Battery pack consistency online diagnosis method based on recursive curve reconstruction

Through the online diagnosis method of battery pack consistency based on recurve curve reconstruction, the problems of high computational complexity and high memory requirements in high voltage application scenarios are solved, and the online consistency index acquisition and ease of actual operating conditions are achieved Improved awesomeness and is suitable for real-time battery management systems.

CN119024217APending Publication Date: 2024-11-26BEIHANG UNIV +1
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411100633.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-12
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

The existing battery pack consistency diagnosis method has high computational complexity in high voltage application scenarios, and cannot achieve real-time online diagnosis, and requires a large amount of memory, which cannot meet the actual application needs.

Method used

The online diagnosis method of battery pack consistency based on recurve curve reconstruction is adopted. By obtaining the battery pack equivalent circuit, calculating the OCV of a single battery, establishing the curve relationship between OCV and SOC, using the extended Kalman filtering algorithm and the mean drift tracking algorithm Perform parameter tracking to reduce single-time computing and memory requirements.

Benefits of technology

It realizes online consistency indicator acquisition under actual operating conditions, improves the robustness and reliability of diagnosis, reduces the computational complexity and memory requirements, and is suitable for real-time battery management systems.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119024217A_ABST
    Figure CN119024217A_ABST
Patent Text Reader

Abstract

The invention provides a battery pack consistency online diagnosis method based on recursive curve reconstruction, and relates to the technical field of battery management, and the method comprises the steps: obtaining a battery pack equivalent circuit, and calculating the OCV of a single battery according to the equivalent circuit; obtaining a curvilinear relationship between the OCV and the SOC according to the OCV; performing curve reconstruction on the curve relation to obtain consistency parameters of the single batteries; and tracking the consistency parameter through a mean shift tracking algorithm to obtain a tracking result. According to the method, the calculated amount and the memory demand of single OCV are reduced, and the robustness and the reliability of diagnosis are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of battery management, and in particular to an online diagnosis method for battery pack consistency based on recursion curve reconstruction. Background Art

[0002] As a carrier for storing and converting electric energy, lithium batteries can convert into energy with extremely high efficiency without pollutant emissions. Among them, diagnosing the capacity consistency and state of charge (SOC) consistency of each single cell in the lithium battery pack is an important means to ensure the safe and efficient operation of the high-voltage battery system. Existing methods mainly include direct measurement method, state estimation method and information fusion method.

[0003] Among them, the direct measurement method tests the battery pack in an offline manner, and uses a full charge and discharge method to measure the capacity and SOC of the single battery in the battery pack. However, this method is time-consuming and can only be performed in a laboratory environment, and cannot be applied in practice. The state estimation method uses various state estimation algorithms, such as SOC estimation and capacity estimation algorithms based on battery models, to estimate the capacity and SOC state of all single batteries in the battery pack, thereby achieving consistency diagnosis. However, this method has a large amount of calculation, especially in high-voltage application scenarios. There are a large number of batteries in series, and the state estimation algorithm is executed on all batteries. Its computational complexity poses a huge challenge to the battery management system. The information fusion method evaluates the series consistency by selecting one or more consistency indicators, but this method needs to obtain complete consistency indicators, which requires a large amount of memory and is not suitable for real-time battery management system applications. Therefore, it is very necessary to design an online diagnosis method for battery pack consistency based on recursive curve reconstruction. Summary of the invention

[0004] The object of the present invention is to provide a battery pack consistency online diagnosis method based on recursive curve reconstruction to reduce the calculation amount and memory requirement of a single OCV and improve the robustness and reliability of the diagnosis.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A battery pack consistency online diagnosis method based on recursion curve reconstruction includes the following steps:

[0007] Obtain the battery pack equivalent circuit, and calculate the OCV of the single cell based on the equivalent circuit;

[0008] The curve relationship between OCV and SOC is obtained according to OCV;

[0009] Reconstruct the curve relationship to obtain the consistency parameters of the single battery;

[0010] The consistency parameters are tracked by the mean shift tracking algorithm to obtain the tracking results.

[0011] Optionally, the OCV of the single cell is obtained by calculating the equivalent circuit, and the specific steps include:

[0012] According to the equivalent circuit, the ratio of the terminal voltage of the single cell to the terminal voltage of the battery pack is regarded as the ratio of the OCV of the single cell to the OCV of the battery pack;

[0013] The OCV of the battery pack is obtained by using a recursive least squares algorithm with a forgetting factor;

[0014] The OCV of the single cell is obtained according to the OCV of the battery pack and the ratio of the OCV of the single cell to the OCV of the battery pack.

