An Online Detection Method for Battery Health State Based on Regional Straight-Line Distance

Abstract

The present invention relates to an online detection method for the state of health of a battery based on the regional straight-line distance, including: obtaining voltage data of a battery with a known SOH during charging at a set rate, and establishing a Q-V curve; interpolating the Q-V curve to draw an IC curve and performing smoothing processing; selecting the highest peak value from the IC curve and recording the voltage peak value V peak ; thereby determining the regional straight-line distance L from the Q-V curve AB ; selecting batteries with different SOHs, repeating the above process, and obtaining the regional straight-line distances L corresponding to the batteries with different SOHs AB ; using a linear fitting method to obtain the L AB -SOH fitting equation; charging the battery to be measured at a set rate to obtain the voltage data to be measured, establishing a Q-V curve of the battery to be measured, and obtaining the regional straight-line distance L of the battery to be measured according to the above steps AB‑t ; substituting the regional straight-line distance L of the battery to be measured AB‑t into the L AB -SOH fitting equation to obtain the SOH value of the battery to be measured. Compared with the prior art, the present invention can simply, efficiently and accurately know the current state of health of the battery.

Description

Technical Field

[0001] The present invention relates to the technical field of battery detection, and in particular to an online detection method for the state of health of a battery based on the regional straight-line distance. Background Art

[0002] Lithium-ion batteries play an important role in many applications such as electrified transportation and smart grid batteries. However, the performance of lithium batteries will decline over time. If severely aged lithium-ion batteries are not replaced in time, battery degradation will affect the normal operation of the battery system and even lead to safety accidents such as explosions. The assessment of the state of health (SOH) of lithium-ion batteries can provide important references for functions such as battery safety protection, charge and discharge control, and thermal management.

[0003] Currently, the estimation methods for battery SOH can be roughly divided into model-based estimation methods, data-driven estimation methods, and direct measurement methods. Regardless of the method, it is crucial to extract appropriate health factors for battery SOH. The use of the battery is one of the main reasons for battery degradation. Therefore, it has become increasingly important to extract and analyze the health state of the battery from on-site measurement data. If a battery management system (BMS) is used for calculation, the model must be simplified.

[0004] For example, Chinese Patent Application CN115144758A discloses a method for evaluating the state of health of a battery based on a convolutional neural network. Starting from the characteristics of the battery itself, according to the battery IC curve, the voltage data in a specific voltage range under the full charge state is extracted as the input to train and optimize the evaluation model as the basic data. Although the battery SOH evaluation can be completed using only one model and one objective function, this method has a large amount of calculation, high requirements for the training set, is not suitable for the SOH evaluation of small-sample batteries, and is not applicable to the online evaluation of SOH due to the limited storage space of the BMS.

[0005] Chinese Patent Application CN113138340A discloses an estimation method for the state of health of a battery based on a model. The method includes the following steps: I. Obtain the electrochemical impedance spectrum of the battery under preset state parameters (preset temperature parameter and preset state of charge parameter). II. Calculate the relaxation time distribution spectrum of the battery according to the electrochemical impedance spectrum; III. Establish an equivalent circuit model of the battery according to the relaxation time distribution spectrum. This method usually requires high computing power and a large amount of work to parameterize the battery model, and the evaluation model has not been well simplified.

[0006] In addition, Chinese Patent Application CN111308378A discloses a method for evaluating the state of health (SOH) of a battery pack based on incremental capacity analysis (ICA). It determines the IC curve of the battery pack according to the charging current and terminal voltage of the battery pack, filters the capacity increment curve of the battery pack using a moving average filtering algorithm, selects the maximum curve value of the capacity increment curvature, and determines the SOH of the battery pack using a regularization method. However, ICA is easily affected by subjective parameters such as cycling conditions, charging rate, sampling frequency, noise components, and filtering, which can cause changes in the IC curve, making the method for evaluating the SOH of the battery based on the IC curve may be inaccurate. Summary of the Invention

[0007] The purpose of the present invention is to overcome the defects of the above-mentioned existing technologies and provide an online detection method for the state of health of a battery based on the regional straight-line distance, which can simply, efficiently, and accurately obtain the current state of health of a lithium battery.

