Battery health state assessment method

Abstract

The invention relates to a battery health state assessment method, which comprises the following steps: carrying out charge and discharge test on a battery to obtain relaxation voltage data in a set time period and fitting the relaxation voltage data to obtain a relaxation voltage change curve; solving an angular bisector based on fitting parameters of the relaxation voltage change curve; solving to obtain inflection point voltage according to the relaxation voltage data and the angular bisector; repeating the process to obtain inflection point voltages corresponding to the plurality of batteries; performing linear fitting on the inflection point voltage of each battery and the corresponding SOH value of the battery to obtain a linear relation between the inflection point voltage and the SOH value; for a to-be-detected battery, the inflection point voltage of the to-be-detected battery is firstly obtained, and then the SOH value of the to-be-detected battery is searched and determined by combining the linear relation between the inflection point voltage and the SOH. Compared with the prior art, the relaxation voltage curve is fitted through the Bacon-Watts method, the inflection point voltage of the relaxation curve is extracted to serve as the health factor reflecting the SOH of the battery, and refined online rapid evaluation can be carried out on the health state of the battery.

Description

Technical Field

[0001] The present invention relates to the technical field of battery performance evaluation, and in particular to a battery health status evaluation method. Background Art

[0002] Batteries have been widely used in DC power supplies for substations and communication base stations, and in low-speed electric vehicles and electric bicycles due to their low cost, stable performance, no memory effect, mature large-capacity technology, strong environmental adaptability, and high recycling value. However, factors such as frequent deep charging and discharging, many charging times, long service life, and incorrect charging habits will lead to battery aging, which will affect the safe use of the battery and its remaining life. In order to ensure that the battery maintains a good working condition in the use environment, the battery's health state SOH (State of Health) must be quickly and accurately evaluated.

[0003] In the prior art, patent CN118376943A discloses a lithium-ion battery state of health (SOH) estimation method based on the characteristic value of the typical curve of Incremental Capacity (IC), including the steps of: for a given sample, carrying out a battery cycle aging test under the selected working temperature and charge and discharge rate conditions, recording the test data, and calculating the SOH value; extracting the typical characteristic value of the IC curve and determining the final characteristic value vector T; using the final characteristic value vector T as the input vector and the SOH value as the output parameter, constructing a Back Propagation Neural Network (BP) neural network model, and training it, and evaluating the battery's SOH value for the charge and discharge data under non-test conditions. This method relies on performing cycle aging tests at specific working temperatures and charge and discharge rates, which may limit the applicability of the model because the battery may operate in different environments and working conditions in actual use, thereby affecting the accuracy of the SOH estimation. In addition, extracting the typical characteristic values ​​of the IC curve may require more complex algorithms and higher professional knowledge, resulting in a more cumbersome implementation process, which may also affect the efficiency of real-time evaluation, especially in large-scale application scenarios.

[0004] Patent CN118311434A discloses a lithium-ion battery SOH estimation method and system based on electrochemical impedance spectroscopy. The electrochemical impedance spectroscopy data is interpreted and feature extracted through an improved relaxation time distribution algorithm, which enhances the robustness to noise and the ability to present abnormal data. The constructed neural network model adopts an improved joint loss convolutional neural network, and uses electrochemical impedance spectroscopy to perform end-to-end estimation of the SOH of the lithium-ion battery. The neural network model fuses the DRT features calculated based on the relaxation time Distribution of relaxation times (DRT) with the DRT features automatically extracted by the neural network model, avoiding the loss of feature information, effectively improving the performance of the neural network model, and improving the accuracy of the neural network model. Then, the SOH of the lithium-ion battery can be estimated quickly and accurately through the modified neural network model. Deep learning models are usually regarded as "black boxes", that is, although they can provide accurate predictions, their internal mechanisms and the importance of features may not be easy to explain. This may make it challenging for users to understand the SOH estimation results, especially when decisions need to be made. The specific data set used in the training of the model may affect its universality. Failure of the model to cover all possible battery types or usage conditions may result in distorted SOH estimates under new or different conditions.