[0015] Optionally, the calculation formula of the recursive least squares algorithm with forgetting factor is:

[0016]

[0017] Among them, K k is the update gain, y k is the battery terminal voltage, is the regression vector, is the transposed matrix of the regression vector, ε k is the prediction error, P k is the error covariance at the current moment, P k-1 is the error covariance of the previous moment, λ k is the forgetting factor, I is the current, θ k is the unknown parameter vector at the current moment, θ k-1 is the unknown parameter vector at the previous moment.

[0018] Optionally, the expression of the curve relationship is:

[0019] z i-1 =z0+Q ah,i / C i ;

[0020] Among them, U oc is OCV, z is SOC, f() is z and U oc The curve relationship, a i are the polynomial coefficients, z i-1 is the SOC of the i-1th battery, z0 is the initial SOC when the battery starts discharging, Q ah,i is the cumulative amount of electricity discharged during the discharge process, C i is the capacity of a single battery.

[0021] Optionally, the curve relationship is reconstructed to obtain the consistency parameter of the single battery, and the specific steps are: expressing the curve relationship as an objective function, and solving the objective function by an extended Kalman filter algorithm;

[0022] The objective function expression is: Among them, J t (ω) is the objective function, is the consistency parameter, U oc,i is the OCV of the ith single cell, t is the time, Q ah,i The cumulative amount of electricity discharged during the discharge process;

[0023] The calculation formula of the extended Kalman filter algorithm is: P t+1 =P t -K t H t P t +Q t ;in, is the consistency parameter, K t is the Kalman filter gain, P t is the error covariance matrix at the current moment, P t+1 is the error covariance matrix at the next moment, H t is the observation matrix, is the transposed matrix of the observation matrix, R t is the measurement variance, Q t is the system variance, y t is the real-time OCV, ω is the consistency parameter to be solved, ω t+1 is the estimated value at the next moment, ω t is the estimated value at the current moment, and x is the OCV and Q of the i-th single cell when taking the partial derivative ah,i The real-time value of .

[0024] Optionally, the consistency parameters include: a battery capacity estimation value and an initial SOC estimation value.

[0025] Optionally, the consistency parameters are tracked by a mean shift tracking algorithm to obtain tracking results. The specific steps are: use the battery capacity estimate and the initial SOC estimate as data points in the mean shift tracking algorithm, select any data point as the starting point, and gradually track the starting point by the mean shift tracking algorithm until the center of the data point with the highest probability density is tracked.

[0026] Optionally, the expression for step-by-step tracking is: t =x t+1 +M h ; where x t is the data point at the current moment, xt+1 is the data point at the next moment, M h is the drift vector; the calculation formula of the drift vector is: Among them, k is the number of data points in the current valid data set, x i is the consistency parameter of the ith single cell battery.

[0027] According to the specific embodiment provided by the present invention, the present invention discloses the following technical effects: the online diagnosis method of battery pack consistency based on recursive curve reconstruction provided by the present invention includes: obtaining the equivalent circuit of the battery pack, and calculating the OCV of the single cell according to the equivalent circuit; obtaining the curve relationship between OCV and SOC according to OCV; reconstructing the curve relationship to obtain the consistency parameter of the single cell; tracking the consistency parameter through the mean drift tracking algorithm to obtain the tracking result. This method reduces the calculation amount and memory requirement of a single OCV, and improves the robustness and reliability of diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] 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.

[0029] Figure 1 The flowchart of the online diagnosis method of battery pack consistency based on recursive curve reconstruction according to an embodiment of the present invention. DETAILED DESCRIPTION

[0030] 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.

[0031] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and understandable, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0032] like Figure 1 As shown, an embodiment of the present invention provides a battery pack consistency online diagnosis method based on recursion curve reconstruction, comprising the following steps:

[0033] Step 100: Obtaining an equivalent circuit of a battery pack, and calculating an open circuit voltage (OCV) of a single cell according to the equivalent circuit;

[0034] Specifically, each single cell in the battery pack is equivalent to a first-order RC model. t The terminal voltage U t,i The expression is:

[0035] U t,i =U oc,i +U p,i +I t R s,i ;

[0036] Among them, U oc,i is the OCV of the ith single cell, R s,i and U p,i are the DC internal resistance and polarization voltage of the ith single cell respectively. The battery terminal voltage U t,p is the sum of the terminal voltages of all single cells, and the expression is:

[0037] Wherein, n is the number of single cells.

[0038] Specifically, the OCV of a single cell is obtained by calculating the equivalent circuit, and the specific steps include:

[0039] According to the equivalent circuit, the ratio of the terminal voltage of the single cell to the terminal voltage of the battery pack is regarded as the ratio of the OCV of the single cell to the OCV of the battery pack. The specific expression is:

[0040] U oc,i / U oc,p =U t,i / U t,p Among them, U oc,p is the OCV of the battery pack.

