Battery SOC and SOH joint estimation method and device, storage medium and vehicle

Abstract

The invention relates to the technical field of battery management, in particular to a battery SOC and SOH joint estimation method and device, a storage medium and a vehicle, and the method comprises the steps that a trained estimation model constructed based on a support vector machine algorithm is acquired, the input of the estimation model is a battery operation parameter feature data set, and the output of the estimation model is battery SOC and SOH; operating parameters of the battery in the current circulation stage are collected, data preprocessing and normalization are carried out, the data are input into the estimation model, and SOC and SOH estimation results of the battery are obtained. The method is low in limitation and small in calculation amount, and the accuracy of SOC and SOH estimation of the battery can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of battery management, and in particular to a method, device, storage medium and vehicle for jointly estimating battery SOC and SOH. Background Art

[0002] With the advancement of battery energy storage system technology, the specific energy of battery energy storage has increased rapidly, and the unit cost has gradually decreased. In addition, the battery energy storage system itself has the characteristics of rapid response and convenient setting, making the application of battery energy storage system more and more extensive. In the access of new energy (such as wind power generation, solar photovoltaic power generation), it is used as an energy buffer device to smooth new energy and improve the penetration rate of new energy; in the automobile, it is used as a braking energy recovery device, and also as an auxiliary power energy provider for special working conditions (such as starting and acceleration) (generally in HEV or PHEV), or directly as the energy source of all the power of the car (i.e. pure electric vehicle). On the one hand, SOC and SOH are the core of BMS, BMS is the core of the battery, and the battery is the core of the battery energy storage system. SOC and SOH estimation are crucial to the battery energy storage system. If there is no accurate SOC and SOH, the following situations may occur: overcharge / over discharge, resulting in shortened battery life, stalling, etc.; the balanced consistency effect is not ideal, the output power is reduced, and the power performance is reduced; in order to avoid stalling, too much redundant power is set to reduce the overall energy output. Therefore, the accurate estimation of SOC and SOH is of great significance. Its algorithm is also one of the core competitive advantages of related companies.

[0003] At present, the mainstream estimation methods of battery SOC include discharge method, ampere-hour integration method, open circuit voltage method, neural network method, and Kalman filter method. The neural network method is too difficult. The Kalman filter method has been studied a lot, but the actual technical operation data is unknown. The discharge method cannot be actually used. The ampere-hour integration and open circuit voltage methods have large errors when used alone. The current mainstream method is a combination of ampere-hour integration and open circuit voltage method, which is easier to practice. The ampere-hour integration method and the open circuit voltage method have many influencing factors and large errors. The present invention uses the support vector machine algorithm (SVM) in the field of artificial intelligence to make estimates.

[0004] CN106772065A discloses a microgrid energy storage SOC estimation method and system based on least squares support vector machine, which first uses a microgrid experimental platform containing an energy storage system to measure the voltage, current, temperature and SOC data sequence of the energy storage system during the charging and discharging process; then obtains the training set and test set samples of the least squares support vector machine based on the measured data; then uses the cross-validation method to select the optimal training parameters of the least squares support vector machine training set, and then obtains the training model; then uses the test set to test the training model and evaluate the results, and finally uses the training model to estimate and determine the SOC of the corresponding type of energy storage system in the microgrid. This technical solution is a useful attempt in this field, but it takes into account the estimation and determination of the battery SOC, and does not combine the battery SOH for joint estimation and analysis. Summary of the invention

[0005] The object of the present invention is to provide a battery SOC and SOH joint estimation method, device, storage medium and vehicle, which have low limitations, small amount of calculation, and can improve the accuracy of battery SOC and SOH estimation.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] In a first aspect, the present invention provides a method for jointly estimating battery SOC and SOH, comprising:

[0008] Obtain a trained estimation model based on a support vector machine algorithm, wherein the input of the estimation model is a battery operating parameter feature data set, and the output of the estimation model is the battery SOC and SOH;

[0009] The operating parameters of the battery in the current cycle stage are collected, and the data are preprocessed and normalized, and input into the estimation model to obtain the battery SOC and SOH estimation results.

[0010] Furthermore, the battery operation parameter characteristic data set includes the discharge voltage difference in the current cycle stage, the battery temperature rise in the current cycle stage, the current at the current moment, the battery SOC at the initial moment of the current cycle stage, and the battery SOH at the initial moment of the current cycle stage.

[0011] Furthermore, the discharge voltage difference in the current cycle stage=the discharge voltage at the initial moment of the current cycle stage−the discharge voltage at the end moment of the current cycle stage.

