Battery health state prediction method and device, detection assembly and vehicle

Abstract

The invention discloses a battery health state prediction method and device, a detection assembly and a vehicle, and belongs to the technical field of new energy. The method comprises the following steps: acquiring a plurality of time-varying expansive force values and residual electric quantity of a to-be-detected battery in a single charging process; and under the condition that the difference between the second residual electric quantity at the end of charging and the first residual electric quantity at the start of charging is greater than the preset charging depth, determining at least two expansive force extreme values and the residual electric quantity corresponding to the sampling time of each expansive force extreme value. The expansion force extreme value corresponds to a key feature point in a single charging process, and the health degree of the battery can be reflected by combining the synchronous residual electric quantity. According to the method, the at least two expansive force extreme values and the corresponding residual electric quantities are input into the preset prediction model, the correlation between the independent variable (expansive force extreme value) and the dependent variable (health state information) is subjected to regression analysis through the preset prediction model, the health state information is predicted, and the accuracy of the health state of the battery is improved.

Description

Technical Field

[0001] The present application relates to the field of new energy technology, and in particular to a method, device, detection component and vehicle for predicting the health status of a battery. Background Art

[0002] With the rapid development of the domestic new energy industry, the use of new energy vehicles and electrochemical energy storage systems has increased significantly year by year, and new energy vehicles have been widely used in people's daily lives. New energy vehicles and energy storage systems usually use rechargeable batteries as power sources. During use, the state of the rechargeable battery needs to be tested to determine its performance and reliability. Among them, the state of health (SOH) of the rechargeable battery is one of the important indicators to characterize the battery state.

[0003] In the related art, the SOH estimation method usually uses the driving mileage or the actual battery capacity to predict the SOH. This judgment method is affected by the actual operating environment of the rechargeable battery and it is difficult to obtain an accurate value. Summary of the invention

[0004] The present application provides a method, device, detection component and vehicle for predicting the health status of a battery, which can improve the accuracy of the health status of the battery. The technical solution is as follows:

[0005] In a first aspect, a method for predicting a battery health state is provided, the method comprising: obtaining a plurality of expansion force values ​​and a plurality of remaining capacities that vary with time during a single charging process of a battery to be detected, and obtaining a first remaining capacities at the start of charging and a second remaining capacities at the end of charging during the single charging process; when the difference between the second remaining capacities and the first remaining capacities is greater than a preset depth of charge, determining at least two expansion force extremes from the plurality of expansion force values ​​that vary with time, and determining the remaining capacities corresponding to the sampling moments of each of the at least two expansion force extremes from the plurality of remaining capacities that vary with time; inputting the at least two expansion force extremes and the remaining capacities corresponding to the expansion force extremes into a preset prediction model, and predicting the health state information of the battery to be detected through the preset prediction model; wherein the preset prediction model is used to learn the correlation between independent variables and dependent variables, the independent variables comprising expansion force parameters, and the dependent variables comprising a battery discharge capacity that is positively correlated with the health state information.

[0006] In a second aspect, a device for predicting a battery health status is provided, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program implements the method described in the first aspect when executed by the processor.

[0007] According to a third aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0008] In a fourth aspect, a computer program product comprising instructions is provided, which, when executed on a computer, enables the computer to execute the method described in the first aspect.

[0009] In a fifth aspect, a detection component is provided, which includes a pressure sensor, a power sensor, and a battery health status prediction device as described in the second aspect above; the pressure sensor is clamped between two adjacent batteries or between a battery and a baffle; the pressure sensor is used to detect multiple expansion force values ​​that change with time during a single charging process of the battery to be detected; the power sensor is used to detect multiple residual powers of the battery to be detected that change with time during the single charging process; the battery health status prediction device is used to implement the method described in the first aspect above according to the multiple expansion force values ​​that change with time and the multiple residual powers that change with time.

[0010] In a sixth aspect, a vehicle is provided, the vehicle comprising the detection component described in the fifth aspect.

[0011] In a seventh aspect, a storage system is provided, wherein the energy storage system comprises the detection component described in the fifth aspect.

[0012] The embodiment of the present application provides a method, device, detection component and vehicle for predicting the health status of a battery. According to the scheme provided by the present application, as the battery to be detected continues to age during the cyclic use process, its SOH value (i.e., battery discharge capacity) will gradually decay, and its appearance will gradually expand irreversibly. Based on this, multiple expansion force values ​​and multiple residual power values ​​that vary with time during a single charging process of the battery to be detected are obtained. And the first residual power at the beginning of charging and the second residual power at the end of charging during a single charging process are obtained. In the case where the difference between the second residual power and the first residual power is greater than the preset charging depth, at least two expansion force extremes are determined from the multiple expansion force values ​​that vary with time, and the residual power corresponding to the sampling time of each expansion force extreme value in the multiple residual power values ​​that vary with time is determined. At least two expansion force extremes correspond to at least two key feature points in the charging process, and combined with the synchronized residual power, the health of the battery can be more accurately reflected. That is to say, there is a certain correlation between these irreversible expansions (which can be used as independent variables) and SOH values ​​(which can be used as dependent variables). Based on this, at least two expansion force extremes and the remaining power corresponding to each expansion force extreme are input into a preset prediction model, and the independent variables and dependent variables are regressed through the preset prediction model to predict the health status information of the battery to be tested. Compared with a single solution that uses mileage or actual battery capacity for prediction, this solution performs regression analysis on the multi-dimensional parameters that cause the SOH value to drop, thereby improving the accuracy of the battery health status. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0014] Figure 1 A schematic diagram of the structure of a rechargeable battery provided in an embodiment of the present application;

[0015] Figure 2 A schematic diagram of the structure of another rechargeable battery provided in an embodiment of the present application;

[0016] Figure 3 It is a schematic diagram of a battery charging and discharging rule provided in an embodiment of the present application;

[0017] Figure 4 It is a schematic diagram of a change rule of discharge capacity information with the number of charge and discharge cycles provided in an embodiment of the present application;

[0018] Figure 5is a schematic diagram of key characteristic points in a battery charging and discharging process provided by an embodiment of the present application;

[0019] Figure 6 It is a schematic diagram of another variation law of discharge capacity information with the number of charge and discharge cycles provided in an embodiment of the present application;

[0020] Figure 7 is a flow chart of a method for predicting battery health status provided in an embodiment of the present application;

[0021] Figure 8 It is a structural schematic diagram of a neural network model provided in an embodiment of the present application;

[0022] Fig. 9 is a flowchart of another method for predicting battery health status provided in an embodiment of the present application;

[0023] Fig.10 It is a structural schematic diagram of a battery health status prediction device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0024] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0025] It should be understood that the "multiple" mentioned in this application refers to two or more. In the description of this application, unless otherwise specified, " / " means or, for example, A / B can mean A or B; "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships, for example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, in order to facilitate the clear description of the technical solution of this application, the words "first" and "second" are used to distinguish between the same items or similar items with basically the same functions and effects. Those skilled in the art can understand that the words "first" and "second" do not limit the quantity and execution order, and the words "first" and "second" do not limit them to be different.

[0026] Before explaining the embodiments of the present application in detail, the application scenarios of the embodiments of the present application are first described.

[0027] The battery to be tested in the embodiment of the present application is a rechargeable battery. The SOH of the rechargeable battery can reflect the actual capacity, health and performance status of the rechargeable battery, that is, the percentage of the full charge capacity of the rechargeable battery relative to the rated capacity. SOH is one of the important indicators to characterize the battery status. It can be understood that the SOH value of a newly manufactured rechargeable battery is 100%. As the rechargeable battery continues to age during the cyclic use process, its SOH value will gradually decay, and its appearance will gradually undergo irreversible expansion, and these irreversible expansions are related to the SOH value of the rechargeable battery.

[0028] For example, taking a rechargeable battery as a lithium battery, the lithium battery will continue to age during the cycle, the capacity will gradually decay, and its appearance will gradually expand irreversibly. The change pattern of these irreversible expansions is correlated to the SOH state of the lithium battery. As the irreversible expansion of the lithium battery increases, the squeezing force between adjacent lithium batteries or between the lithium battery and the inner wall of its adjacent shell will gradually increase. Based on this, the embodiment of the present application detects the expansion force of the lithium battery and learns the correlation between it and the SOH state corresponding to the lithium battery, so as to predict the SOH state of the lithium battery.