[0041] The OCV of the battery pack is obtained by using a recursive least squares algorithm with a forgetting factor. The calculation formula of the recursive least squares algorithm with a forgetting factor is:

[0042]

[0043] θ k =θ k-1 +K k ε k ;

[0044]

[0045] Among them, K k is the update gain, y k is the battery terminal voltage, is the regression vector, is the transposed matrix of the regression vector, ε k is the prediction error, Pk is the error covariance at the current moment, P k-1 is the error covariance of the previous moment, λ k is the forgetting factor, I is the current, θ k is the unknown parameter vector at the current moment, θ k-1 is the unknown parameter vector at the previous moment. The OCV of the battery pack is extracted by recursively calculating the parameter θ of the least squares algorithm based on the forgetting factor. The calculation formula is:

[0046] U oc,p =θ6 / (1-θ1-θ2);

[0047] Among them, θ1, θ2 and θ6 are the 1st, 2nd and 6th values ​​of parameter θ respectively.

[0048] The OCV of a single cell is obtained by the ratio of the OCV of the battery pack and the OCV of the single cell to the OCV of the battery pack. The calculation formula is:

[0049] Step 200: Obtain a curve relationship between OCV and SOC according to OCV; the curve relationship is a nonlinear relationship, and its calculation formula is:

[0050]

[0051] Among them, U oc is OCV, z is SOC, a i are the polynomial coefficients, z i-1 is the SOC of the i-1th battery. The SOC of the discharge process is related to the initial SOC value and the accumulated discharge capacity. The SOC calculation formula of the i-1th battery is:

[0052] z i-1 =z0+Q ah,i / C i ;

[0053] Among them, z0 is the initial SOC when the battery starts to discharge, Q ah,i is the cumulative amount of electricity discharged during the discharge process of the ith battery, C i is the capacity of the ith battery.

[0054] Step 300: reconstructing the curve relationship to obtain the consistency parameter of the single cell;

[0055] Specifically, the specific steps of curve reconstruction are: expressing the curve relationship as an objective function, and solving the objective function through an extended Kalman filter (EKF) algorithm.

[0056] More specifically, the objective function is calculated as: Among them, J t (ω) is the objective function, is the consistency parameter, U oc,i is the OCV of the ith single cell. Methods for solving the objective function include particle swarm optimization, Gauss-Newton method, genetic algorithm, etc., but these methods require all data to be put into the optimization solution algorithm after the discharge process is completed, which is not suitable for real-time applications. Therefore, this embodiment uses the extended Kalman filter algorithm to solve the objective function.

[0057] The calculation formula of the extended Kalman filter algorithm is:

[0058]

[0059] P t+1 =P t -K t H t P t +Q t ;

[0060] in, is the consistency parameter, K t is the Kalman filter gain, P t is the error covariance matrix at the current moment, P t+1 is the error covariance matrix at the next moment, H t is the observation matrix, is the transposed matrix of the observation matrix, R t is the measurement variance, Q t is the system variance, y t is the real-time OCV, ω is the consistency parameter to be solved, ω t+1 is the estimated value at the next moment, ω t is the estimated value at the current moment, and x is the OCV and Q of the i-th single cell when taking the partial derivative ah,i The real-time value of .

[0061] Step 400: Track the consistency parameter through the mean shift tracking algorithm (MS) to obtain the tracking result. The specific steps include:

[0062] The calculation result of EKF, that is, It is represented by x and is considered as a data point in the MS algorithm;

[0063] The MS algorithm selects any data point as the starting point and gradually tracks to the center of the data point with the highest probability density through iterative updates.

[0064] Specifically, the center of the data point with the highest probability density is gradually tracked through iterative updates. The specific steps include:

[0065] Use the drift vector to determine the movement distance of the data point. The calculation formula is:

[0066] x t =x t+1 +M h ;

[0067]

[0068] Among them, x t is the data point at the current moment, x t+1 is the data point at the next moment, M h is the drift vector and k is the number of data points in the valid data set of the current iteration.

[0069] Replace the average form of the drift vector with a Gaussian weighted form and simplify the data points of the current iteration step. The calculation formula is:

[0070]

[0071] K(x i -x) = exp(-c||x i -x|| 2 );

[0072] S k (x) = {y:(yx i ) T (yx i )<h 2};

[0073] Among them, K is the Gaussian kernel function, c is a constant, S k (x) is the set of data points within the radius h of the current value x, y is the coordinates of the current center point when valid data is defined, () T is the transpose of the coordinate vector.

[0074] The beneficial effects of the present invention are as follows:

[0075] 1) The consistency index can be extracted online under actual operating conditions, which improves the adaptability of the working conditions for obtaining the consistency index;

[0076] 2) Reduce the memory requirements and computational complexity of a single solution process, and improve the real-time performance of the algorithm;

[0077] 3) The mean shift tracking algorithm is used to reduce the impact of abnormal data and calculation results and improve robustness.