[0012] Furthermore, the regression function of the estimation model is

[0013] In the formula, is the estimated value of SOC or SOH, ai is the Lagrange coefficient, K(x i , x) is the kernel function, b is the intercept, x i is the input value of the i-th battery operating parameter characteristic data, and C is the penalty factor.

[0014] Furthermore, the kernel function is a Gaussian kernel function, which is defined as i To a certain center x c A monotonic function of the Euclidean distance between Where δ is the width parameter of the Gaussian kernel function.

[0015] Furthermore, obtaining the trained support vector machine algorithm to construct an estimation model includes:

[0016] Designing a long cycle experiment, and obtaining battery operating parameters based on the long cycle experiment;

[0017] Based on the long cycle experiment, the battery operation parameters are obtained to construct n sample data pairs {(x i ,y i ), i = 1, 2, ..., n}, x i is the i-th input sample data, y i For x i The corresponding output values ​​are the SOC and SOH of the battery;

[0018] Divide the constructed sample data pairs into training sets and test sets;

[0019] The initial estimation model based on the support vector machine algorithm is trained using a training set. The input of the support vector machine is the characteristic data of the battery operating parameters, and the output is the battery SOC and SOH. The support vector machine establishes a mapping model between the input vector and the output vector. After a given training set input, the estimated value of the corresponding output variable under the mapping relationship can be obtained. The parameters of the model are optimized and solved through training with the training set.

[0020] After the training is completed, the estimation model based on the support vector machine algorithm is tested and analyzed using the test set, and the model optimization parameters are adjusted based on the test results until an estimation model based on the support vector machine algorithm that meets the estimation accuracy requirements is obtained.

[0021] Furthermore, the normalization is Min-Max normalization.

[0022] In a second aspect, the present invention provides a battery SOC and SOH joint estimation device, comprising:

[0023] An acquisition module is used to acquire a trained estimation model based on a support vector machine algorithm, wherein the input of the estimation model is a battery operating parameter feature data set, and the output of the estimation model is a battery SOC and SOH;

[0024] The acquisition module is used to collect the operating parameters of the battery in the current cycle stage.

[0025] The estimation module is used to preprocess and normalize the collected operating parameter data of the battery in the current cycle stage, input it into the estimation model, and obtain the battery SOC and SOH estimation results.

[0026] In a third aspect, the present invention provides a readable storage medium having a computer program stored thereon, which implements the steps of the above-mentioned estimation method when executed by a processor.

[0027] In a fourth aspect, the present invention provides a vehicle, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned estimation method when executing the program.

[0028] The present invention has the following unexpected beneficial effects:

[0029] 1. The present invention uses a trained support vector machine (SVM) algorithm to construct an estimation model to estimate the operating parameters of the battery in the current cycle stage, and obtain the battery SOC and SOH estimation results. Under the SVM algorithm, the data set is mapped to a high-dimensional feature space by a nonlinear function, and a linear regression operation is performed on it, thereby realizing the transformation of the nonlinear regression to the linear regression problem. Compared with the conventional ampere-hour integration method, the estimation method of the present invention is subject to fewer constraints, does not need to be converted into a corresponding equivalent circuit model in the actual estimation process, and does not require high initial values ​​of the estimated state. Therefore, the limitations are low and the applicability can be improved.

[0030] 2. The battery operating parameter characteristic data set described in the present invention includes the discharge voltage difference in the current cycle stage, the battery temperature rise in the current cycle stage, the current at the current moment, the battery SOC at the initial moment of the current cycle stage, and the battery SOH at the initial moment of the current cycle stage, which not only ensures the estimation accuracy, but also simplifies the extraction of characteristic quantities, reduces unnecessary calculations, and improves the estimation efficiency.

[0031] 3. The present invention utilizes the coupling relationship between SOC and SOH, and the proposed joint estimation method can make fuller use of the historical data of SOC and SOH, further improving the accuracy of estimation. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention.

[0033] Figure 1 A schematic diagram of a process of a battery SOC and SOH joint estimation method provided by an embodiment of the present invention is shown;

[0034] Figure 2 A schematic diagram of the structure of a battery SOC and SOH joint estimation device provided by an embodiment of the present invention is shown;

[0035] Figure 3 A schematic diagram of an estimation model constructed based on a support vector machine algorithm provided in an embodiment of the present invention is shown.