[0029] In the embodiment of the present application, the expansion force of the rechargeable battery can be detected by a pressure sensor disposed on the surface of the battery, and the pressure sensor is clamped between two adjacent batteries or between a battery and a baffle. The positional relationship between the pressure sensor and the rechargeable battery is as follows: Figure 1 As shown, Figure 1 A schematic diagram of the structure of a rechargeable battery provided in an embodiment of the present application is provided. For ease of description, the following description uses a battery to represent a rechargeable battery. The battery has a front side and a side surface adjacent to the front side. The front sides of two batteries (including battery 11 and battery 12) are arranged opposite to each other. The pressure sensor 20 (which includes multiple pressure detection areas or multiple pressure detection points) is clamped between the front sides of the two batteries (including battery 11 and battery 12); when a baffle is provided between the front sides of the two batteries, the pressure sensor 20 is clamped between the front side of any one of the batteries and the baffle (also referred to as a partition or end plate). In other words, the pressure sensor 20 is provided between two adjacent batteries. When there are no other components between the two adjacent batteries, the pressure sensor 20 is clamped between the two adjacent batteries; when a baffle is provided between the two adjacent batteries, the pressure sensor 20 is clamped between the baffle and the battery. The adjacent object of the battery can be another adjacent battery or an adjacent baffle.

[0030] In some embodiments, the pressure sensor 20 may be a thin film pressure sensor. A plurality of pressure detection areas or a plurality of pressure detection points are arranged on it, and the pressure values ​​at different areas or points of the battery can be obtained. The thin film pressure sensor is used to obtain a plurality of expansion force values ​​of the battery to be detected that vary with time, and the thin film pressure sensor is communicated with a battery health state prediction device (used to execute a battery health state prediction method) to transmit the obtained plurality of expansion force values ​​that vary with time to the battery health state prediction device. The thin film pressure sensor can be arranged between the battery to be detected and other batteries, or between the battery to be detected and its adjacent shell (or baffle). In practical applications, taking lithium-ion batteries as an example, when lithium-ion batteries are charged, as lithium ions are continuously embedded, the active particles on the negative electrode undergo corresponding volume expansion, and these volume expansions will cause the battery to squeeze the thin film pressure sensor, and then the size of the squeezing force, that is, the expansion force value, can be obtained through the thin film pressure sensor.

[0031] The present application does not specifically limit the type of thin film pressure sensor, as long as the thin film pressure sensor can obtain the pressure on the battery surface. For example, the thin film pressure sensor clamped between two batteries can be a thin film flexible pressure sensor with a thickness of less than 0.5 mm.

[0032] Based on the above Figure 1 ,like Figure 2 As shown, Figure 2 A schematic diagram of the structure of another rechargeable battery provided in an embodiment of the present application, Figure 2 In the figure, two adjacent batteries (including battery 11 and battery 12) are exhibited in a disassembled form so as to more fully see the arrangement position of the pressure sensor 20. The pressure sensor 20 includes a pressure sensing sheet (also referred to as an expansion force sensing portion), on which a plurality of pressure detection areas or a plurality of pressure detection points may be arranged, and the pressure sensing sheet is clamped between the front surfaces of the two batteries (including battery 11 and battery 12).

[0033] For example, the pressure sensor is a thin film pressure sensor, and the expansion force is the total expansion force of the battery surface. In the initial state, the thin film pressure sensor is clamped between two batteries or between the battery and other devices (e.g., baffles), and the thin film pressure sensor is fixed to the battery surface by the warning force. By performing multiple cycles of charge and discharge experiments on the battery, the change trend of the total expansion force on the battery surface can be obtained, such as Figure 3 As shown, Figure 3 It is a schematic diagram of a battery charging and discharging rule provided in an embodiment of the present application.

[0034] Figure 3The horizontal axis represents the time of battery charging and discharging, and the vertical axis represents the data after the expansion force data is processed. It is only for illustrating the trend and does not represent the actual size. Constant current charging and constant current discharging can be used in the experiment. Therefore, there is a correlation between the charging and discharging time and the remaining capacity (state of charge, SOC) of the battery. That is, SOC increases with the increase of charging time and decreases with the increase of discharging time. Figure 3 It can be seen that in the initial state, the battery is not charged or discharged, and the battery will not expand, so the extrusion force on the film pressure sensor does not change. When the battery starts to charge from 0%, the battery will shrink within a period of time, resulting in a decrease in the pressure applied to the film pressure sensor, that is, the expansion force tends to decrease in the 0-1000 time period. As the battery continues to be charged, the battery shrinks to a certain extent and gradually expands. The expanded battery will increase the squeezing of the film pressure sensor, and the expansion force begins to increase.

[0035] Above Figure 3 In the graph, lighter colors indicate earlier stages of the cycle, and darker colors indicate later stages of the cycle. Figure 3 From the variation trend of the total expansion force on the battery surface, it can be seen that with the increase of the number of battery charge and discharge cycles, the expansion force of the battery at the same SOC continues to increase, and the growth is slow in the early stage of the cycle and rapid in the later stage of the cycle. By analyzing the variation trend of the battery expansion force at each stage, it can be seen that with the increase of the number of cycles, the curve shape, inflection point (i.e., extreme point) position, and inflection point size change significantly.

[0036] When the battery is cyclically charged and discharged during the experimental stage, the battery is discharged after charging is completed, and the discharge capacity information of the battery in the charge and discharge tester is extracted to obtain the change law of the discharge capacity with the number of battery cycle charge and discharge cycles, such as Figure 4 As shown, Figure 4 It is a schematic diagram of the change law of discharge capacity information with the number of charge and discharge cycles provided in an embodiment of the present application. Figure 4 In the A figure, the horizontal axis represents the number of cycles (i.e., the number of charge and discharge cycles), and the vertical axis represents the discharge battery capacity of the battery, in Ah; Figure 4 In the B figure, the horizontal axis represents the number of cycles (i.e., the number of charge and discharge cycles), and the vertical axis represents the rate of change of the discharge battery capacity of the battery. The discharge battery capacity and the rate of change of the discharge battery capacity decrease as the number of charge and discharge cycles increases. The discharge battery capacity can be regarded as the actual capacity of the battery after it is fully charged, and the discharge battery capacity is positively correlated with the SOH. Based on this, the embodiment of the present application can predict the SOH by learning the relationship between the expansion force and the discharge battery capacity.

[0037] Continue reading Figure 4It can be seen that the discharge capacity of the battery is related to its charge and discharge times (duration of use) and is linearly related. As the charge and discharge times of the battery increase, its discharge capacity (SOH value) gradually decreases. In other words, the SOH value of the battery will not drop suddenly in a short period of time. When the detection interval between the two SOH values ​​of the battery is short, the difference between the two SOH values ​​should be small; and when the detection interval between the two SOH values ​​of the battery is long, the difference between the two SOH values ​​should be large. Therefore, the correspondence between the SOH change amplitude of the battery and its usage time can be further obtained to form a battery time-SOH change amplitude limit table (that is, the correspondence between the health status value change amplitude and the time interval). The time-SOH change amplitude limit table can be used to set the maximum change amplitude of the SOH value within a certain interval time period to form a preset amplitude.

[0038] In the present application embodiment, based on the above Figure 3 When the battery is cyclically charged and discharged, the expansion force curve of each cycle is marked according to the changing characteristics of the battery expansion force during the process, and the two peaks and two valleys of the expansion force curve in the charging stage are marked, such as Figure 5 As shown, Figure 5 It is a schematic diagram of key characteristic points in a battery charging and discharging process provided in an embodiment of the present application. Figure 5 , key characteristic point 1 (valley value), key characteristic point 2 (peak value), key characteristic point 3 (valley value) and key characteristic point 4 (peak value) in the charging stage of one cycle are shown.

[0039] Based on the above Figure 5 The marking method of the four key characteristic points of the other number of cycles is used to mark the four key characteristic points. That is, for the expansion force curve of each number of cycles, the positions and expansion force data of the two peaks and two valleys are marked, so as to obtain the change law of the expansion force data of the two peaks and two valleys with the number of cycles, such as Figure 6 As shown, Figure 6 It is a schematic diagram of another type of discharge capacity information provided in an embodiment of the present application that changes with the number of charge and discharge cycles. Figure 6 The horizontal axis represents the number of cycles (i.e., the number of charge and discharge cycles), and the vertical axis represents the data after processing the expansion force data. It is only for illustrating the trend and does not represent the actual size.

