An Online SOC Estimation Method for Power Battery Packs Considering Inconsistencies

By establishing an equivalent circuit model of the power battery pack and using the entropy weight-TOPSIS method, the problem of SOC estimation accuracy caused by the differences in individual cells was solved, and high-precision online estimation of the SOC of the power battery pack was achieved.

CN117607693BActive Publication Date: 2026-11-10NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202311534994.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-11-17
Publication Date
2026-11-10
Estimated Expiration
2043-11-17

AI Technical Summary

Technical Problem

Existing technologies fail to consider the differences between individual cells in the SOC estimation of power battery packs, resulting in poor estimation accuracy.

Method used

By establishing an equivalent circuit model of the power battery pack that considers inconsistencies, the average performance of the battery pack and the performance differences of individual cells are obtained. SOC estimation is performed using a filtering algorithm and the entropy weight-TOPSIS method. The entropy weight method is then combined to evaluate the inconsistency of individual cells, thus achieving SOC fusion estimation.

Benefits of technology

It achieves accurate estimation of the SOC of the power battery pack while considering the inconsistency of individual cells, with an error of less than 2% and an RMSE value of approximately 1.8%, thus improving the accuracy of SOC estimation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117607693B_ABST
    Figure CN117607693B_ABST
Patent Text Reader

Abstract

This invention provides an online SOC estimation method for power battery packs that considers inconsistencies. It addresses secondary inconsistencies in voltage, internal resistance, and capacity, enabling accurate online SOC estimation. The method includes: 1) establishing an equivalent model of the power battery pack based on a mean difference model; 2) proposing an inconsistency characterization method based on entropy weight-TOPSIS; and 3) proposing a fusion estimation method for power battery pack SOC that considers inconsistencies. This invention considers the impact of individual cell inconsistencies on the SOC of the power battery pack, more accurately reflecting the state of charge of the power battery system. It has significant engineering and application value for improving the accuracy of state estimation and prediction of power battery systems.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of battery management technology, and in particular to an online method for estimating the state of charge (SOC) of a power battery pack that takes into account inconsistencies. Background Technology

[0002] SOC (State of Charge) refers to the state of charge of a battery, which describes the remaining capacity of the battery. It is one of the most important parameters of a battery management system, and the accuracy of its estimation is a necessary requirement for a high-performance BMS (Battery Management System), directly affecting the performance of the battery system.

[0003] Currently, the SOC estimation of power battery packs mainly adopts the method of using the highest or lowest single cell voltage in the battery pack as the threshold value and using the ampere-hour integration method to estimate the SOC of the entire battery pack. This method does not consider the impact of the differences between individual cells on the SOC of the entire battery pack, and cannot accurately reflect the overall state of charge of the battery pack, resulting in poor estimation accuracy. Summary of the Invention

[0004] Objective of the Invention: The technical problem to be solved by the present invention is to address the shortcomings of the prior art by providing an online SOC estimation method for power battery packs that takes into account inconsistencies, comprising the following steps:

[0005] Step 1: Considering the impact of secondary inconsistencies during actual use of the battery pack, obtain the real-time parameters of the power battery pack and each individual cell. The real-time parameters include terminal voltage and discharge current.

[0006] Step 2: Establish an equivalent circuit model of the power battery pack considering inconsistencies. Based on the mean difference model of the power battery pack, obtain the average performance of the battery pack and the difference performance of each individual cell. Use a filtering algorithm to obtain the SOC estimate of the mean model. m The deviation value ΔSOC of the SOC of each individual cell;

[0007] Step 3: Correct the SOC estimate of the mean model using the SOC deviation value of each individual cell to obtain the SOC estimate of each individual cell in the current power battery pack. i ;

[0008] Step 4: Combine the entropy weight-TOPSIS method to evaluate the inconsistency of individual cells in the power battery pack, and obtain the fusion weight C of the power battery pack's SOC. i This enables the fusion estimation of the SOC of the power battery pack.

[0009] In step 1, the influencing factors of the secondary inconsistency include internal resistance, voltage, and capacitance.