[0036] In the figure, 1 is an estimation device, 11 is an acquisition module, 12 is a collection module, and 13 is an estimation module. DETAILED DESCRIPTION

[0037] The following will describe the embodiments of the present invention with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention, not for limiting the scope of protection of the present invention.

[0038] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention. The drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0039] In one embodiment, see Figure 1 As shown, the present invention provides a method for jointly estimating battery SOC and SOH, comprising:

[0040] Obtain a trained estimation model based on a support vector machine algorithm, wherein the input of the estimation model is a battery operating parameter feature data set, and the output of the estimation model is the battery SOC and SOH;

[0041] The operating parameters of the battery in the current cycle stage are collected, and the data are preprocessed and normalized, and input into the estimation model to obtain the battery SOC and SOH estimation results.

[0042] The present invention uses a trained estimation model based on a support vector machine (SVM) algorithm to estimate the operating parameters of the battery in the current cycle stage, and obtains the battery SOC and SOH estimation results. Under the SVM algorithm, the data set is mapped to a high-dimensional feature space by a nonlinear function, and a linear regression operation is performed on it, thereby realizing the transformation of the nonlinear regression to the linear regression problem. Compared with the conventional ampere-hour integration method, the estimation method of the present invention is subject to fewer constraints, does not need to be converted into a corresponding equivalent circuit model in the actual estimation process, and does not require high initial values ​​of the estimated state. Therefore, the limitations are low and the applicability can be improved. At the same time, the present invention utilizes the coupling relationship between SOC and SOH, and the proposed joint estimation method can make fuller use of the historical data of SOC and SOH, further improving the accuracy of the estimation.

[0043] The ampere-hour integration method determines the input and output power by integrating the charging and discharging currents by ampere-hours, and estimates the SOC of the battery by the above means. Because this method is very subject to the initial value of the SOC, if the initial value deviation is large, the error of this method is large, and this method is an open-loop measurement, the error will accumulate over time, and will be greatly affected by noise, temperature and battery aging. The estimation method described in the present invention can use a data set, which can ensure that the optimization process starting from any initial point can eventually converge to the global optimal solution.

[0044] In a preferred embodiment, the battery operation parameter characteristic data set includes the discharge voltage difference of the current cycle stage, the battery temperature rise of the current cycle stage, the current at the current moment, the battery SOC at the initial moment of the current cycle stage, and the battery SOH at the initial moment of the current cycle stage. Further, the discharge voltage difference of the current cycle stage = the discharge voltage at the initial moment of the current cycle stage - the discharge voltage at the end moment of the current cycle stage.

[0045] The present invention simplifies the extraction of feature quantities, reduces unnecessary computational complexity, and improves estimation efficiency. If the battery voltage, current, temperature, SOC, SOH and other data sets are not processed and all used as input feature quantities, the amount of input data will be too large, increasing the difficulty of calculation. Therefore, redundant processing is performed on the input feature quantities, dimensionality reduction is performed, and important feature quantities are retained to improve estimation efficiency.

[0046] Discharge voltage difference in the current cycle stage: This parameter reflects the voltage change of the battery during the discharge process, which is related to factors such as the remaining capacity (i.e. SOC) and internal resistance of the battery. The change in discharge voltage difference can be used as an important indicator of the battery status.

[0047] Battery temperature rise during the current cycle phase: The temperature of a battery has a significant impact on its performance. Temperature rise data can help the model account for the effect of temperature on SOC and SOH estimates, especially in hot or cold conditions.

[0048] Current at the current moment: Current is a direct reflection of the battery's operating status and is particularly important for estimating SOC.

[0049] Battery SOC at the beginning of the current cycle phase: As the historical data of SOC, combined with the historical data of SOH as the input of the estimation model, it can improve the estimation accuracy.

[0050] Battery SOH at the initial moment of the current cycle stage: SOH reflects the health status of the battery, including factors such as capacity decay and internal resistance increase. The SOH value at the initial moment is used as a reference for the model to estimate subsequent SOH changes.

[0051] When building the SVM model, these features will be used as input variables, and the estimated values ​​of SOC and SOH will be used as output variables. The goal of the model is to learn the complex relationship between the input features and the output variables so that SOC and SOH can be accurately predicted when given new input data.

[0052] During the actual model training or model estimation process, ensure that all feature data are synchronized in time to accurately reflect the state of the battery at the same time. And try to collect high-quality data to avoid the impact of noise and outliers on model performance. Use an independent validation data set to evaluate the performance of the model to ensure that the model performs well on unseen data.