[0040] in, Figure 6 Figure A shows the change of key feature point 1 (valley value) with the number of cycles. Figure 6 Figure B shows the change of key feature point 2 (peak value) with the number of cycles. Figure 6 Figure C shows the change of key feature point 3 (valley value) with the number of cycles. Figure 6 Figure D in FIG. 4 shows the change pattern of key feature point 4 (peak value) with the number of cycles.

[0041] from Figure 6 It can be found that as the number of battery charge and discharge cycles increases, the changes in the peak and valley values ​​in the battery expansion force curve basically show a downward trend, and the rate of decline slows down in the later period, and the curve becomes slower. The changes in key characteristic points are strongly correlated with the battery SOH (i.e., the discharge battery capacity). Based on this, the embodiment of the present application can establish the correlation between the two peaks and two valleys and the battery SOH (discharge battery capacity) through the above data.

[0042] In some embodiments, based on the above analysis of the strong correlation between the expansion force data and the discharge battery capacity, the initial prediction model is trained so that the trained prediction model can predict the SOH based on the expansion force data and its corresponding SOC. Based on this, the training process of the preset prediction model is described. Figure 7 As shown, Figure 7 This is a flowchart of a method for predicting battery health status provided in an embodiment of the present application.

[0043] S101, obtaining expansion force change data samples, remaining power change data samples, and battery discharge capacity samples of a battery sample during a charging process and a discharging process after charging.

[0044] In the embodiment of the present application, the expansion force of the battery sample can be detected by a pressure sensor to obtain an expansion force change data sample, and the SOC of the battery sample can be detected by a power sensor or a battery management system (BMS) to obtain an SOC change data sample. During the experiment, the charge and discharge tester is connected to the battery sample, and the battery sample is discharged after charging, and the discharge capacity sample of the battery in the charge and discharge tester is extracted.

[0045] It should be noted that the expansion force change data samples include not only the expansion force change data samples of different battery samples during the charging process and the process of discharging after charging (i.e., the charging and discharging process), but also the expansion force change data samples of the same battery sample during different charging and discharging processes. Similarly, the remaining power change data samples and the battery discharge capacity samples correspond to the charging and discharging process.

[0046] S102. Determine a plurality of expansion force extreme value samples according to the expansion force change data samples.

[0047] In the embodiment of the present application, the expansion force change data sample includes a plurality of expansion force value samples that change over time. Figure 3 and Figure 5A marking method for key feature points in a method for selecting multiple expansion force extreme value samples from multiple expansion force value samples, wherein the expansion force extreme value samples are expansion force maximum value samples or expansion force minimum value samples.

[0048] Exemplarily, the expansion force change data samples can be plotted into a curve, and the valley and peak values ​​can be found in the curve, thereby determining multiple expansion force extreme value samples. Alternatively, the expansion force change rate can be obtained, and the expansion force extreme value samples can be determined in the expansion force change data samples according to the expansion force change rate. For example, if the expansion force change rate changes from a negative value to a positive value, it means that the point is a valley value, and if the expansion force change rate changes from a positive value to a negative value, it means that the point is a peak value, thereby determining multiple expansion force extreme value samples.

[0049] S103 . Determine, in the remaining power change data samples, the remaining power samples corresponding to the sampling moments of the respective expansion force extreme value samples among the plurality of expansion force extreme value samples.

[0050] The expansion force change data samples and the remaining power change data samples both change with time. For each charging and discharging process, at least two expansion force extreme value samples can be determined from the multiple expansion force value samples that change with time included in the expansion force change data samples, and the remaining power samples corresponding to the sampling moments of the respective expansion force extreme value samples can be determined from the multiple remaining power samples that change with time. By analogy, multiple expansion force extreme value samples and their corresponding remaining power samples in multiple charging and discharging processes can be obtained.

[0051] According to the multiple expansion force extreme value samples and the remaining power samples corresponding to their respective sampling moments, sample pairs of expansion force extreme value and remaining power can be constructed, thus obtaining multiple sample pairs. Figure 3 and Figure 5 For the key characteristic points in each charging and discharging process, if the extreme values ​​in the charging stage are counted, 4 sample pairs can be obtained, if the extreme values ​​in the discharging stage are counted, 2 sample pairs can be obtained, and if the extreme values ​​in the entire charging and discharging stage are counted, 6 sample pairs can be obtained. The number of sample pairs can be appropriately set by those skilled in the art according to actual conditions, and the sample pairs can be located in the charging stage, the discharging stage, or the entire charging and discharging stage.

[0052] In the embodiment of the present application, after a charge and discharge process of the same battery sample, the battery discharge capacity sample is obtained, and the multiple sample pairs obtained above and the battery discharge capacity sample are used as a group of samples of the battery sample. In a complete charge and discharge process, multiple sample pairs correspond to the same battery discharge capacity sample, and so on, multiple groups of samples are constructed. In other words, the expansion force change data sample, the remaining power change data sample and the battery discharge capacity sample are used as a group of samples.

[0053] It should be noted that, by comprehensively considering the different charging and discharging processes of battery samples, or the charging and discharging processes of different battery samples, the data of the same key feature point can also be taken as a group of samples. Figure 3 and Figure 5 , multiple sample pairs of key feature point 1 and the battery discharge capacity samples of each sample pair are taken as a group of samples, multiple sample pairs of key feature point 2 and the battery discharge capacity samples of each sample pair are taken as a group of samples, and so on, to construct multiple groups of samples. The present application embodiment does not limit the construction of each group of samples.

[0054] S104. Through the initial prediction model, regression analysis is performed on multiple expansion force extreme value samples, and the remaining power samples and battery discharge capacity samples corresponding to each of the expansion force extreme value samples, until the health status information of the battery samples predicted by the initial prediction model meets the preset training conditions, thereby obtaining the preset prediction model.

[0055] In the embodiment of the present application, the expansion force extreme value sample and its corresponding remaining power sample are used as independent variables, and the battery discharge capacity sample is used as the dependent variable. The independent variable and the dependent variable are subjected to regression analysis through the initial prediction model to learn the relationship between the independent variable and the dependent variable until the preset training condition is reached. The preset training condition can be that the difference between the predicted health status information of the battery sample and the battery discharge capacity sample is within a preset range, or that the training reaches a preset number of times to obtain a preset prediction model. The embodiment of the present application does not limit the preset number of times and the preset range.

[0056] For example, the above Figure 3 and Figure 5 Taking the four key characteristic points in the middle charging stage as an example, the regression analysis method is used to analyze the correlation between the two peak values, the two valley values ​​and the battery SOH (ie, the battery discharge capacity) to obtain the SOH prediction model (ie, the preset prediction model).

[0057] Among them, regression analysis is used to study the relationship between one or more independent variables (e.g., explanatory variables, independent variables) and dependent variables (e.g., response variables, dependent variables), by building a mathematical model to predict the value of the dependent variable and explain the impact of the independent variable on the dependent variable.

[0058] In the embodiment of the present application, the expansion force extreme value and its corresponding remaining power can reflect the health of the battery. The expansion force extreme value and its corresponding remaining power are used as independent variables, and the battery discharge capacity which is positively correlated with SOH is used as the dependent variable. The relationship between the independent variable and the dependent variable is regressed through the initial prediction model to obtain a preset prediction model that can predict SOH. This improves the accuracy of SOH when the preset prediction model is used to predict SOH in the future.

[0059] In some embodiments, the preset prediction model is a linear regression model, a nonlinear regression model, an interactive regression model, a decision tree regression model, or a neural network regression model. That is, the correlation between the two peak values, the two valley values, and the battery SOH can be analyzed by using the analysis methods of linear regression, nonlinear regression, interactive regression, decision tree regression, or neural network regression to obtain the preset prediction model. The following are respectively explained.

[0060] Linear regression is a method used to model the linear relationship between independent variables and dependent variables, as shown in formula (1).