[0010] Step 2 includes:

[0011] Step 2-1: Based on the second-order Thevenin model, establish a mean value model to characterize the mean state of the power battery pack.

[0012] Step 2-2: Based on the internal resistance model, establish a single-cell difference model to characterize the deviation of each battery cell from the mean state.

[0013] Steps 2-3: Using test data, identify the unknown parameters in the mean model and the difference model offline and online;

[0014] Steps 2-4: Combine the filtering algorithm to obtain the SOC estimate of the mean model, and use the difference model to correct the difference of the mean model estimate to obtain the SOC estimate of each individual cell.

[0015] Step 2-1 includes: The state equation of the mean model is:

[0016] U t =U ocv -R0I-U1-U2

[0017]

[0018]

[0019] Among them, U t U represents the battery terminal voltage. o c v R0 represents the battery open-circuit voltage, I represents the battery equivalent ohmic internal resistance, U1 and U2 represent the polarization voltage inside the battery, t represents the corresponding operating time of the battery, R1 and R2 represent the electrochemical polarization resistance and concentration polarization resistance of the battery, respectively, C1 and C2 represent the electrochemical polarization capacitance and concentration polarization capacitance, respectively, and d is the integral symbol.

[0020] Step 2-2 includes: The state equation of the monomer difference model is:

[0021]

[0022] in, This represents the average discharge voltage U of the i-th individual cell in the battery pack. t The difference between them This represents the voltage U of the i-th battery cell and the average open-circuit voltage. OCV The difference between them This represents the difference between the i-th battery cell and the average internal resistance R0;

[0023] The input to the differential model is the difference in voltage and discharge current between the current average state of the power battery pack and the individual cells, which are acquired in real time. The output is the state-of-charge (SOC) estimation deviation value ΔSOC of the individual cell differential model. i .

[0024] Steps 2-3 include:

[0025] The terminal voltage and discharge current of the current power battery pack and individual cells are acquired in real time, and parameter identification is performed using offline parameter identification and online parameter identification methods.

[0026] The parameter to be identified in the mean model is the open-circuit voltage U. ocv The parameters to be identified in the difference model are: ohmic internal resistance R0, electrochemical polarization capacitance C1, concentration polarization capacitance C2, electrochemical polarization resistance R1 and concentration polarization resistance R2 of the battery, and the single-cell open-circuit voltage difference. and internal resistance difference

[0027] Steps 2-4 include: The mean model simulates the average performance of the battery pack, and the state of charge (SOC) of the power battery pack under the mean state is obtained using the ampere-hour integral method and the Kalman filter algorithm. m The calculation formula is:

[0028]

[0029] Among them, SOC0 and SOC m (t) represents the SOC at the initial time and the SOC of the mean model at time t, respectively; Q N Rated capacity; η is the integral of the battery operating current from the initial time to time t; η is the charge / discharge efficiency.

[0030] The single-cell difference model characterizes the deviation value of each cell in the power battery pack. After correcting the mean model, the single-cell state-of-the-art (SOC) is obtained; combined with the mean state-of-the-art (SOC) estimate, the final SOC is obtained. m ΔSOC obtained from the difference model i Estimate the State of Charge (SOC) of each individual cell in the power battery pack. i =SOC m +ΔSOC i SOC i This represents the estimated state of charge of the i-th individual cell.

[0031] Step 4 proposes a method for characterizing inconsistencies in power battery packs, including the following steps:

[0032] Step 4-a1: Establish the evaluation index system and calculate the entropy value of each index. Where S jThis represents the entropy value corresponding to the j-th indicator, where j is 1, 2, or 3, specifically representing indicators such as battery voltage, internal resistance, and capacity; i refers to each individual cell in the power battery pack; m is the total number of indicators in the indicator evaluation system; n is the number of battery cells; y ij The value of the j-th index of the i-th individual is obtained after dimensionless processing.