[0005] CN107367692A discloses a method and device for evaluating the health status of a power battery, the method comprising: obtaining historical data of the power battery; calculating the capacity, internal resistance, charge cut-off cell pressure difference, and constant current charge cut-off cell pressure difference of the power battery based on the historical data; using the capacity to calculate the capacity health status of the power battery, using the internal resistance to calculate the internal resistance health status of the power battery, using the charge cut-off cell pressure difference to calculate the charge cut-off cell pressure difference health status of the power battery, and using the constant current charge cut-off cell pressure difference to calculate the constant current charge cut-off cell pressure difference health status of the power battery; determining the health status of the power battery based on the capacity health status, internal resistance health status, charge cut-off cell pressure difference health status, and constant current charge cut-off cell pressure difference health status. This method relies on the historical data of the power battery to evaluate the health status. If the historical data is incomplete or affected by noise, it may lead to inaccurate calculation results. In addition, the historical data may not fully reflect the actual status of the current battery, especially when the use environment and working conditions change. This method comprehensively evaluates the battery health status through multiple indicators (such as capacity, internal resistance, charge cut-off cell pressure difference, etc.). Although this multidimensional evaluation method can provide more comprehensive information, in actual operation, it may be relatively complicated to reasonably weigh the impact of various indicators and set appropriate weights.

[0006] In summary, existing research on online battery SOH evaluation generally has the problems of low accuracy and cumbersome analysis and processing. Summary of the invention

[0007] The purpose of the present invention is to provide a battery health status assessment method in order to overcome the defects of the above-mentioned prior art, which can perform refined online rapid assessment on the battery health status.

[0008] The purpose of the present invention can be achieved by the following technical solution: A battery health status assessment method comprises the following steps:

[0009] S1. Perform charge and discharge tests on the battery to obtain relaxation voltage data within a set time period;

[0010] S2, fitting the relaxation voltage data within a set time period to obtain a relaxation voltage change curve;

[0011] S3, based on the fitting parameters of the relaxation voltage change curve, solving to obtain the angle bisector;

[0012] S4. Solve the inflection point voltage based on the relaxation voltage data and the angle bisector;

[0013] S5, repeating steps S1 to S4 to obtain the inflection point voltages corresponding to the multiple batteries;

[0014] S6. Perform linear fitting on the inflection point voltage of each battery and the corresponding battery SOH value to obtain a linear relationship between the inflection point voltage and SOH;

[0015] S7. For the battery to be tested, execute steps S1 to S4 to obtain the inflection point voltage of the battery to be tested, and then find and determine the SOH value of the battery to be tested based on the linear relationship between the inflection point voltage and SOH.

[0016] Furthermore, the specific process of step S1 is as follows:

[0017] Perform discharge test on the battery at a set rate;

[0018] When the battery is discharged to the cut-off voltage, it is left to stand for a set time and the relaxation voltage data within the set time period is recorded;

[0019] After the discharge is completed, charge the battery to the cut-off voltage at the set charging rate;

[0020] The battery is charged at a constant voltage with a cut-off voltage. When the battery reaches the cut-off condition, the battery is stopped and allowed to stand for a set time. The relaxation voltage data within the set time period is recorded.

[0021] Furthermore, the set standing time is specifically 30 minutes.

[0022] Furthermore, the cut-off condition is specifically that the current charging current is less than or equal to the cut-off current.

[0023] Furthermore, the step S2 specifically adopts the least square method to fit the data of the early and late relaxation voltage to obtain the relaxation voltage change curve.

[0024] Furthermore, the fitting formula of the relaxation voltage change curve in step S2 is:

[0025]

[0026] where Z is a normally distributed random variable centered at zero representing the residual, and α 0 is the leftmost line segment intercept, k 1 The straight line L 1 The slope, k 2 The straight line L 2 The slope, L 1 and L 2 is the linear fitting of the relaxation voltage curve at the early and late stages, V k is the voltage inflection point, and γ is the inflection point transition mutation control quantity.

[0027] Furthermore, step S3 includes the following process:

[0028] Taking the minimum residual sum of squares as the optimization goal, the optimal fitting parameter α is obtained 0 , k 1 , k 2 ;

[0029] Based on the optimal fitting parameter α 0 , k 1 , k 2 , the angle bisector is solved as: y=k×x+b, where k is the slope of the angle bisector and b is the intercept of the angle bisector.