[0061] Y=β0+β1X1+β2X2+β3X3+…+β n X n +∈ (1)

[0062] In the above formula (1), X1, X2, ...X n represents the independent variable, β0, β1…β n Represents coefficient, ∈ represents constant, and Y represents dependent variable. The linear regression model assumes that the dependent variable is a linear combination of independent variables, and determines the coefficient by fitting to minimize the error between the predicted value and the actual observed value. The calculation is fast and easy to implement, but it can only fit linear relationships, and the fitting effect for nonlinear relationships is poor, and it is sensitive to outliers.

[0063] Non-linear regression can also be called multivariate nonlinear regression, which expands the linear regression model and allows the relationship between the independent variable and the dependent variable to be nonlinear. It is usually modeled by introducing nonlinear functions or polynomials, as shown in formula (2).

[0064] Y=f0+f1(X1)+f2(X2)+f3(X3)+…+f n (X n ) (2)

[0065] In the above formula (2), X1, X2, ...X n Represents the independent variable, f represents the function, and Y represents the dependent variable. Nonlinear regression models can fit more complex nonlinear relationships and are more flexible than linear models, but the model is more complex and prone to overfitting.

[0066] Interactive regression takes into account the interaction between independent variables and allows the change of the dependent variable to be affected by the mutual influence between different independent variables, as shown in formula (3).

[0067] Y=β0+β1X1+β2X2+f3(X1·X2)+f4(X1+X3)+…+∈ (3)

[0068] In the above formula (3), X1, X2, ...X n Represents independent variables, f represents function, β0, β1 and β2 represent coefficients, ∈ represents constant, and Y represents dependent variable. The interactive regression model can capture the complex relationship between independent variables, improve the expressive power of the model, increase the complexity of the model, and may require more data to stabilize the estimated parameters.

[0069] Decision tree regression divides data into different regions by building a decision tree, and uses a simple model (e.g., a constant) to make predictions in each region. It is suitable for processing complex nonlinear relationships and high-dimensional data, and has no strict assumptions about data distribution. It can be used in data mining and machine learning.

[0070] Neural network regression is a network structure composed of multiple neurons, such as Figure 8 As shown, Figure 8 is a structural diagram of a neural network model provided in an embodiment of the present application, Figure 8 In this case, X represents the independent variable, Y represents the dependent variable, and [1] represents the first layer, a [2] represents the second layer, a [3] Indicates the third layer, a [4] represents the fourth layer, and each layer from the first to the third layer includes multiple neurons (for example, n). Complex nonlinear relationships are fitted through the connection relationship between multiple layers of neurons. A fully connected multilayer perceptron (MLP) structure is usually used, in which the neurons in each hidden layer are connected to all neurons in the previous layer. This enables the neural network regression model to capture more complex nonlinear relationships.

[0071] It should be noted that in the embodiment of the present application, when analyzing the correlation between the expansion force and the battery discharge capacity, it is not limited to the linear regression, multivariate nonlinear regression, interactive regression, decision tree regression, neural network regression, etc. listed above. The embodiment of the present application does not impose any restrictions on this, as long as the relationship between the independent variable and the dependent variable can be learned.

[0072] In some embodiments, a suitable model can be further selected for regression analysis. Linear regression, multivariate nonlinear regression, interactive regression, decision tree regression, and neural network regression models are used to predict SOH. A data set is constructed using 4 highly correlated expansion force data and their corresponding SOC and battery discharge capacity. 80% of the data set is used for prediction and 20% of the data set is used for testing to further compare the accuracy and estimation error of each model. The accuracy of the model can be judged by comparing the mean square error. The experimental results show that the prediction effect of the neural network regression model is better, and the accuracy of SOH can be greater than 95%. Based on this, the embodiment of the present application can use a neural network regression model to analyze the correlation between 2 peaks, 2 valleys and battery SOH (battery discharge capacity) to obtain a SOH prediction model (i.e., a preset prediction model) to improve the accuracy of the prediction results, that is, to improve the accuracy of the health status information.

[0073] In some embodiments, the training process may also include the following steps: obtaining the influencing attributes of the battery sample, the influencing attributes including the ambient temperature and / or the battery type. Based on this, the above S104 may also be implemented in the following manner. Through the initial prediction model, a regression analysis is performed on multiple expansion force extreme value samples belonging to the target influencing attribute, the remaining power samples corresponding to each of the expansion force extreme value samples, and the battery discharge capacity samples, until the health status information of the battery sample predicted by the initial prediction model meets the preset training conditions, and the preset prediction model corresponding to the target influencing attribute is obtained; the target influencing attribute is any influencing attribute; or, through the initial prediction model, a regression analysis is performed on the influencing attribute, multiple expansion force extreme value samples, the remaining power samples corresponding to each of the expansion force extreme value samples, and the battery discharge capacity samples, until the health status information of the battery sample predicted by the initial prediction model meets the preset training conditions, and the preset prediction model is obtained.

[0074] In the embodiment of the present application, different ambient temperatures have a greater impact on the expansion of the battery. Exemplarily, due to the lower ambient temperature, lithium ions will accumulate on the surface of the graphite layer, the battery thickness will increase, and the expansion force will also increase. Therefore, the increase in the expansion force is small. For example, the increase in the expansion force is small when charging at 25°C. Higher ambient temperatures not only increase the volume of the battery, but also generate a solid electrolyte interface membrane (SEI) film during the charging process, and therefore have a greater expansion force. Based on this, when training the initial prediction model, the ambient temperature during the charging and discharging process of the battery sample is also obtained.

[0075] In one implementation, the initial prediction model is used to perform regression analysis on multiple groups of samples at the same ambient temperature to obtain the SOH prediction model at the ambient temperature (i.e., the preset prediction model), and so on, the SOH prediction models at multiple different ambient temperatures can be obtained. In this way, there can be multiple SOH prediction models corresponding to ambient temperatures. When predicting the SOH of the battery to be tested, the ambient temperature of the environment in which the battery to be tested is located is obtained, and the SOH prediction model corresponding to the ambient temperature is called for prediction, which can improve the accuracy of the prediction results.

[0076] Among them, the multiple different ambient temperatures can be several common ambient temperatures, which can be appropriately set by technicians in this field according to actual conditions, and the embodiments of the present application are not limited to this.

[0077] Exemplarily, taking the expansion force extreme value sample in each charge and discharge process including 2 peaks and 2 valleys as an example, the battery is subjected to multiple cycles of charge and discharge experiments at different ambient temperatures to obtain the correlation between the 2 peaks, 2 valleys and the battery SOH (battery discharge capacity) at different ambient temperatures. For example, multiple cycles of charge and discharge experiments are performed at an ambient temperature of 25°C to obtain the correlation between the 2 peaks, 2 valleys and the battery SOH (battery discharge capacity) at 25°C; multiple cycles of charge and discharge experiments are performed at an ambient temperature of 45°C to obtain the correlation between the 2 peaks, 2 valleys and the battery SOH (battery discharge capacity) at 45°C; multiple cycles of charge and discharge experiments are performed at an ambient temperature of 5°C to obtain the correlation between the 2 peaks, 2 valleys and the battery SOH (battery discharge capacity) at 5°C.

[0078] In another implementation, the ambient temperature is used as an independent variable and added to the multiple groups of samples constructed above, and then the multiple groups of samples after adding the ambient temperature are regressed and analyzed by the initial prediction model to obtain the SOH prediction model (i.e., the preset prediction model). The SOH prediction model has the ability to predict the SOH based on the expansion force parameters under a variety of different ambient temperatures. In this way, when predicting the SOH of the battery, the ambient temperature of the environment where the battery is located is obtained, and the ambient temperature and at least two expansion force extremes of the battery and the remaining power corresponding to each expansion force extreme are input into the SOH prediction model. The accuracy of the prediction result can be improved by predicting the SOH through the SOH prediction model.

[0079] Exemplarily, the ambient temperature is used as one of the dependent variables, and the correlation between the corresponding two peak values, two valley values ​​and the battery SOH (battery discharge capacity) is combined to train the initial prediction model to obtain the SOH prediction model.