[0008] The purpose of the present invention can be achieved by the following technical solutions: An online detection method for the state of health of a battery based on the regional straight-line distance, comprising the following steps:

[0009] S1. Obtain the voltage data of a battery with a known SOH during charging at a set rate to establish a corresponding Q-V (capacity-voltage) curve;

[0010] S2. Perform interpolation processing on the Q-V curve and then plot an IC (incremental capacity) curve, and smooth the IC curve;

[0011] S3. Select the highest peak value from the smoothed IC curve and record the corresponding voltage peak value V peak ;

[0012] S4. According to the voltage peak value V peak , determine the regional straight-line distance L from the Q-V curve AB ;

[0013] S5. Select batteries with different SOHs and repeat steps S1 to S4 to obtain the regional straight-line distances L corresponding to the batteries with different SOHs AB ;

[0014] S6. Adopt a linear fitting method to obtain an L AB -SOH fitting equation;

[0015] S7. Charge the battery to be tested at a set rate to obtain the voltage data to be tested, and establish a Q-V curve of the battery to be tested;

[0016] S8. For the Q-V curve of the battery under test, perform steps S2 to S4 to obtain the regional straight-line distance L of the battery under test. AB-t ;

[0017] S9. Substitute the regional straight-line distance L of the battery under test AB-t into the L AB -SOH fitting equation to obtain the SOH value of the battery under test.

[0018] Further, step S1 specifically includes the following steps:

[0019] S11. Calibrate the actual capacity of the battery;

[0020] S12. Determine the SOH value of the battery according to the actual capacity and the rated capacity of the battery;

[0021] S13. Charge the battery at a set rate and obtain the corresponding voltage data to establish the corresponding Q-V curve.

[0022] Further, step S11 specifically includes the following steps:

[0023] S111. Charge the battery under constant current and constant voltage conditions until the set first voltage threshold is reached;

[0024] S112. After the battery is left standing for a set time, discharge the battery under the same constant current condition as in step S111 until the set second voltage threshold is reached;

[0025] S113. After the battery is left standing for a set time, measure the discharge capacity of the battery, which is the actual capacity of the battery.

[0026] Further, the SOH value of the battery in step S12 is specifically:

[0027]

[0028] Further, the first voltage threshold is specifically the charging upper limit voltage of the battery, and the second voltage threshold is specifically the discharging lower limit voltage of the battery.

[0029] Further, step S4 specifically includes the following steps:

[0030] S41. Find the point with voltage V peak in the Q-V curve and denote it as A;

[0031] S42. According to the set regional voltage range ΔV, starting from point A, determine the voltage value V B of the end point B;

[0032] S43. Find the point with voltage V in the Q-V curveB If the end point is B, then the straight line segment AB is the regional straight line, and the straight-line distance between point A and point B is the regional straight-line distance L AB .

[0033] Furthermore, the voltage value of the end point B in the step S42 is specifically: V B =V peak +ΔV.

[0034] Furthermore, the regional straight-line distance in the step S43 is specifically:

[0035]

[0036] wherein, L AB is the regional straight-line distance within the regional voltage range ΔV, x A and y A are the abscissa and ordinate of the starting point A respectively, and x B and y B are the abscissa and ordinate of the end point B respectively.

[0037] Furthermore, the step S6 specifically uses the regional straight-line distance L AB as the independent variable and the battery SOH as the dependent variable to establish the L AB -SOH linear regression equation.

[0038] Furthermore, the establishment of the Q-V curve, the interpolation processing of the Q-V curve, the plotting of the IC curve, and the selection of the highest peak value from the IC curve are all realized by operating the Origin software.

[0039] Compared with the prior art, based on the existing ICA method, the present invention introduces the concept of regional straight-line distance into the original charging curve and extracts the regional side length L AB as a health factor to detect and evaluate the SOH of the battery. The method of the present invention has a simple process and does not require a large amount of calculation. Once the starting voltage of the regional straight-line segment is determined, the distance of the regional straight-line segment can be determined in the original charging curve; in addition, the working data during the battery charging process can be collected in real time by the existing battery management system, without the need for offline experiments on lithium batteries or additional data collection, which can greatly reduce the workload. The method of the present invention is not sensitive to the sampling frequency, can still obtain high accuracy at low sampling frequencies, ensure the accuracy of detecting and evaluating the SOH, and has simple calculation and is applicable to the real-time online SOH detection and evaluation of the battery system. Description of the Drawings

[0040] Figure 1 is a schematic flow chart of the method of the present invention;

[0041] Figure 2 Q-V curves of nickel-cobalt-aluminum lithium-ion (NCA) single cells at different SOH values in the embodiments;

[0042] Figure 3 Battery charging IC curves with a sampling point every three minutes at different SOH values in the embodiments;

[0043] Figure 4 The straight-line segment AB in the Q-V curve of the nickel-cobalt-aluminum lithium-ion (NCA) single cell at different SOH values in the embodiments (taking ΔV = 100 mV as an example);