[0030] Furthermore, the residual sum of squares is specifically:

[0031] f = ∑(y i -k i × i -α 0 ) 2

[0032] The optimal fitting parameters are specifically:

[0033]

[0034] Where N is the number of relaxation voltage sampling points, y i is time x i The collected relaxation voltage.

[0035] Furthermore, the step S4 specifically inputs the relaxation voltage data and the solved angle bisector into Matlab, and uses the fzero function in Matlab to solve the intersection of the straight line and the curve, that is, to solve the inflection point voltage value.

[0036] Furthermore, the step S6 specifically inputs the inflection point voltage value and the corresponding SOH value as variables X and Y respectively into the Origin software, and obtains the linear relationship between the inflection point voltage and SOH through linear fitting of the two variables.

[0037] Compared with the prior art, the present invention has the following advantages:

[0038] The present invention determines the inflection point of the relaxation voltage curve through Bacon-Watts: first, the battery is charged and discharged, and the relaxation voltage change curve is obtained by collecting the relaxation voltage data within a set time period and fitting, and then the inflection point voltage is obtained by solving the angle bisector; the inflection point voltage of each battery and the corresponding battery SOH value are linearly fitted to obtain the linear relationship between the inflection point voltage and SOH. Therefore, for the battery to be tested, as long as the relaxation voltage data of the battery at the same charge and discharge rate is collected, the inflection point voltage size is obtained through the model established by the Bacon-Watts method, and then according to the linear relationship between the inflection point voltage and SOH, the battery SOH value corresponding to the inflection point voltage can be found, thereby realizing the online refined evaluation of the battery health status SOH. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a schematic diagram of the method flow of the present invention;

[0040] Figure 2a This is the SOH distribution diagram of 80 400Ah lead-acid batteries in the embodiment;

[0041] Figure 2b This is the SOH distribution diagram of 20 300Ah lead-acid batteries in the embodiment;

[0042] Figure 3 Schematic diagram of the inflection point voltage in the embodiment;

[0043] Figure 4a is the relaxation inflection point voltage V of 80 400Ah lead-acid batteries in the embodiment k Distribution map;

[0044] Figure 4b is the relaxation inflection point voltage V of 20 300Ah lead-acid batteries in the embodiment k Distribution map;

[0045] Figure 5a The figure is a linear fitting diagram of the inflection point voltage and SOH of 80 400Ah lead-acid batteries in the embodiment;

[0046] Figure 5b The figure is a linear fitting diagram of the inflection point voltage and SOH of 20 300Ah lead-acid batteries in the embodiment. DETAILED DESCRIPTION

[0047] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0048] Example

[0049] like Figure 1 As shown, a battery health status assessment method includes the following steps:

[0050] S1. Perform charge and discharge tests on the battery to obtain relaxation voltage data within a set time period;

[0051] S2, fitting the relaxation voltage data within a set time period to obtain a relaxation voltage change curve;

[0052] S3, based on the fitting parameters of the relaxation voltage change curve, solving to obtain the angle bisector;

[0053] S4. Solve the inflection point voltage based on the relaxation voltage data and the angle bisector;

[0054] S5, repeating steps S1 to S4 to obtain the inflection point voltages corresponding to the multiple batteries;

[0055] S6. Perform linear fitting on the inflection point voltage of each battery and the corresponding battery SOH value to obtain a linear relationship between the inflection point voltage and SOH;

[0056] S7. For the battery to be tested, execute steps S1 to S4 to obtain the inflection point voltage of the battery to be tested, and then find and determine the SOH value of the battery to be tested based on the linear relationship between the inflection point voltage and SOH.

[0057] This embodiment applies the above solution, and the main process includes:

[0058] (1) Perform a charge and discharge test on the battery. The test steps are as follows:

[0059] a. Perform discharge test on the battery at a certain rate;

[0060] b. Discharge to the cut-off voltage. When the battery reaches the cut-off voltage, let it stand for 30 minutes and record the voltage, current and other data;

[0061] c. After the rest is completed, charge the battery to the cut-off voltage at the same charging rate;

[0062] d. Charging at a constant voltage with a cut-off voltage until the cut-off condition is that the current is less than or equal to the cut-off current;

[0063] e. After charging is completed, let it stand for 30 minutes and record the voltage, current and other data.