[0080] In the embodiment of the present application, the battery type indicates the battery structure and installation method. Different battery types have slightly different battery structures, performance, installation methods, preload forces applied during installation, and usage conditions. These differences will cause inconsistent pressure changes during battery use. When predicting the SOH of batteries of different battery types, if the same preset prediction model is used for prediction, the predicted SOH will be different. Based on this, when training the initial prediction model, the battery type of the battery sample is also obtained.

[0081] In one implementation, multiple groups of samples under the same battery type are regressed and analyzed by the initial prediction model to obtain the SOH prediction model under the battery type (i.e., the preset prediction model), and so on, the SOH prediction models under multiple different battery types can be obtained. In this way, there can be SOH prediction models corresponding to multiple battery types. When predicting the SOH of the battery to be tested, the battery type of the battery to be tested is obtained, and the SOH prediction model corresponding to the battery type is called for prediction, which can improve the accuracy of the prediction results.

[0082] In another implementation, the battery type is added as an independent variable to the multiple groups of samples constructed above, and then the multiple groups of samples after adding the battery type are regressed and analyzed by the initial prediction model to obtain the SOH prediction model (i.e., the preset prediction model). The SOH prediction model has the ability to predict the SOH based on the expansion force parameters under a variety of different battery types. In this way, when predicting the SOH of the battery, the battery type of the battery is obtained, and the battery type and at least two expansion force extremes of the battery and the remaining power corresponding to each expansion force extreme are input into the SOH prediction model. The accuracy of the prediction result can be improved by predicting the SOH through the SOH prediction model.

[0083] It should be noted that when training the initial prediction model, the initial prediction model can be used to perform regression analysis on multiple groups of samples of the same battery type and the same ambient temperature to obtain the SOH prediction model of the battery type and ambient temperature. By analogy, the SOH prediction models of multiple different battery types at different ambient temperatures can be obtained. In this way, there can be SOH prediction models corresponding to each battery type and each ambient temperature. When predicting the SOH of the battery to be tested, the battery type of the battery to be tested and the ambient temperature of the environment in which it is located are obtained, and the SOH prediction model corresponding to the battery type and ambient temperature is called for prediction, which can improve the accuracy of the prediction results.

[0084] When training the initial prediction model, the battery type and ambient temperature are used as independent variables and added to the multiple groups of samples constructed above. Then, the multiple groups of samples after adding the battery type and ambient temperature are regressed through the initial prediction model to obtain the SOH prediction model. In this way, when predicting the SOH of the battery, the battery type and the ambient temperature of the environment in which the battery is located are obtained, and the battery type, ambient temperature, at least two expansion force extremes of the battery, and the remaining power corresponding to each expansion force extreme are input into the SOH prediction model. The SOH is predicted by the SOH prediction model, which can improve the accuracy of the prediction result.

[0085] Based on the above preset prediction model, the present application embodiment provides a method for predicting the health status of a battery, such as Fig. 9 As shown, Fig. 9 1 is a flowchart of another method for predicting the health status of a battery provided in an embodiment of the present application. The method for predicting the health status of a battery includes:

[0086] S201, obtaining multiple expansion force values ​​and multiple remaining capacities that vary with time during a single charging process of the battery to be tested, and obtaining a first remaining capacities at the beginning of charging and a second remaining capacities at the end of charging during the single charging process.

[0087] When charging a battery, the battery's operating condition is relatively stable. For example, when charging a vehicle's battery, the vehicle is generally stationary and the battery is in a relatively stable operating condition. At this time, the change in the battery's expansion force is less affected by external factors and will not produce large fluctuations. When the vehicle is driving, due to the influence of the vehicle's driving conditions and various external factors, the battery's expansion force changes during discharge will fluctuate, and these fluctuations will affect the accuracy of the expansion force. Based on this, the embodiment of the present application obtains the expansion force value of the battery to be tested during a single charging process.

[0088] In an embodiment of the present application, the expansion force of the battery to be tested during a single charging process is detected by a pressure sensor to obtain multiple expansion force values ​​that change with time, and the SOC of the battery to be tested during the charging process or the discharging process is detected by a power sensor or a BMS to obtain multiple SOC values ​​that change with time.

[0089] S202. When the difference between the second remaining power and the first remaining power is greater than a preset depth of charge, at least two expansion force extremes are determined from a plurality of expansion force values ​​that vary with time, and a remaining power corresponding to a sampling moment of each of the at least two expansion force extremes is determined from a plurality of remaining power that varies with time.

[0090] During use, the battery will often go through the charging and discharging process. When to charge, how much charge to charge, when to discharge, and how much charge to discharge are all affected by external factors, and these influences will cause a sudden change in the battery's expansion force. Therefore, if only one expansion force extreme value, that is, a key characteristic point, is determined during a single charging process, the SOH value predicted based on the key characteristic point cannot accurately reflect the health of the battery. Based on this, in the embodiment of the present application, the charging process that can determine at least two expansion force extreme values ​​is regarded as an effective charging process. By obtaining at least two expansion force extreme values ​​(i.e., two key characteristic points) under a single charging condition, predicting the SOH value according to a preset prediction model, and judging whether the predicted SOH value is accurate by comparing the difference between the SOH value and the historical SOH value, the accuracy of the prediction result is improved.

[0091] Normally, during the battery charging stage, when the single charge depth of the battery reaches the preset charge depth, there are at least two expansion force extremes in the charging process. The single charge depth indicates the difference between the second remaining power at the end of charging and the first remaining power at the start of charging, that is, the difference between the SOC at the start mark of the charging condition and the SOC at the stop mark of charging.

[0092] Among them, the preset charging depth can be appropriately set by those skilled in the art according to actual conditions, for example, 60%, 50%, 70%, etc., as long as it can ensure that at least two expansion force extremes can be obtained during a single charging process.

[0093] In the embodiment of the present application, the expansion force extreme value selected from the multiple expansion force values ​​that change with time is the expansion force maximum value or the expansion force minimum value. Exemplarily, the multiple expansion force values ​​that change with time can be plotted into a curve, and the valley value and the peak value can be found in the curve, so as to determine at least two expansion force extreme values. Alternatively, the expansion force change rate can be obtained, and the expansion force extreme value can be determined from the multiple expansion force values ​​that change with time according to the expansion force change rate. For example, the expansion force change rate changes from a negative value to a positive value, indicating that the point is a valley value, and the expansion force change rate changes from a positive value to a negative value, indicating that the point is a peak value, so as to determine at least two expansion force extreme values. After determining at least two expansion force extreme values, for each expansion force extreme value, the remaining power corresponding to its sampling time is obtained, so as to determine the remaining power corresponding to the sampling time of each expansion force extreme value in the at least two expansion force extreme values.

[0094] In an embodiment of the present application, at least two expansion force extreme values ​​(i.e., at least two key characteristic points) under a single charging condition and the remaining power corresponding to the sampling time are obtained to improve the data validity, so as to improve the accuracy of the prediction result when performing SOH prediction subsequently.

[0095] S203, inputting at least two expansion force extreme values ​​and the remaining power corresponding to each of the expansion force extreme values ​​into a preset prediction model, and predicting the health status information of the battery to be tested through the preset prediction model; wherein the preset prediction model is used to learn the correlation between independent variables and dependent variables, the independent variables include the expansion force parameters, and the dependent variables include the battery discharge capacity that is positively correlated with the health status information.

[0096] In an embodiment of the present application, the preset prediction model has the function of learning the correlation between the expansion force extreme value and the battery discharge capacity, at least two expansion force extreme values ​​and their corresponding remaining power are input into the preset prediction model, and regression analysis is performed through the preset prediction model to output the SOH value (i.e., health status information) of the battery to be tested.

[0097] According to the solution provided by the present application, as the battery to be tested continues to age during the cyclic use process, its SOH value (i.e., battery discharge capacity) will gradually decay, and its appearance will gradually expand irreversibly. Based on this, multiple expansion force values ​​and multiple residual power values ​​that vary with time during a single charging process of the battery to be tested are obtained. And the first residual power at the beginning of charging and the second residual power at the end of charging during a single charging process are obtained. In the case where the difference between the second residual power and the first residual power is greater than the preset charging depth, at least two expansion force extremes are determined from the multiple expansion force values ​​that vary with time, and the residual power corresponding to the sampling moment of each expansion force extreme value in the at least two expansion force extremes is determined from the multiple residual power values ​​that vary with time. At least two expansion force extremes correspond to at least two key feature points in a single charging process, and combined with the synchronized residual power, the health of the battery can be more accurately reflected. That is to say, there is a certain correlation between these irreversible expansions (which can be used as independent variables) and SOH values ​​(which can be used as dependent variables). Based on this, at least two expansion force extremes and the remaining power corresponding to each expansion force extreme are input into a preset prediction model, and the independent variables and dependent variables are regressed through the preset prediction model to predict the health status information of the battery to be tested. Compared with a single solution that uses mileage or actual battery capacity for prediction, this solution performs regression analysis on the multi-dimensional parameters that cause the SOH value to drop, thereby improving the accuracy of the battery health status.