[0053] It should be noted that although this implementation lists multiple features, feature selection may be required in the actual modeling process to exclude irrelevant or redundant features and improve the generalization ability of the model.

[0054] Real-time: Consider the computational efficiency and response time of the model in practical applications to ensure that it can meet the needs of real-time estimation.

[0055] By comprehensively considering these factors, an accurate and reliable battery SOC and SOH joint estimation model can be constructed to provide strong support for the optimization and decision-making of the battery management system.

[0056] In a preferred embodiment, the regression function of the estimation model is

[0057] In the formula, is the estimated value of SOC or SOH, a i is the Lagrange coefficient, K(x i, x) is the kernel function, b is the intercept, x i is the input value of the i-th battery operating parameter characteristic data, and C is the penalty factor.

[0058] See also Figure 3 As mentioned above, when solving nonlinear problems, under the SVM algorithm, the data set X i The nonlinear function φ(x i ) is mapped to a high-dimensional feature space and a linear regression operation is performed on it, thereby realizing the transformation of nonlinear regression to linear regression problems. For lithium-powered battery state estimation, the input data set can be voltage, current, temperature, etc., and the output is the state of charge SOC and the state of health SOH.

[0059] Suppose the training set contains n training samples {(x i ,y i ), i = 1, 2, ..., n}, x i is the i-th input sample data, y i For x i The corresponding output values ​​are the SOC and SOH of the battery.

[0060] The corresponding fitting relationship is: f(x) = w·φ(x i )+b. Where w is the normal vector and b is the intercept.

[0061] In order to obtain the optimal coefficients of w and b, the slack variable ξ is introduced i ,

[0062] According to the support vector regression principle, linear regression can be transformed into a solvable optimization problem, namely:

[0063]

[0064] Constraints include

[0065] In the formula, C is the penalty factor, is the training error greater than ε, ξ i To make the training error smaller than ε, ε specifies the error requirement of the regression function, that is, ε = |f(x)-y|, where f(x) is the estimated value returned by the regression function and y is the corresponding true value. The smaller ε is, the smaller the error of the regression function is.

[0066] The larger the C, the greater the penalty for samples with training errors greater than ε.

[0067] In operations research, the optimal solution of nonlinear programming is related to the saddle point of the Lagrange function, so the Lagrange operator is introduced to obtain the support vector machine regression function

[0068] Furthermore, the kernel function is a Gaussian kernel function, which is defined as i To a certain center x c A monotonic function of the Euclidean distance between Where δ is the width parameter of the Gaussian kernel function.

[0069] The kernel function determines the structure and complexity of the high-dimensional feature space, which will affect the prediction accuracy of the support vector machine. Therefore, it is necessary to select a suitable kernel function from a variety of kernel functions. By adjusting the parameters of the Gaussian kernel function, the true value can be approximated with arbitrary accuracy, while having a strong anti-interference ability to noise. Therefore, the present invention selects the SVM of the Gaussian kernel function to jointly estimate the SOC and SOH of the lithium power battery.

[0070] In a preferred embodiment, obtaining a trained estimation model based on a support vector machine algorithm includes:

[0071] Designing a long cycle experiment, and obtaining battery operating parameters based on the long cycle experiment;

[0072] Based on the long cycle experiment, the battery operation parameters are obtained to construct n sample data pairs {(x i ,y i ), i = 1, 2, ..., n}, x i is the i-th input sample data, y i For x i The corresponding output values ​​are the SOC and SOH of the battery;

[0073] Divide the constructed sample data pairs into training sets and test sets;

[0074] The initial estimation model based on the support vector machine algorithm is trained using a training set. The input of the support vector machine is the characteristic data of the battery operating parameters, and the output is the battery SOC and SOH. The support vector machine establishes a mapping model between the input vector and the output vector. After a given training set input, the estimated value of the corresponding output variable under the mapping relationship can be obtained. The parameters of the model are optimized and solved through training with the training set.

[0075] After the training is completed, the estimation model based on the support vector machine algorithm is tested and analyzed using the test set, and the model optimization parameters are adjusted based on the test results until an estimation model based on the support vector machine algorithm that meets the estimation accuracy requirements is obtained.

[0076] Here is a detailed explanation of the steps in the process:

[0077] 1. Design long cycle experiments: First, a series of long cycle battery test experiments need to be designed and executed. These experiments are designed to simulate various working conditions of the battery in actual use in order to fully capture the performance changes of the battery. During the experiment, various battery operating parameters will be recorded, such as voltage, current, temperature, number of charge and discharge cycles, etc.