[0064] In this embodiment, 80 400Ah and 20 300Ah lead-acid batteries are tested at an ambient temperature of 25°C. The SOH values ​​of the 80 400Ah and 20 300Ah lead-acid batteries are as follows: Figure 2a and Figure 2b shown.

[0065] First, perform a 1 / 10C discharge test on the lead-acid battery;

[0066] After that, the discharge cut-off voltage is 1.75V. When the lead-acid battery reaches the cut-off voltage, it is left to stand for 30 minutes and the voltage and current data are recorded.

[0067] After the rest is complete, charge the battery at a charging rate of 1 / 10C until the voltage reaches 2.35V;

[0068] Then charge at a constant voltage of 2.35V until the current is less than or equal to 4A;

[0069] After charging is completed, let it stand for 30 minutes and record the voltage, current and other data.

[0070] (2) The least squares method is used to fit the data of the early and late relaxation voltage to obtain a curve fitting the change of relaxation voltage:

[0071]

[0072] where Z is a normally distributed random variable centered at zero representing the residual; α 0 is the leftmost line segment (at x=V k A type of intercept at 1 and k 2 Adjust the slope of the two intersecting lines, where k 1 The straight line L 1 The slope of k 2 The straight line L 2 The slope of the transition is γ, while γ controls the abruptness of the transition. In this example, γ is set to 10 -5 , to obtain the k The parameter α is learned using the least squares method. 0 , k 1 , k 2 and V k , to fit the relaxation voltage data.

[0073] (3) When a series of relaxation voltage data is given, time x i , voltage y iSince the inflection point is determined by linear fitting, we can use y=k i ×x+α 0 , the task of this scheme is to find the optimal parameter k i (i=1,2) and α 0 Make the relaxation voltage data as close to the linear equation as possible.

[0074] (4) When the fitting effect is the best, it is obvious that the residual sum of squares is minimized, and the problem is transformed into finding the minimum residual sum of squares.

[0075] f = ∑(y i -k i × i -α 0 ) 2

[0076] (5) To find the optimal slope and intercept, we only need to find the partial derivative of the residual sum of squares, and finally k i With α 0 The expression is:

[0077]

[0078] Where N is the number of relaxation voltage sampling points.

[0079] (6) Figure 3 As shown, two fitting straight lines L are solved. 1 and L 2 The intersection point (x 0 ,y 0 ), let the angle bisector line L 3 It is y=k×x+b, according to the definition of angle bisector:

[0080] |kk 1 | / (1+k 1 ×k)=|k 2 -k| / (1+k×k 2 )

[0081] Solve for the slope k of the angle bisector and connect the two straight lines L 1 and L 2 Substitute the intersection point into the equation y=k×x+b, and then get the intercept b;

[0082] (7) Input the relaxation voltage data and the solved angle bisector into Matlab, and use the fzero function in Matlab to solve the intersection of the straight line and the curve, and then solve the inflection point voltage value V k The relaxation inflection point voltage V of 80 400Ah lead-acid batteries and 20 300Ah lead-acid batteries in this embodiment is k The distribution is as follows Figure 4a and Figure 4b As shown;

[0083] (8) The inflection point voltage value and the battery health status value are input into the Origin software as variables X and Y, respectively, and a linear fit is performed on the two variables (the linear fitting results of the inflection point voltage and SOH of 80 400Ah lead-acid batteries and 20 300Ah lead-acid batteries in this embodiment are shown in Figure 2). Figure 5a and Figure 5b As shown in Figure 2 ). Once the linear relationship between the inflection point voltage and SOH is obtained, the SOH of the battery can be evaluated through the linear relationship by simply knowing the inflection point voltage of any relaxation curve.

[0084] In summary, this solution determines the inflection point voltage of the relaxation curve based on the Bacon-Watts method, and uses the inflection point voltage as a health factor to achieve online evaluation of the battery health status. Among them, the battery relaxation voltage can be directly collected in real time according to the battery management system, without the need for additional collection, and will not increase the workload.