[0044] Figure 5 L-SOH fitting curve graphs at different regional voltages in the embodiments AB -SOH fitting curve graphs. Detailed implementation manners

[0045] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0046] Embodiment

[0047] As Figure 1 shown, an online battery health state detection method based on regional straight-line distance includes the following steps:

[0048] S1. Obtain voltage data of a battery with a known SOH during charging at a set rate to establish a corresponding Q-V (capacity-voltage) curve. Before obtaining the voltage data, it is necessary to calibrate the actual capacity of the battery to determine the SOH of the battery. Specifically:

[0049] First, charge the battery under constant current and constant voltage conditions until the charging upper limit of the battery is reached;

[0050] After the battery is left standing for a period of time, then discharge the battery under the same constant current condition until the discharge lower limit of the battery is reached;

[0051] After leaving the battery standing for a period of time, measure the discharge capacity of the battery, and take the discharge capacity as the actual capacity of the battery;

[0052] Calculate the SOH of the battery:

[0053]

[0054] S2. Perform interpolation processing on the Q-V curve and then plot an IC (incremental capacity) curve, that is, a ΔQ / ΔV-V curve, and smooth the IC curve;

[0055] S3. Select the highest peak value from the smoothed IC curve and record the corresponding voltage peak value V peak ;

[0056] S4. Determine the regional straight-line distance L from the Q-V curve according to the voltage peak V peak , specifically: AB Specifically:

[0057] First, find the point with voltage V in the Q-V curve, denoted as A. The abscissa and ordinate of point A are x peak and y A respectively. Then, according to the set regional voltage range ΔV, taking point A as the starting point, the voltage of the end point B is V A = V B + ΔV; peak + ΔV;

[0058] Mark the point on the Q-V curve with abscissa V B as point B. Point B is the end point, and the abscissa and ordinate of point B are x B and y B respectively;

[0059] Then the straight-line segment AB is the regional straight line, and calculate the straight-line distance L between point A and point B AB :

[0060]

[0061] where L AB is the regional straight-line distance within the regional voltage range ΔV;

[0062] S5. Select batteries with different SOHs, repeat steps S1 to S4, and obtain the regional straight-line distances L corresponding to the batteries with different SOHs AB ;

[0063] S6. Adopt a linear fitting method to obtain the L AB -SOH fitting equation, where the regional straight-line distance L AB is the independent variable and the battery SOH is the dependent variable;

[0064] S7. Charge the battery under test at a set rate to obtain the voltage data under test and establish the Q-V curve of the battery under test;

[0065] S8. For the Q-V curve of the battery under test, execute steps S2 to S4 to obtain the regional straight-line distance L of the battery under test AB-t ;

[0066] S9. Substitute the regional straight-line distance L of the battery under test AB-t into the L AB -SOH fitting equation to obtain the SOH value of the battery under test.

[0067] In practical applications, establishing the Q-V curve, interpolating the Q-V curve, performing incremental capacity analysis, and peak seeking can all be achieved using Origin software operations.

[0068] In this embodiment, 21700 ternary lithium-ion batteries with different SOHs retired from a certain brand of electric vehicle are detected. Among them, the rated capacity of a brand-new 21700 ternary battery is 4.8 Ah, and its SOH value is defined as 100%. The batteries are cycled and aged at a 1 / 2C rate, and the battery data sampling frequency is one sample every 3 minutes. Table 1 shows the change of the battery SOH with the number of cycles in this embodiment. An online evaluation method for the state of health of the battery based on the regional straight-line distance is used to model the SOH of the 21700 ternary lithium-ion battery.

[0069] Table 1 shows the change of the battery SOH with the number of cycles in this embodiment

[0070] Number of cycles 50 100 150 200 250 SOH / % 71.88 70.21 67.92 65.42 63.54

[0071] The main processes include:

[0072] First, obtain the charging voltage data of the 21700 ternary lithium-ion battery at a 1 / 2C rate, and establish a battery capacity-voltage curve (Q-V curve) based on the voltage data and the actual capacity (as Figure 2 shown).

[0073] Second, use Origin software to interpolate the Q-V curve, differentiate the interpolated Q-V curve, and obtain the (ΔQ / ΔV)-V curve at different SOH values and smooth it to obtain the IC smooth curve (as Figure 3 shown), and automatically seek the peak to obtain the voltage V corresponding to its highest peak peak . The point corresponding to V on the Q-V curve is denoted as point A, and its coordinate values are (x peak , y A , y A ).