[0033] Step 4-a2: Calculate the weight values ​​of each indicator within the evaluation system, and the weight value W of the j-th indicator. j The calculation formula is

[0034] Step 4-a3: Assign the weights determined by the entropy weight method to the standard mean to obtain the weighted evaluation matrix V = (v ij ) m×n Determine the positive ideal solution V + Negative ideal solution V - :

[0035] V + ={(maxv ij / j∈j1), (minv ij / j∈j2)}

[0036] V - ={(minv ij / j∈j1), (minv ij / j∈j2)}

[0037] Where the intermediate parameter v ij v is obtained by using the j-th index of the i-th individual and the weight value calculated by the entropy weight method. ij =w i y ij j1 is the set of profit-generating indicators; j2 is the set of loss-generating indicators.

[0038] Step 4-a4: Calculate the Euclidean distance between each evaluation object and the positive ideal solution. Euclidean distance between each evaluation object and the negative ideal solution

[0039]

[0040]

[0041] Step 4-a4, calculate the relative proximity coefficient C. i Characterizing the inconsistency of power battery packs:

[0042]

[0043] Step 4-a1 specifically includes:

[0044] The evaluation index system takes each individual cell in the power battery pack as the evaluation object. Inconsistent index parameters include battery voltage, internal resistance, and capacity. Among them, voltage and internal resistance are more favorable when they are close to the mean, and capacity is more favorable when it is larger. The dimensionless processing method is as follows:

[0045]

[0046]

[0047] Where x ij This represents the value of the j-th index of the i-th individual; This represents the average value of the j-th indicator;

[0048] In step 4, the SOC fusion estimation of the power battery pack is achieved using the following formula:

[0049]

[0050] SOC Pack This refers to the state of charge of the battery pack.

[0051] When individual battery cells are connected in series and parallel to form a battery pack, the states of each cell influence each other, making it impossible to characterize the state of the battery pack by individual cells. Therefore, this invention considers the inconsistencies among the individual cells in the battery pack and their characterization methods, studies their impact on the overall charge performance of the power battery pack, and proposes a power battery pack SOC estimation method that considers inconsistencies. Achieving accurate online estimation of this method has significant research and engineering application value.

[0052] Beneficial effects:

[0053] (1) The present invention provides an online SOC estimation method for power battery packs that considers inconsistency. It takes into account the impact of inconsistency of individual cells on the SOC estimation of power battery packs and proposes an SOC estimation method for power battery packs by combining three indicators: voltage, internal resistance and capacity.

[0054] (2) This invention proposes characterization methods for voltage, internal resistance, and capacity to address inconsistencies in power battery packs, which have certain practical value and reference significance. Based on this, the definitions of other inconsistencies in power battery packs can be expanded, thereby achieving a more accurate state estimation of the battery pack. Attached Figure Description

[0055] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments, and the advantages of the present invention in the above and / or other aspects will become clearer.

[0056] Figure 1 A flowchart for the fusion estimation of the SOC of a power battery pack to account for inconsistencies.

[0057] Figure 2a This is a schematic diagram of a power battery pack model based on the mean value model.

[0058] Figure 2b This is a schematic diagram of a power battery pack model based on a difference model.

[0059] Figure 3 This is a flowchart for characterizing inconsistencies in power battery packs.

[0060] Figure 4a This is a schematic diagram illustrating the estimated SOC value under test conditions.

[0061] Figure 4b This is a schematic diagram illustrating the estimation error of SOC under test conditions. Detailed Implementation

[0062] like Figure 1 As shown, this invention provides an online SOC estimation method for power battery packs that considers inconsistencies, including establishing a mean difference model that considers the inconsistencies of power battery packs, proposing a method for characterizing the inconsistencies of power battery packs, and proposing a SOC fusion estimation method for power battery packs that considers inconsistencies.

[0063] Establishing a mean difference model that considers the inconsistency of power battery packs includes the following steps:

[0064] Step 1: First, establish a model based on the second-order Thevenin model, such as... Figure 2a The power battery pack mean value model shown represents the mean state of each individual cell in the power battery pack. The state equation of the model is expressed as follows:

[0065] U t =U ocv -R0I-U1-U2

[0066]

[0067]

[0068] Among them, U t U represents the battery terminal voltage. ocv R0 represents the battery open-circuit voltage, I represents the battery equivalent internal resistance in ohms, U1 and U2 represent the polarization voltage inside the battery, t represents the corresponding operating time of the battery, R1 and R2 represent the electrochemical polarization resistance and concentration polarization resistance of the battery, and C1 and C2 represent the electrochemical polarization capacitance and concentration polarization capacitance.

