historical degradation path in the vertical direction. The translation variable is a scalar, represented by x t ; Calculate the initial prediction path based on the weight variables and translation variables The formula is: In the formula, is the weight variable of the i-th historical degradation path, N h is the number of historical degenerate paths, is the i-th historical degradation path at t k The performance degradation variable value at time.

3. The method for predicting remaining battery life based on historical degradation paths according to claim 2, characterized in that: In S1, the mean square error between the online degradation path and the initial prediction path is minimized by an optimization method to determine the optimal weight vector and translation variable, including: The BFGS optimizer is used for optimization, and the optimal weight vector and translation variable are obtained by repeated iterative calculation. The formula is: In the formula, is the optimal weight vector, is the optimal weight variable of the rth historical degradation path; is the optimal translation variable; P o is the online degradation path, where t1=0,t k is the online degradation path P o The time scale of P o The length of the vector.

4. A method for predicting remaining battery life based on historical degradation paths according to claim 3, characterized in that: In S2, the final prediction path is calculated based on the optimal weight vector and translation variable, including: In the formula, is the final predicted path, is the optimal weight variable of the i-th historical degradation path, is the i-th historical degradation path in The performance degradation variable value at time , For the smallest N h is the number of historical degenerate paths.

5. The method for predicting remaining battery life based on historical degradation paths according to claim 4, characterized in that: In S3, the deviation between the final predicted path and the online degraded path in the latest time period is compensated by setting compensation variables, including: Setting a compensation variable Used to compensate the final predicted path With the online degradation path P o The deviation between the latest time stage, where the online degradation path P o Expressed as Final predicted path From t1 to t k The degradation path of t1=1,t k represents the online degenerate path P o Length; Among them, the compensation variable Obtained by the following formula: len=min(20,int(0.2*t k )); Where len is the time window length of the latest time stage, int(·) is the function of rounding down, is the jth performance degradation variable value of the online degradation path, is the jth performance degradation variable value of the final predicted path.

6. A method for predicting remaining battery life based on historical degradation paths according to claim 5, characterized in that: In S3, applying the compensation variable to the final predicted path to obtain a final predicted degradation path includes: The compensation variable will be obtained To compensate the final predicted path The error is obtained to obtain the nearest time stage of the final predicted path: In the formula, is the final predicted path.

7. A method for predicting remaining battery life based on historical degradation paths according to claim 6, characterized in that: In S4, the remaining life of the battery is calculated according to the final predicted degradation path and the set failure threshold, including: The failure threshold is set and the remaining life of the battery is calculated based on the final predicted degradation path: Where ξ is the failure threshold, The final predicted path In t k The performance degradation variable value at time +l, inf is the lower bound, The remaining battery life.

Images