[0078] 2. Obtain battery operating parameters and construct sample data pairs: Based on the data of long cycle experiments, construct n sample data pairs as {(x i ,y i ), i = 1, 2, ..., n}, x i is the i-th input sample data, y i For x i The corresponding output values ​​are the battery's SOC and SOH. Each sample data pair consists of an input vector (containing multiple battery operating parameter feature data) and an output vector (containing the battery's SOC and SOH values). These sample data pairs will be used for training and support vector machine models.

[0079] 3. Divide the sample data into training set and test set: Divide the constructed sample data into training set and test set randomly or in a specific ratio. The training set is used to train the support vector machine model so that it can learn the mapping relationship between input vectors and output vectors; the test set is used to evaluate the performance of the trained model to ensure that the model has sufficient generalization ability.

[0080] 4. Use the training set to train the support vector machine model: Use the training set to train the initial estimation model based on the support vector machine algorithm. In this step, the support vector machine will try to find an optimal hyperplane (or a set of hyperplanes, for nonlinear problems) that can best classify or regress the sample points in the input space according to their output values ​​(i.e., the battery's SOC and SOH). The parameters of the model (such as the coefficients and bias terms of the hyperplane) are trained and optimized through the training set.

[0081] 5. Test and analyze the model: After training is completed, use the test set to test and analyze the estimation model based on the support vector machine algorithm. Evaluate the estimation accuracy and performance of the model by comparing the battery SOC and SOH values ​​predicted by the model with the actual values ​​in the test set. If the model performance is found to be poor, it may be necessary to adjust the optimization parameters of the model (such as kernel function type, regularization parameters, etc.) and retrain.

[0082] 6. Adjust model optimization parameters Based on the test results, continuously adjust the model optimization parameters until a model that meets the estimation accuracy requirements is obtained. This process may require multiple iterations, each of which involves retraining and testing the model.

[0083] 7. Validation and application of models: Ultimately, the fully validated model can be applied in real-world scenarios to estimate the battery’s SOC and SOH in real time, providing support for the battery management system’s decision making.

[0084] Through this preferred implementation, an efficient and accurate battery state estimation model based on the support vector machine algorithm can be constructed, providing strong support for battery health management and performance optimization.

[0085] In a preferred embodiment, the normalization is Min-Max normalization, that is, In the formula, x is the original data, min(x) is the minimum value in the data set, max(x) is the maximum value in the data set, and x new is the normalized data.

[0086] The Min-Max normalization is simple, easy to understand and implement, and the relationship between the original data is preserved after scaling.

[0087] The following is an explanation with reference to specific examples.

[0088] Embodiment 1, a method for jointly estimating battery SOC and SOH, comprising:

[0089] S1, obtaining a trained estimation model based on a support vector machine algorithm, wherein the input of the estimation model is a battery operating parameter feature data set, and the output of the estimation model is the battery SOC and SOH;

[0090] The regression function of the estimation model is

[0091] In the formula, is the estimated value of SOC or SOH, a i is the Lagrange coefficient, K(x i , x) is the Gaussian kernel function, b is the intercept, x i is the input value of the i-th battery operating parameter characteristic data, and C is the penalty factor.

[0092] The battery operation parameter characteristic data set includes the discharge voltage difference in the current cycle stage, the battery temperature rise in the current cycle stage, the current at the current moment, the battery SOC at the initial moment of the current cycle stage, and the battery SOH at the initial moment of the current cycle stage.

[0093] S2, collecting the operating parameters of the battery in the current cycle stage, performing data preprocessing and Min-Max normalization, and inputting the data into the estimation model to obtain the battery SOC and SOH estimation results.

[0094] Conduct comparative verification analysis and set up comparative example 1 and comparative example 2.

[0095] Comparative Example 1 is different from the method described in the present invention in that the kernel function of the regression function is a non-Gaussian kernel function.

[0096] Comparative Example 2 is different from the method of the present invention in that the input parameters of the estimation model are voltage, current and temperature.

[0097] See Table 1 for the estimation results of Example 1, Comparative Example 1 and Comparative Example 2 for a certain power battery.

[0098] Table 1 Estimation results of different methods

[0099]

[0100] The absolute error is calculated as…actual value – estimated value….

[0101] It can be seen from Table 1 that the absolute error of the estimated value obtained by the method of the present invention is lower than that of Comparative Example 1 and Comparative Example 2. It can be seen that the present invention improves the accuracy of the estimation result by limiting the regression function, kernel function of the estimation model and the specific category of the battery operating parameter characteristic data set, so that the estimation result is closer to the actual value.