[0069] Step 2: Real-time acquisition of the terminal voltage and discharge current of the current power battery pack and individual cells to obtain the terminal voltage and discharge current in the mean state. Based on offline HPPC experiments and the least squares online parameter identification method, identify parameters such as the ohmic internal resistance and polarization capacitor resistance of the model in the mean state.

[0070] Considering the impact of the inconsistency in the initial state of the power battery pack's SOC, the initial state SOC0 and open-circuit voltage U are used. ocv The relationship between resting time t and the following is obtained:

[0071]

[0072] In the formula, k1 and k2 are the dynamic change parameters for SOC estimation, and I is the discharge rate.

[0073] Step 3: Next, establish a model based on the internal resistance model, such as... Figure 2b The differential model of the power battery pack shown takes into account accuracy and computational complexity. It uses an internal resistance model to characterize the deviation of each battery cell from the mean state. The state equation is:

[0074]

[0075] in, This represents the average discharge voltage U of the i-th individual cell in the battery pack. t The difference between them This represents the voltage U of the i-th battery cell and the average open-circuit voltage. OCV The difference between them This represents the difference between the i-th cell and the average internal resistance R0.

[0076] Step 4: Use the least squares method to perform online parameter identification for the differential model, and combine it with the Kalman filter algorithm to obtain the state of charge estimation deviation value ΔSOC for each individual differential model. i For the mean state SOC estimate, SOC m Make corrections and estimate the State of Charge (SOC) of each individual cell in the power battery pack. i =SOC m +ΔSOC i .

[0077] For the characterization method of inconsistency in power battery packs, such as Figure 3 As shown, it includes the following steps:

[0078] Step 1: Taking each individual cell in the power battery pack as the evaluation object, and the inconsistent index parameters as battery voltage, internal resistance, and capacity, establish a dimensionless evaluation index system Y = (y ij ) m×nAmong them, voltage and internal resistance are more favorable when they are close to their average values, while capacitance is more favorable when it is larger. The processing methods are as follows:

[0079]

[0080]

[0081] Where x ij Let represent the evaluation value corresponding to the j-th indicator of the i-th evaluated object.

[0082] Step 2: Based on the evaluation index system, calculate the entropy value of each index. In the formula, 0≤S j ≤1, which means when y ij When = 0 or 1, ln(y) ij ) = 0.

[0083] Step 3: Based on the concept of entropy, the difference between the j-th indicator corresponding to different evaluation objects reflects the amount of information. The greater the difference, the greater the amount of information reflected. Therefore, the difference coefficient of the j-th indicator is defined as d. j =1-S j .

[0084] Step 4: Calculate the weight value of each indicator within the evaluation system.

[0085] Step 5: Assign the weights determined by the entropy weight method to the standard mean value to obtain the weighted evaluation matrix V = (v ij ) m×n Thus, the positive and negative ideal solutions V are determined. ± :

[0086] V + ={(maxv ij / j∈j1), (minv ij / j∈j2)}

[0087] V-={(minv ij / j∈j1), (minv ij / j∈j2)}

[0088] Step 6: Calculate the Euclidean distance between each evaluation object and the positive and negative ideal solutions:

[0089]

[0090]

[0091] Step 7, Relative Proximity Coefficient C i This characterizes the inconsistency of the power battery pack, namely:

[0092] The SOC fusion estimation method for power battery packs that considers inconsistencies includes the following steps:

[0093] Step 1: Obtain the parameters and real-time status of the individual cells in the power battery pack.

[0094] Step 2: Establish a mean value model for the power battery pack, and use a filtering algorithm to estimate the SOC of the mean value model. The input of the mean value model is the mean voltage U. CMM The current I is output as the battery SOC estimate under the mean model. m .

[0095] Step 3: Establish a single-cell difference model for the power battery pack to characterize the deviation of each battery cell from the mean state. The input to the difference model is the difference voltage U. CDM Current I, SOC under the mean model m The estimated output is the state of charge (SOC) of a single battery cell. i =SOC m +ΔSOC.