[0078] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0079] The present invention uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only used to help understand the method and core ideas of the present invention. At the same time, for those skilled in the art, according to the ideas 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 battery pack consistency online diagnosis method based on recursion curve reconstruction, characterized in that: The steps include: Obtaining an equivalent circuit of a battery pack, and calculating the OCV of a single battery according to the equivalent circuit; Obtaining a curve relationship between OCV and SOC according to the OCV; Reconstructing the curve relationship to obtain the consistency parameter of the single cell; The consistency parameter is tracked by a mean shift tracking algorithm to obtain a tracking result.

2. The battery pack consistency online diagnosis method based on recursion curve reconstruction according to claim 1 is characterized in that: The OCV of the single cell is calculated according to the equivalent circuit, and the specific steps include: According to the equivalent circuit, the ratio of the terminal voltage of the single cell to the terminal voltage of the battery pack is regarded as the ratio of the OCV of the single cell to the OCV of the battery pack; The OCV of the battery pack is obtained by using a recursive least squares algorithm with a forgetting factor; The OCV of the single cell is obtained according to the OCV of the battery pack and the ratio of the OCV of the single cell to the OCV of the battery pack.

3. The battery pack consistency online diagnosis method based on recursion curve reconstruction according to claim 2 is characterized in that: The calculation formula of the recursive least squares algorithm with forgetting factor is: i k =θ k-1 +K k e k ; Among them, K k is the update gain, y k is the battery terminal voltage, is the regression vector, is the transposed matrix of the regression vector, ε k is the prediction error, P k is the error covariance at the current moment, P k-1 is the error covariance of the previous moment, λ k is the forgetting factor, I is the current, θ k is the unknown parameter vector at the current moment, θ k-1 is the unknown parameter vector at the previous moment.

4. The battery pack consistency online diagnosis method based on recursion curve reconstruction according to claim 1 is characterized in that: The expression of the curve relationship is: Among them, U oc is OCV, z is SOC, f() is z and U oc The curve relationship, a i are the polynomial coefficients, z i-1 is the SOC of the i-1th battery, z0 is the initial SOC when the battery starts discharging, Q ah,i is the cumulative amount of electricity discharged during the discharge process, C i is the capacity of a single battery.

5. The battery pack consistency online diagnosis method based on recursion curve reconstruction according to claim 4 is characterized in that: Reconstructing the curve relationship to obtain the consistency parameter of the single battery, the specific steps are: expressing the curve relationship as an objective function, and solving the objective function by using an extended Kalman filter algorithm; The objective function expression is: Among them, J t (ω) is the objective function, is the consistency parameter, U oc,i is the OCV of the ith single cell, t is the time, Q ah,i The cumulative amount of electricity discharged during the discharge process; The calculation formula of the extended Kalman filter algorithm is: P t+1 =P t -K t H t P t +Q t ;in, is the consistency parameter, K t is the Kalman filter gain, P t is the error covariance matrix at the current moment, P t+1 is the error covariance matrix at the next moment, H t is the observation matrix, is the transposed matrix of the observation matrix, R t is the measurement variance, Q t is the system variance, y t is the real-time OCV, ω is the consistency parameter to be solved, ω t+1 is the estimated value at the next moment, ω t is the estimated value at the current moment, and x is the OCV and Q of the i-th single cell when taking the partial derivative ah,i The real-time value of .

6. The battery pack consistency online diagnosis method based on recursion curve reconstruction according to claim 5 is characterized in that: The consistency parameters include: a battery capacity estimation value and an initial SOC estimation value.

7. The battery pack consistency online diagnosis method based on recursion curve reconstruction according to claim 6 is characterized in that: The consistency parameter is tracked by a mean shift tracking algorithm to obtain a tracking result. The specific steps are: taking the battery capacity estimation value and the initial SOC estimation value as data points in the mean shift tracking algorithm, selecting any of the data points as a starting point, and gradually tracking the starting point by the mean shift tracking algorithm until the center of the data point with the highest probability density is tracked.

8. The battery pack consistency online diagnosis method based on recursion curve reconstruction according to claim 7 is characterized in that: The expression of the step-by-step tracking is: t =x t+1 +M h ; where x t is the data point at the current moment, x t+1 is the data point at the next moment, M h is the drift vector; the calculation formula of the drift vector is: Among them, k is the number of data points in the current valid data set, x i is the consistency parameter of the ith single cell.

Citation Information

Cited By

  • Intelligent management method and system for general sodium ion battery for electric vehicle and large-scale energy storage

    CN119550877A