[0102] In one embodiment, see Figure 2 As shown, the present invention provides a battery SOC and SOH joint estimation device, and the estimation device 1 includes an acquisition module 11, a collection module 12 and an estimation module 13. The acquisition module 11 is used to acquire a trained estimation model based on a support vector machine algorithm, the input of the estimation model is a battery operating parameter feature data set, and the output of the estimation model is the battery SOC and SOH. The acquisition module 12 is used to collect the operating parameters of the battery in the current cycle stage. The estimation module 13 is used to preprocess and normalize the collected operating parameter data of the battery in the current cycle stage, and input it into the estimation model to obtain the battery SOC and SOH estimation results.

[0103] In one embodiment, the present invention provides a readable storage medium having a computer program stored thereon, which implements the steps of the above estimation method when executed by a processor.

[0104] In one embodiment, the present invention provides a vehicle, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above estimation method when executing the program.

[0105] The above embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Any equivalent substitution or change made by a person skilled in the art based on the present invention is within the protection scope of the present invention.

Claims

1. A method for jointly estimating battery SOC and SOH, characterized in that: include: Obtain a trained estimation model based on a support vector machine algorithm, wherein the input of the estimation model is a battery operating parameter feature data set, and the output of the estimation model is the battery SOC and SOH; The operating parameters of the battery in the current cycle stage are collected, and the data are preprocessed and normalized, and input into the estimation model to obtain the battery SOC and SOH estimation results.

2. The battery SOC and SOH combined estimation method according to claim 1, characterized in that: The battery operation parameter characteristic data set includes the discharge voltage difference in the current cycle stage, the battery temperature rise in the current cycle stage, the current at the current moment, the battery SOC at the initial moment of the current cycle stage, and the battery SOH at the initial moment of the current cycle stage.

3. The method for jointly estimating battery SOC and SOH according to claim 2, characterized in that: The discharge voltage difference of the current cycle stage=the discharge voltage at the initial moment of the current cycle stage-the discharge voltage at the end moment of the current cycle stage.

4. The battery SOC and SOH combined estimation method according to claim 1, characterized in that: The regression function of the estimation model is In the formula, is the estimated value of SOC or SOH, a i is the Lagrange coefficient, K(x i , x) is the kernel function, b is the intercept, x i is the input value of the i-th battery operating parameter characteristic data, and C is the penalty factor.

5. The method for jointly estimating battery SOC and SOH according to claim 4, characterized in that: The kernel function is a Gaussian kernel function.

6. The battery SOC and SOH joint estimation method according to claim 1, characterized in that: Obtaining the trained support vector machine algorithm to build an estimation model includes: Designing a long cycle experiment, and obtaining battery operating parameters based on the long cycle experiment; Based on the long cycle experiment, the battery operation parameters are obtained to construct n sample data pairs {(x i ,y i ), i = 1, 2, ..., n}, x i is the i-th input sample data, y i For x i The corresponding output values ​​are the SOC and SOH of the battery; Divide the constructed sample data pairs into training sets and test sets; The initial estimation model based on the support vector machine algorithm is trained using a training set. The input of the support vector machine is the characteristic data of the battery operating parameters, and the output is the battery SOC and SOH. The support vector machine establishes a mapping model between the input vector and the output vector. After a given training set input, the estimated value of the corresponding output variable under the mapping relationship can be obtained. The parameters of the model are optimized and solved through training with the training set. After the training is completed, the estimation model based on the support vector machine algorithm is tested and analyzed using the test set, and the model optimization parameters are adjusted based on the test results until an estimation model based on the support vector machine algorithm that meets the estimation accuracy requirements is obtained.

7. The battery SOC and SOH combined estimation method according to claim 1, characterized in that: The normalization is Min-Max normalization.

8. A battery SOC and SOH joint estimation device, characterized in that: include: An acquisition module is used to acquire a trained estimation model based on a support vector machine algorithm, wherein the input of the estimation model is a battery operating parameter feature data set, and the output of the estimation model is a battery SOC and SOH; A collection module is used to collect the operating parameters of the battery in the current cycle stage; The estimation module is used to preprocess and normalize the collected operating parameter data of the battery in the current cycle stage, input it into the estimation model, and obtain the battery SOC and SOH estimation results.

9. A readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the estimation method according to any one of claims 1 to 7 are implemented.

10. A vehicle, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the estimation method according to any one of claims 1 to 7 when executing the program.

Images