[0098] In some embodiments, the categories of expansion force extreme values ​​include a first category, a second category, a third category, and a fourth category, and the remaining power corresponding to the first category, the second category, the third category, and the fourth category gradually increases, the first category and the third category represent valley values, and the second category and the fourth category represent peak values.

[0099] In the embodiment of the present application, the expansion force value corresponding to the first category is smaller than the expansion force value corresponding to the third category, and the expansion force value corresponding to the second category is smaller than the expansion force value corresponding to the fourth category. For multiple expansion force values ​​that change with time during a single charging process, the expansion force extreme value is selected in chronological order. If the time of the charging process is long enough, for example, from any SOC between 0% and 10% to any SOC between 90% and 100%, then 4 expansion force extreme values ​​can be obtained. The first category corresponds to the above Figure 3 The key feature point 1 in is the first valley value; the second category corresponds to the above Figure 3 The key feature point 2 in is the first peak; the third category corresponds to the above Figure 3 The key feature point 3 in is the second valley value; the fourth category corresponds to the above Figure 3 For a single charging process, at least two expansion force extremes are determined from a plurality of expansion force values ​​that vary with time, and the expansion force extremes may be of the first category, the second category, the third category, or the fourth category.

[0100] The step of determining at least two expansion force extreme values ​​in the above S202 can also be implemented in the following manner: determining at least two initial expansion force extreme values ​​from a plurality of expansion force values ​​that change with time, and determining the initial remaining power corresponding to the sampling moment of each of the at least two initial expansion force extreme values ​​from a plurality of remaining power that change with time; determining the category of each initial expansion force extreme value according to the interval where the initial remaining power corresponding to the sampling moment of each initial expansion force extreme value is located; and taking the initial expansion force extreme value that belongs to a preset category as the expansion force extreme value, so as to obtain at least two expansion force extreme values.

[0101] In the embodiment of the present application, the category of the expansion force extreme value is related to the remaining power, as described above Figure 3 and Figure 5 As shown. Key feature point 1 (corresponding to the first category), key feature point 2 (corresponding to the second category), key feature point 3 (corresponding to the third category), and key feature point 4 (corresponding to the fourth category) correspond to different SOC intervals. Therefore, based on the interval where the initial remaining power is located, it can be determined which category the initial expansion force extreme value belongs to. For example, the SOC of key feature point 1 is about 20%-30%; the SOC of key feature point 2 is about 40%; the SOC of key feature point 3 is about 60%-70%; and the SOC of key feature point 4 is about 90%.

[0102] Among them, the preset category can be any one of the first category, the second category, the third category and the fourth category, or any two, or any three. It can be specifically set by technical personnel in this field according to actual conditions, and the embodiments of the present application are not limited to this.

[0103] According to the above Figure 3 , it can be seen that with the increase in the number of battery charge and discharge cycles, compared with key feature point 1, the expansion force changes of key feature point 2, key feature point 3 and key feature point 4 are more significant, that is to say, they are more representative and can better reflect the health status of the battery to be tested. In the actual charging process, most users start charging before the battery SOC reaches 30%, and stop charging before the battery reaches 90%. Based on this, key feature point 2 and key feature point 3 can be selected for SOH prediction. In other words, the preset category can be set to the second category and the third category.

[0104] Exemplarily, three initial expansion force extreme values ​​(including key feature point 2, key feature point 3 and key feature point 4) are determined from multiple expansion force values ​​that change with time, and the preset categories include the second category and the third category. The initial remaining power corresponding to the sampling time of each of the three initial expansion force extreme values ​​is determined from multiple remaining power that change with time; according to the interval where the initial remaining power corresponding to the sampling time of each initial expansion force extreme value is located, it can be determined that the categories of the three initial expansion force extreme values ​​are the second category, the third category and the fourth category. The first two initial expansion force extreme values ​​are selected as the two expansion force extreme values ​​that are finally input into the preset prediction model.

[0105] In the embodiment of the present application, a preset category is pre-set, and the category of the initial expansion force extreme value is determined according to the interval of the initial remaining power corresponding to the sampling time of the initial expansion force extreme value. The initial expansion force extreme value of the preset category is used as the expansion force extreme value, that is, the initial expansion force extreme value in a specific SOC interval is used as the expansion force extreme value, and the remaining power corresponding to its sampling time is determined. By selecting the expansion force extreme value of the preset category, the amount of data processing can be reduced, the prediction efficiency can be improved, and the expansion force extreme value of the preset category can better reflect the health status of the battery to be tested and improve the accuracy of the prediction result.

[0106] In some embodiments, the method for predicting the battery health status also includes the following steps: obtaining the influencing attributes of the battery to be detected during a single charging process; the influencing attributes indicate factors that are relevant to the battery health status assessment, and the influencing attributes include ambient temperature and / or battery type. Based on this, the above S203 can also be implemented in the following manner. Input the influencing attributes, at least two expansion force extremes, and the remaining power corresponding to each expansion force extreme value into a preset prediction model, and predict the health status information of the battery to be detected through the preset prediction model; the independent variables include ambient temperature parameters, battery type parameters, and expansion force parameters; or, determine a prediction model that matches the influencing attributes; input at least two expansion force extremes and the remaining power corresponding to each expansion force extreme value into a prediction model that matches the influencing attributes, and predict the health status information of the battery to be detected through the prediction model that matches the influencing attributes.

[0107] The above-mentioned trained prediction model can have multiple preset prediction models for different battery types. In this way, when predicting the SOH of the battery to be tested, the battery type of the battery to be tested is obtained, and the preset prediction model corresponding to the battery type is called for prediction, which can improve the accuracy of the prediction result.

[0108] When training the initial prediction model, the battery type can also be used as an independent variable for prediction. The preset prediction model has the ability to predict SOH based on the expansion force parameters of a variety of different battery types. In this way, when predicting the SOH of the battery, the battery type of the battery is obtained, and the battery type and at least two expansion force extremes of the battery and the remaining power corresponding to each expansion force extreme are input into the preset prediction model for prediction, which can improve the accuracy of the prediction result.

[0109] Similarly, for the ambient temperature, when predicting the SOH of the battery to be tested, the ambient temperature of the environment in which the battery to be tested is located is obtained, and a preset prediction model corresponding to the ambient temperature is called for prediction, which can improve the accuracy of the prediction result. Alternatively, the ambient temperature and at least two expansion force extremes of the battery and the remaining power corresponding to each expansion force extreme are input into a preset prediction model for prediction, which can improve the accuracy of the prediction result.

[0110] It should be noted that, in the case where the influencing attributes include ambient temperature and battery type, when predicting the SOH of the battery to be detected, the battery type of the battery to be detected and the ambient temperature of the environment in which it is located are obtained, and the preset prediction model corresponding to the battery type and ambient temperature is called for prediction, which can improve the accuracy of the prediction result. Alternatively, the battery type and ambient temperature, as well as at least two expansion force extremes of the battery and the remaining power corresponding to each expansion force extreme value, are input into the preset prediction model for prediction, which can improve the accuracy of the prediction result.

[0111] In some embodiments, the above S203 can also be implemented in the following ways: predict the health status value of the battery to be tested by using a preset prediction model; obtain the historical health status value predicted last time for the battery to be tested, as well as the last prediction time; determine the preset amplitude based on the interval between the last prediction time and the current time, and the correspondence between the amplitude of the health status value change and the time interval; when the health status value is less than the historical health status value, and the difference between the health status value and the historical health status value is less than the preset amplitude, use the health status value as health status information.