[0085] This solution uses Bacon-Watts to determine the inflection point of the relaxation voltage curve, that is, the voltage magnitude of the critical point where the relaxation voltage changes from instantaneous change to slow change, and the inflection point voltage value is linearly negatively correlated with the SOH value. The Bacon-Watts method is simple and convenient for processing battery relaxation voltage data, and the acquisition of the inflection point voltage is also very simple and convenient. Once a linear fitting model of the inflection point voltage and SOH is established, for any battery to be tested with unknown SOH, it is only necessary to collect the relaxation voltage data of the battery at the same charge and discharge rate, and then obtain the inflection point voltage magnitude through the model established by the Bacon-Watts method. Based on the linear relationship between the inflection point voltage and SOH, the battery SOH value corresponding to the inflection point voltage can be found, thereby realizing the online evaluation of the battery health status SOH. This solution is easy to implement in engineering and can ensure the accuracy of the results.

Claims

1. A battery health status assessment method, characterized in that: The following steps are involved: S1. Perform charge and discharge tests on the battery to obtain relaxation voltage data within a set time period; S2, fitting the relaxation voltage data within a set time period to obtain a relaxation voltage change curve; S3, based on the fitting parameters of the relaxation voltage change curve, solving to obtain the angle bisector; S4. Solve the inflection point voltage based on the relaxation voltage data and the angle bisector; S5, repeating steps S1 to S4 to obtain the inflection point voltages corresponding to the multiple batteries; S6. Perform linear fitting on the inflection point voltage of each battery and the corresponding battery SOH value to obtain a linear relationship between the inflection point voltage and SOH; S7. For the battery to be tested, execute steps S1 to S4 to obtain the inflection point voltage of the battery to be tested, and then find and determine the SOH value of the battery to be tested based on the linear relationship between the inflection point voltage and SOH.

2. A battery health status assessment method according to claim 1, characterized in that: The specific process of step S1 is as follows: Perform discharge test on the battery at a set rate; When the battery is discharged to the cut-off voltage, it is left to stand for a set time and the relaxation voltage data within the set time period is recorded; After the discharge is completed, charge the battery to the cut-off voltage at the set charging rate; The battery is charged at a constant voltage with a cut-off voltage. When the battery reaches the cut-off condition, the battery is stopped and allowed to stand for a set time. The relaxation voltage data within the set time period is recorded.

3. A battery health status assessment method according to claim 2, characterized in that: The static setting time is specifically 30 minutes.

4. A battery health status assessment method according to claim 2, characterized in that: The cut-off condition is specifically that the current charging current is less than or equal to the cut-off current.

5. A battery health status assessment method according to claim 1, characterized in that: The step S2 specifically adopts the least square method to fit the data of the early and late relaxation voltage to obtain the relaxation voltage change curve.

6. A battery health status assessment method according to claim 5, characterized in that: The fitting formula of the relaxation voltage change curve in step S2 is: Where Z is a normal distribution random variable centered at zero representing the residual, α0 is the intercept of the leftmost line segment, k1 is the slope of line L1, k2 is the slope of line L2, L1 and L2 are the linear fitting relaxation voltage curve early and late change lines, V k is the voltage inflection point, and γ is the inflection point transition mutation control quantity.

7. A battery health status assessment method according to claim 6, characterized in that: Step S3 The process includes: Taking the minimum residual sum of squares as the optimization goal, the optimal fitting parameters α0, k1, k2 are obtained; Based on the optimal fitting parameters α0, k1, and k2, the angle bisector is solved to be: y=k×x+b, where k is the slope of the angle bisector and b is the intercept of the angle bisector.

8. A battery health status assessment method according to claim 7, characterized in that: The residual sum of squares is specifically: f=∑(y i -k i ×x i -α0) 2 The optimal fitting parameters are specifically: Where N is the number of relaxation voltage sampling points, y i is time x i The collected relaxation voltage.

9. A battery health status assessment method according to claim 1, characterized in that: The step S4 specifically inputs the relaxation voltage data and the angle bisector to be solved into Matlab, and uses the fzero function in Matlab to solve the intersection of the straight line and the curve, that is, to solve the inflection point voltage value.

10. A battery health status assessment method according to claim 1, characterized in that: The step S6 specifically inputs the inflection point voltage value and the corresponding SOH value as variables X and Y respectively into the Origin software, and obtains the linear relationship between the inflection point voltage and SOH through linear fitting of the two variables.

Images