[0074] Third, the regional voltages ΔV are set to 25 mV, 50 mV, 75 mV, and 100 mV respectively; starting from point A, take V B = V peak + ΔV. The point on the Q-V curve with the abscissa of V B is denoted as point B, and point B is the end point, and the coordinates of point B are (x B , y B ); calculate the regional straight-line distance L AB of the straight-line segment AB under different regional voltages (as shown in Table 2 and Figure 4 ).

[0075] Table 2

[0076] Coordinates of point A Coordinates of point B <![CDATA[L AB > The 50th cycle (0.894，3.758) (1.306，3.858) 0.424 The 100th cycle (0.708，3.729) (1.104，3.829) 0.408 The 150th cycle (0.701，3.744) (1.089，3.844) 0.401 The 200th cycle (0.688，3.765) (1.064，3.865) 0.389 The 250th cycle (0.682，3.776) (1.045，3.876) 0.377

[0077] IV. According to the L AB value obtained in Step III, establish an L AB -SOH model by linear fitting. Taking the SOH value of the battery as the ordinate and the regional straight-line distance L AB as the abscissa, obtain the L AB -SOH fitting curve (as Figure 5 shown). The fitting results are shown in Table 3. It can be seen that the fitting degree of this model can reach 0.99, with high precision.

[0078] Table 3

[0079] Regional voltage ΔV / mV <![CDATA[L AB -SOH model]]> <![CDATA[Fitting degree R 2 > 25 <![CDATA[SOH = 19.11 + 468.48L AB > 0.9427 50 <![CDATA[SOH = 14.91 + 261.13L AB > 0.9959 75 <![CDATA[SOH=6.91+200.46L AB > 0.9910 100 <![CDATA[SOH=-5.81+184.17L AB > 0.9803

[0080] As Figure 5 shown, there is a good linear positive correlation between the regional straight-line distance L AB and SOH. Therefore, the regional straight-line distance L AB can be extracted from the charging curve as the health factor of the battery SOH, and the SOH value of the battery to be tested can be quickly detected through their positive correlation.

[0081] According to the online detection method of battery health status based on the regional straight-line distance L AB involved in this embodiment, on the basis of ICA, the concept of regional straight-line distance is introduced into the original charging curve, and the regional side length L AB is extracted as the health factor to evaluate the SOH evaluation of the battery. Using the charging voltage data of NCA single cells at a 1 / 2C rate, an SOH model at a sampling frequency of 1 / 180Hz is established. The results show that SOH and L AB are linearly positively correlated, and the R square (R 2 ) of linear fitting can reach more than 0.99. In the narrow voltage range with a sampling frequency as low as one point every 3 minutes, when SOH modeling is adopted, the proposed health factor still has a high R 2 , which provides the possibility for the online evaluation of battery SOH in engineering. The sampling frequency of this embodiment is 1 / 180Hz. The lower the sampling frequency, the fewer data points that can be collected during the charging test, corresponding to less computational effort, and a lower sampling frequency does not require expensive test units, and the price of the BMS is also lower. While the higher the sampling frequency, the finer the data and the more accurate the estimation model. However, the results of this embodiment show that the regional straight-line distance method is very effective for SOH modeling under low sampling frequencies and small regional voltage ranges, which is very helpful for the engineering application of online battery SOH evaluation, because the sampling frequency of battery data is not necessarily high, and the voltage window of battery operation is not necessarily wide.

[0082] In summary, in this technical solution, the lithium-ion battery is first charged and discharged at a certain rate, the battery charging voltage data is collected, and the Q-V curve is obtained; then the charging voltage data is converted into an IC curve, and the voltage V corresponding to the maximum peak of the IC curve is searched peak ; then, in the Q-V curve, the point A corresponding to V peak is selected as the starting point, a suitable regional voltage range ΔV is selected, and V B = V peak + ΔV. The point on the Q-V line with the abscissa of V B is denoted as point B, and point B is the end point, thus forming the regional straight line L AB ; then, using the distance formula between two points, the distance L AB of the regional straight line is obtained; with the distance L AB of the regional straight line as the independent variable and the battery SOH as the dependent variable, a linear regression equation of L AB - SOH is established; finally, another lithium-ion sample battery to be measured is selected, and the above steps are repeated to collect the charging voltage data of the lithium-ion sample battery to be measured at the same rate, and the distance L AB of the regional straight line corresponding to the regional voltage ΔV of the lithium-ion sample battery to be measured is obtained. The distance L AB of the regional straight line of the battery to be measured is used as a health factor index and substituted into the linear regression equation to calculate the SOH value of the battery to be measured corresponding to the distance L AB of the regional straight line, thereby realizing the evaluation of the battery SOH of the lithium-ion sample battery to be measured.