[0112] In an embodiment of the present application, the difference judgment between the health status value and the historical health status value can be achieved in the following manner: obtain the time interval between the last predicted historical health status value of the battery to be tested and the health status value predicted this time, and obtain the preset amplitude corresponding to the time interval through the time-SOH change amplitude limit table (that is, the correspondence between the health status value change amplitude and the time interval); determine whether the difference between the health status value and the historical health status value is less than the preset amplitude.

[0113] Based on the above Figure 4 In the analysis of the time-SOH variation range limit table of the battery (i.e., the correspondence between the variation range of the health status value and the time interval), in an embodiment of the present application, the time interval between the last predicted historical health status value and the health status value predicted this time is obtained, and the preset range corresponding to the time interval is obtained through the time-SOH variation range limit table. Then, it is determined whether the difference between the health status value and the historical health status value is less than the preset range. The accuracy of the health status value predicted this time can be further judged, thereby improving the accuracy of the health status information.

[0114] As the number of battery charge and discharge times increases, its SOH value gradually decreases. Based on this, the historical SOH value predicted last time is obtained to determine whether the SOH value predicted this time is less than the historical SOH value predicted last time, and whether the difference between the SOH value predicted this time and the historical SOH value predicted last time is less than the preset range. If so, it means that the SOH value predicted this time is valid, and the SOH value predicted this time is used as the health status information of the battery to be tested, which improves the accuracy of the health status information.

[0115] If the SOH value is greater than or equal to the historical SOH value, or the difference between the SOH value and the historical SOH value is greater than or equal to the preset amplitude, it means that the predicted SOH value is invalid, and multiple expansion force values ​​that change with time in the next charging process of the battery to be tested are continued to be collected to predict the next SOH value.

[0116] Furthermore, the health status information can also be sent to other systems or other devices, and the actual SOH value of the battery to be tested (SOH value calculated by other means) can also be updated or calibrated, thereby solving the technical problem of inaccurate calculation results when calculating the SOH value of the battery to be tested.

[0117] It should be noted that the difference between the currently predicted SOH value and the previously predicted historical SOH value may refer to the difference between the two SOH values, or the percentage difference between the two SOH values. The preset amplitude may be obtained by searching the battery time-SOH variation amplitude limit table (i.e., the correspondence between the health state value variation amplitude and the time interval), wherein the health state value variation amplitude may be a difference or percentage positively correlated with the time interval.

[0118] In some embodiments, the expansion force value is the total expansion force, the single point expansion force, or the total expansion force within a preset detection area.

[0119] In the present application embodiment, as mentioned above Figure 2 As shown, the pressure sensor may be provided with multiple pressure detection areas or multiple pressure detection points, and the pressure values ​​at different detection areas or different detection points of the battery may be obtained. Based on this, the expansion force value may be the total expansion force of the battery to be detected, that is, the sum of the pressure values ​​at multiple detection areas or multiple detection points is taken as the expansion force value of the battery to be detected.

[0120] The expansion force value may also be a single-point expansion force at a certain detection area or a certain detection point that is sensitive to pressure. For example, if the pressure sensor is a thin-film pressure sensor, the thin-film pressure sensor may be set as a single-point pressure sensor to obtain the pressure at a certain point of the battery to be detected.

[0121] The expansion force value may also be the total expansion force of a preset detection area that is sensitive to the expansion force, and the preset detection area may include multiple detection points or multiple detection areas. For example, taking the pressure sensor as a thin film pressure sensor as an example, the thin film pressure sensor may be set as a pressure sensor covering a partial area of ​​the battery to be detected to obtain the pressure magnitude of the partial area of ​​the battery to be detected.

[0122] By setting the expansion force values ​​under different conditions to suit different application scenarios, the richness and diversity of the prediction methods are improved.

[0123] The above respectively introduces the embodiments of selecting the expansion force extreme value belonging to the preset category; the embodiment of increasing the ambient temperature and / or battery type of the battery to be detected; the embodiment of judging the SOH value predicted this time, etc. Based on this, the embodiment of the present application can also provide an embodiment of combining any two or more of the above embodiments to further improve the accuracy of the prediction result.

[0124] Exemplarily, the SOH value is output according to the preset prediction model combined with the ambient temperature, the battery type and at least two key feature points. The preset prediction model can learn the correspondence between the ambient temperature, the battery type and the key feature points and the SOH value. It is determined whether the predicted SOH value is less than the historical SOH value predicted last time, and whether the difference between the SOH value and the historical SOH value is less than a preset range; if so, the SOH value is used as the current SOH value.

[0125] It should be noted that the embodiments of the present application can also provide embodiments in other combinations, which are not listed here one by one.

[0126] Based on the battery health status prediction method provided in the above embodiment, Fig.10 A schematic diagram of a battery health status prediction device provided in an embodiment of the present application is shown in FIG. Fig.10 As shown, the battery health status prediction device 100 includes: a processor 1001, a memory 1002, and a computer program 1003 stored in the memory 1002 and executable on the processor 1001. When the processor 1001 executes the computer program 1003, the steps in the battery health status prediction method in the above embodiment are implemented.

[0127] The battery health state prediction device 100 may be a general-purpose computer device or a dedicated computer device. In a specific implementation, the battery health state prediction device 100 may be a desktop computer, a portable computer, a network server, a PDA, a mobile phone, a tablet computer, a wireless terminal device, a communication device or an embedded device. The embodiment of the present application does not limit the type of the battery health state prediction device 100. Those skilled in the art will understand that Fig.10 It is only an example of the battery health status prediction device 100 and does not constitute a limitation of the battery health status prediction device 100. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components, such as input and output devices, network access devices, etc.

[0128] The processor 1001 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0129] In some embodiments, the memory 1002 may be an internal storage unit of the device 100 for predicting the battery health state, such as a hard disk or memory of the device 100 for predicting the battery health state. In other embodiments, the memory 1002 may also be an external storage device of the device 100 for predicting the battery health state, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the device 100 for predicting the battery health state. Further, the memory 1002 may also include both an internal storage unit of the device 100 for predicting the battery health state and an external storage device. The memory 1002 is used to store an operating system, an application program, a boot loader (Boot Loader), data, and other programs. The memory 1002 may also be used to temporarily store data that has been output or is to be output.

[0130] The memory 1002 stores a preset prediction model. When executing the steps in the method for predicting the battery health status, the preset prediction model is directly called from the memory 1002 to perform regression analysis on at least two expansion force extreme values ​​and the remaining power corresponding to each expansion force extreme value, so as to obtain the health status information of the battery to be detected. Of course, the preset prediction model can also be stored in a cloud server. When executing the steps in the method for predicting the battery health status, the preset prediction model is obtained from the cloud server to perform regression analysis on at least two expansion force extreme values ​​and the remaining power corresponding to each expansion force extreme value, so as to obtain the health status information of the battery to be detected.

[0131] It should be noted that the embodiment of the present application does not impose any specific restrictions on the executor of the battery health status prediction method, and the battery health status prediction device 100 can be a battery management system (BMS) or other systems.

[0132] Exemplarily, the battery health status prediction device is a BMS as an example. The BMS is connected to a pressure sensor to obtain multiple expansion force values ​​detected by the pressure sensor that vary with time. The BMS can also directly obtain multiple SOC values ​​of the battery to be detected that vary with time, or the BMS is connected to a power sensor to obtain SOC change data detected by the power sensor. The preset prediction model (including the prediction model corresponding to each battery type and the prediction model corresponding to each ambient temperature) can be pre-stored in the BMS. The BMS calls the preset prediction model to predict the SOH value of the battery to be detected based on multiple expansion force values ​​that vary with time and multiple SOC values ​​that vary with time. Of course, the preset prediction model can also be stored in the cloud server. When predicting the SOH value of the battery to be detected, the BMS obtains the preset prediction model from the cloud server and predicts the SOH value of the battery to be detected based on multiple expansion force values ​​that vary with time and multiple SOC values ​​that vary with time.

[0133] An embodiment of the present application also provides a device for predicting the health status of a battery, which includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, and the processor implements the steps in any of the above-mentioned method embodiments when executing the computer program.