[0096] Step 4: Obtain the inconsistency characterization parameters of each individual cell in the power battery pack using the entropy weight-TOPSIS method. The specific process is as follows:

[0097] ① Obtain the index weight W using the entropy weight method j ;

[0098] ②Use the weights to obtain the weighted evaluation matrix, obtain the set of positive and negative ideal solutions, and calculate the Euclidean distance between each evaluation object and the positive and negative ideal solutions;

[0099] ③ Calculate the relative proximity coefficient C i This characterizes the inconsistency of the power battery pack.

[0100] Step 5: Establish the SOC obtained from the mean difference model of the power battery pack. i The relationship between the inconsistent characterization parameters obtained from the entropy-weighted TOPSIS method and the SOC of the power battery pack is used to perform a fusion estimation.

[0101] Step 6: Verify the accuracy of the model and method using HPPC conditions. The verification results are as follows: Figure 4a , Figure 4b As shown, the battery pack SOC estimation method proposed in this invention has an estimation error of no more than 2% and an RMSE value of approximately 1.8%, which can achieve effective estimation of the power battery pack SOC even when considering inconsistencies.

[0102] This invention provides an online SOC estimation method for power battery packs that considers inconsistencies. Many methods and approaches exist for implementing this technical solution; the above description is merely a preferred embodiment of the invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this invention, and these improvements and modifications should also be considered within the scope of protection of this invention. All components not explicitly stated in this embodiment can be implemented using existing technologies.

Claims

1. A method for online estimation of the State of Charge (SOC) of a power battery pack considering inconsistencies, characterized in that, Includes the following steps: Step 1: Considering the impact of secondary inconsistencies during actual use of the battery pack, obtain the real-time parameters of the power battery pack and each individual cell. The real-time parameters include terminal voltage and discharge current. Step 2: Establish an equivalent circuit model of the power battery pack considering inconsistencies. Based on the mean difference model of the power battery pack, obtain the average performance of the battery pack and the difference performance of each individual cell. Use a filtering algorithm to obtain the SOC estimate of the mean model. m The deviation value ΔSOC of the SOC of each individual cell; Step 3: Correct the SOC estimate of the mean model using the SOC deviation value of each individual cell to obtain the SOC estimate of each individual cell in the current power battery pack. i ; Step 4: Combine the entropy weight-TOPSIS method to evaluate the inconsistency of individual cells in the power battery pack, and obtain the fusion weight C of the power battery pack's SOC. i This enables the fusion estimation of the SOC of the power battery pack.

2. The method according to claim 1, characterized in that, In step 1, the influencing factors of the secondary inconsistency include internal resistance, voltage, and capacitance.

3. The method according to claim 2, characterized in that, Step 2 includes: Step 2-1: Based on the second-order Thevenin model, establish a mean value model to characterize the mean state of the power battery pack. Step 2-2: Based on the internal resistance model, establish a single-cell difference model to characterize the deviation of each battery cell from the mean state. Steps 2-3: Using test data, identify the unknown parameters in the mean model and the difference model offline and online; Steps 2-4: Combine the filtering algorithm to obtain the SOC estimate of the mean model, and use the difference model to correct the difference of the mean model estimate to obtain the SOC estimate of each individual cell.

4. The method according to claim 3, characterized in that, Step 2-1 includes: The state equation of the mean model is: U t =U ocv -R0I-U1-U2 Among them, U t U represents the battery terminal voltage. ocv R0 represents the battery's open-circuit voltage, I represents the battery's equivalent ohmic internal resistance, U1 and U2 represent the battery's internal polarization voltages, t represents the battery's corresponding operating time, R1 and R2 represent the battery's electrochemical polarization resistance and concentration polarization resistance, respectively, C1 and C2 represent the battery's electrochemical polarization capacitance and concentration polarization capacitance, respectively, and s is the integral symbol.