[0083] The battery health state detection method based on the regional straight line distance proposed by the present invention is simple to implement and does not require a large amount of calculation. Once the starting voltage of the regional straight line segment is determined, the distance of the regional straight line segment can be determined in the original charging curve. The working data during the battery charging process can be collected in real time by the battery management system, without the need for offline experiments on lithium batteries and without additional data collection, reducing the workload. And this method is not sensitive to the sampling frequency and can still obtain high accuracy at a low sampling frequency (such as 1 / 180 Hz, that is, sampling once every 3 minutes), and the calculation is simple, which is suitable for real-time online SOH evaluation of battery systems. Therefore, it has more advantages for the BMS of large-scale energy storage power stations with limited computing and storage capabilities, and this method is easier to implement.

Claims

1. An online detection method for the state of health of a battery based on the regional straight-line distance, characterized in that, It includes the following steps: S1. Obtain the voltage data of the battery with known SOH during charging at a set rate to establish the corresponding Q-V curve, i.e., the capacity-voltage curve; S2. After interpolating the Q-V curve, plot the IC curve, i.e., the ΔQ / ΔV—V curve, and smooth the IC curve; S3. Select the highest peak value from the smoothed IC curve and record the corresponding voltage peak value V peak ; S4. Determine the regional straight-line distance L from the Q-V curve according to the voltage peak value V peak ; AB ; S5. Select batteries with different SOH, and repeat steps S1 to S4 to obtain the regional straight-line distance L corresponding to the batteries with different SOH AB ; S6. Obtain L by using the linear fitting method AB -SOH fitting equation S7. Charge the battery to be measured at a set rate to obtain the voltage data to be measured and establish the Q-V curve of the battery to be measured; S8. For the Q-V curve of the battery under test, perform steps S2 to S4 to obtain the regional straight-line distance L of the battery under test AB-t ; S9. Substitute the linear distance L of the area of the battery under test AB-t into L AB -SOH fitting equation to obtain the SOH value of the battery under test; Step S4 specifically includes the following steps: S41. Find the point with voltage V in the Q-V curve, denoted as A; peak ​ S42. Determine the voltage value V of the end point B starting from point A according to the set regional voltage range ΔV. B ; S43. Find the end point B where the voltage is V in the Q-V curve B Then the straight line segment AB is the regional straight line, and the straight-line distance between point A and point B is the regional straight-line distance L AB ; The voltage value of the end point B in step S42 is specifically: V B = V peak + ΔV.

2. An online detection method for the state of health of a battery based on the regional straight-line distance according to claim 1, characterized in that, The specific steps of step S1 are as follows: S11. Calibrate the actual capacity of the battery; S12. Determine the SOH value of the battery according to the actual capacity and the rated capacity of the battery; S13. Charge the battery at a set rate and obtain the corresponding voltage data to establish the corresponding Q-V curve.

3. The online detection method for battery health state based on regional straight-line distance according to claim 2, wherein, The specific steps of step S11 are as follows: S111. Charge the battery under constant current and constant voltage conditions until the set first voltage threshold is reached; S112. After the battery stands for a set time, discharge the battery under the same constant current condition as in step S111 until the set second voltage threshold is reached; S113. After the battery stands for a set time, measure the discharge capacity of the battery, which is the actual capacity of the battery.

4. An online detection method for the state of health of a battery based on the regional straight-line distance according to claim 2, characterized in that, The SOH value of the battery in step S12 is specifically:

5. The online detection method for the battery health state based on the regional straight-line distance according to claim 3, characterized in that, The first voltage threshold is specifically the upper charging voltage of the battery, and the second voltage threshold is specifically the lower discharging voltage of the battery.

6. The on-line detection method for the battery health state based on the regional straight-line distance according to claim 1, wherein The specific regional straight-line distance in step S43 is: Wherein, L AB is the regional straight-line distance within the regional voltage range ΔV, x A and y A are the abscissa and ordinate of the starting point A respectively, x B and y B are the abscissa and ordinate of the ending point B respectively.

7. An online detection method for battery health status based on regional straight-line distance according to claim 1, characterized in that Specifically, in step S6, the regional straight-line distance L AB is used as the independent variable, and the battery SOH is used as the dependent variable to establish an L AB -SOH linear regression equation.

8. An online detection method for the battery health state based on the regional straight-line distance according to claim 1, characterized in that The establishment of the Q-V curve, the interpolation processing of the Q-V curve, the plotting of the IC curve, and the selection of the highest peak value from the IC curve are all realized by operating the Origin software.

Images