[0134] The present application embodiment provides a detection component, which includes a pressure sensor, a power sensor, and Fig.10 The battery health status prediction device 100 in the embodiment of the present invention; the pressure sensor is clamped and arranged between two adjacent batteries or between the battery and the baffle, as described above Figure 1 and Figure 2 As shown; a pressure sensor is used to detect multiple expansion force values ​​that change with time during a single charging process of the battery to be tested; a power sensor is used to detect multiple remaining power values ​​that change with time during a single charging process of the battery to be tested; a battery health status prediction device is used to implement the battery health status prediction method in any of the above embodiments based on multiple expansion force values ​​that change with time and multiple remaining power values ​​that change with time.

[0135] An embodiment of the present application provides a vehicle, and the vehicle includes a detection component as described in the above embodiment.

[0136] An embodiment of the present application provides an energy storage system, and the energy storage system includes a detection component as described in the above embodiment.

[0137] An embodiment of the present application further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments can be implemented.

[0138] An embodiment of the present application provides a computer program product, which, when executed on a computer, enables the computer to execute the steps in the above-mentioned method embodiments.

[0139] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above method embodiments, which can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of each of the above method embodiments can be implemented. Among them, the computer program includes computer program code, which can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may at least include: any entity or device that can carry the computer program code to the camera / terminal device, recording medium, computer memory, ROM (Read-Only Memory), RAM (Random Access Memory), CD-ROM (Compact Disc Read-Only Memory), magnetic tape, floppy disk and optical data storage device. The computer-readable storage medium mentioned in the present application can be a non-volatile storage medium, in other words, it can be a non-transient storage medium.

[0140] It should be understood that all or part of the steps to implement the above embodiments can be implemented by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. The computer instructions can be stored in the above-mentioned computer readable storage medium.

[0141] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0142] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0143] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for predicting battery health status, characterized in that: The method comprises: Acquire multiple expansion force values ​​and multiple remaining capacities of the battery to be tested that vary with time during a single charging process, and acquire a first remaining capacities at the beginning of charging and a second remaining capacities at the end of charging during the single charging process; When the difference between the second remaining power and the first remaining power is greater than a preset depth of charge, at least two expansion force extremes are determined from the plurality of expansion force values ​​that vary with time, and the remaining power corresponding to the sampling moment of each of the at least two expansion force extremes is determined from the plurality of remaining power that vary with time; Inputting the at least two expansion force extreme values ​​and the remaining power corresponding to each of the expansion force extreme values ​​into a preset prediction model, and predicting the health status information of the battery to be tested by using the preset prediction model; The preset prediction model is used to learn the correlation between independent variables and dependent variables, the independent variables include expansion force parameters, and the dependent variables include battery discharge capacity that is positively correlated with the health status information.

2. The method according to claim 1, characterized in that The categories of the expansion force extreme values ​​include a first category, a second category, a third category and a fourth category, the remaining power corresponding to the first category, the second category, the third category and the fourth category gradually increases, the first category and the third category represent valley values, and the second category and the fourth category represent peak values; The step of determining at least two expansion force extremes from the plurality of expansion force values ​​varying with time comprises: Determining at least two initial expansion force extreme values ​​from the multiple expansion force values ​​that vary with time, and determining, from the multiple remaining capacities that vary with time, an initial remaining capacity corresponding to a sampling moment of each of the at least two initial expansion force extreme values; Determining the category of each of the initial expansion force extreme values ​​according to the interval of the initial remaining power corresponding to the sampling moment of each of the initial expansion force extreme values; The initial expansion force extreme value belonging to the preset category is used as the expansion force extreme value to obtain the at least two expansion force extreme values.

3. The method according to claim 1 or 2, characterized in that The method further comprises: Acquire the influencing attributes of the battery to be detected during a single charging process; the influencing attributes indicate factors that are relevant to the battery health status assessment, and the influencing attributes include ambient temperature and / or battery type; The step of inputting the at least two expansion force extreme values ​​and the remaining power corresponding to each of the expansion force extreme values ​​into a preset prediction model, and predicting the health status information of the battery to be tested by using the preset prediction model includes: The influencing attribute, the at least two expansion force extreme values, and the remaining power corresponding to each of the expansion force extreme values ​​are input into the preset prediction model, and the health status information of the battery to be tested is predicted by the preset prediction model; the independent variables include the ambient temperature parameter, the battery type parameter, and the expansion force parameter; Alternatively, a prediction model matching the influencing attribute is determined; the at least two expansion force extremes and the remaining power corresponding to each of the expansion force extremes are input into the prediction model matching the influencing attribute, and the health status information of the battery to be tested is predicted by the prediction model matching the influencing attribute.

4. The method according to claim 1 or 2, characterized in that: The predicting the health status information of the battery to be tested by the preset prediction model includes: Predicting the health status value of the battery to be tested by using the preset prediction model; Obtain the last predicted historical health status value of the battery to be tested and the last prediction time; Determine the preset amplitude according to the interval between the last predicted time and the current time, and the corresponding relationship between the amplitude of the health status value change and the time interval; When the health status value is smaller than the historical health status value and the difference between the health status value and the historical health status value is smaller than the preset amplitude, the health status value is used as health status information.

5. The method according to claim 1 or 2, characterized in that: Obtain expansion force change data samples, remaining power change data samples, and battery discharge capacity samples of the battery sample during the charging process and the process of discharging after charging; Determining a plurality of expansion force extreme value samples according to the expansion force change data samples; Determining, from the remaining power change data samples, a remaining power sample corresponding to a sampling moment of each of the plurality of expansion force extreme value samples; Through the initial prediction model, regression analysis is performed on the multiple expansion force extreme value samples, the remaining power samples corresponding to each of the expansion force extreme value samples, and the battery discharge capacity samples, until the health status information of the battery samples predicted by the initial prediction model meets the preset training conditions, thereby obtaining the preset prediction model.

6. The method according to claim 5, characterized in that The method further comprises: Acquire influencing attributes of the battery sample, where the influencing attributes include ambient temperature and / or battery type; The initial prediction model is used to perform regression analysis on the multiple expansion force extreme value samples, the remaining power samples corresponding to the expansion force extreme value samples, and the battery discharge capacity samples until the health status information of the battery samples predicted by the initial prediction model meets the preset training conditions, thereby obtaining the preset prediction model, including: Through the initial prediction model, a regression analysis is performed on a plurality of expansion force extreme value samples belonging to the target influencing attribute, the remaining power samples corresponding to the expansion force extreme value samples, and the battery discharge capacity samples, until the health status information of the battery samples predicted by the initial prediction model meets the preset training condition, and a preset prediction model corresponding to the target influencing attribute is obtained; the target influencing attribute is any influencing attribute; Alternatively, through the initial prediction model, regression analysis is performed on the influencing attributes, the multiple expansion force extreme value samples, the remaining power samples corresponding to each of the expansion force extreme value samples, and the battery discharge capacity samples, until the health status information of the battery samples predicted by the initial prediction model meets the preset training conditions, thereby obtaining the preset prediction model.

7. The method according to claim 1 or 2, characterized in that: The expansion force value is the total expansion force, the single point expansion force or the total expansion force in a preset detection area.

8. The method according to claim 1 or 2, characterized in that: The preset prediction model is a linear regression model, a nonlinear regression model, an interactive regression model, a decision tree regression model or a neural network regression model.

9. A battery health status prediction device, characterized in that: The battery health status prediction device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program implements the method according to any one of claims 1 to 8 when executed by the processor.

10. The device according to claim 9, characterized in that The memory stores a preset prediction model; Alternatively, the preset prediction model is obtained from a cloud server.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

12. A detection component, characterized in that: The detection component includes a pressure sensor, a power sensor, and a battery health status prediction device as claimed in claim 9 or 10; the pressure sensor is clamped between two adjacent batteries or between a battery and a baffle; The pressure sensor is used to detect multiple expansion force values ​​of the battery to be tested that vary with time during a single charging process; The power sensor is used to detect a plurality of remaining power levels of the battery to be tested that vary with time during the single charging process; The battery health status prediction device is used to implement the method described in any one of claims 1 to 8 based on the multiple expansion force values ​​that change with time and the multiple remaining capacities that change with time.

13. A vehicle, characterized in that: The vehicle includes the detection assembly of claim 12.

14. An energy storage system, characterized in that: The energy storage system comprises the detection assembly as claimed in claim 12.

Images