5. The method according to claim 4, characterized in that, Step 2-2 includes: The state equation of the monomer difference model is: in, This represents the average discharge voltage U of the i-th individual cell in the battery pack. t The difference between them This represents the voltage U of the i-th battery cell and the average open-circuit voltage. OCV The difference between them This represents the difference between the i-th battery cell and the average internal resistance R0; The input to the differential model is the difference in voltage and discharge current between the current average state of the power battery pack and the individual cells, which are acquired in real time. The output is the state-of-charge (SOC) estimation deviation value ΔSOC of the individual cell differential model. i .

6. The method according to claim 5, characterized in that, Steps 2-3 include: The terminal voltage and discharge current of the current power battery pack and individual cells are acquired in real time, and parameter identification is performed using offline parameter identification and online parameter identification methods. The parameter to be identified in the mean model is the open-circuit voltage U. ocv The parameters to be identified in the difference model are: ohmic internal resistance R0, electrochemical polarization capacitance C1, concentration polarization capacitance C2, electrochemical polarization resistance R1 and concentration polarization resistance R2 of the battery, and the single-cell open-circuit voltage difference. and internal resistance difference 7. The method according to claim 6, characterized in that, Steps 2-4 include: The mean model simulates the average performance of the battery pack, and the state of charge (SOC) of the power battery pack under the mean state is obtained using the ampere-hour integral method and the Kalman filter algorithm. m The calculation formula is: Among them, SOC0 and SOC m (t) represents the SOC at the initial time and the SOC of the mean model at time t, respectively; Q N Rated capacity; η is the integral of the battery operating current from the initial time to time t; η is the charge / discharge efficiency. The single-cell difference model characterizes the deviation value of each cell in the power battery pack. After correcting the mean model, the single-cell state-of-the-art (SOC) is obtained; combined with the mean state-of-the-art (SOC) estimate, the final SOC is obtained. m ΔSOC obtained from the difference model i Estimate the State of Charge (SOC) of each individual cell in the power battery pack. i =SOC m +ΔSOC i SOC i This represents the estimated state of charge of the i-th individual cell.

8. The method according to claim 7, characterized in that, Step 4 proposes a method for characterizing inconsistencies in power battery packs, including the following steps: Step 4-a1: Establish the evaluation index system and calculate the entropy value of each index. Where S j Let y be the entropy value corresponding to the j-th indicator, where j is 1, 2, 3; i refers to each cell in the power battery pack; m is the total number of indicators in the indicator evaluation system; n is the number of battery cells; y ij The value of the j-th index of the i-th individual is obtained after dimensionless processing. Step 4-a2: Calculate the weight values ​​of each indicator within the evaluation system, and the weight value W of the j-th indicator. j The calculation formula is Step 4-a3: Assign the weights determined by the entropy weight method to the standard mean to obtain the weighted evaluation matrix V = (v ij ) m×n Determine the positive ideal solution V + Negative ideal solution V - : V + ={(max v ij / j∈j1),(min v ij / j∈j2)} V - ={(min v ij / j∈j1),(min v ij / j∈j2)} Where the intermediate parameter v ij v is obtained by using the j-th index of the i-th individual and the weight value calculated by the entropy weight method. ij =w i y ij j1 is the set of profit-generating indicators; j2 is the set of loss-generating indicators. Step 4-a4: Calculate the Euclidean distance between each evaluation object and the positive ideal solution. Euclidean distance between each evaluation object and the negative ideal solution Step 4-a4, calculate the relative proximity coefficient C. i Characterizing the inconsistency of power battery packs:

9. The method according to claim 8, characterized in that, Step 4-a1 specifically includes: The evaluation index system takes each individual cell in the power battery pack as the evaluation object. Inconsistent index parameters include battery voltage, internal resistance, and capacity. The dimensionless processing method is as follows: Where x ij This represents the value of the j-th index of the i-th individual; This represents the average value of the j-th indicator.

10. The method according to claim 9, characterized in that, In step 4, the SOC fusion estimation of the power battery pack is achieved using the following formula: SOC Pack This refers to the state of charge of the battery pack.

Citation Information

Patent Citations

  • Method for estimating SOC (State of Charge) of single battery in battery pack by considering influence of multiple factors

    CN114200327A

  • Energy storage battery system state estimation method considering inconsistency

    CN115201679A