SOC estimation method and device, model training method and device, vehicle and storage medium

Abstract

The invention relates to an SOC estimation method and device, a model training method and device, a vehicle and a storage medium, and relates to the technical field of data. The method comprises the following steps: acquiring charging data of a battery of a target vehicle; according to the charging data, determining voltage change data and battery characteristic parameters when current jumps in the charging process of the battery; and determining an SOC estimation result corresponding to the charging data according to the change data, the battery characteristic parameters and a target SOC estimation model. Therefore, the SOC of the battery can be accurately evaluated.

Description

Technical Field

[0001] The present application relates to the field of data technology, in particular to the field of machine learning technology, and specifically to a SOC estimation method, a model training method, a device, a vehicle and a storage medium. Background Art

[0002] The battery state of charge (SOC), as a core indicator of the power battery management system, its accurate estimation is of vital importance for the optimization and control of the charging efficiency and discharge performance of electric vehicles.

[0003] SOC is not only directly related to the range and dynamic response capabilities of electric vehicles, but also the key to ensuring the user's driving experience and preventing adverse situations such as power attenuation or sudden interruption caused by misjudgment of power. Therefore, accurate estimation of SOC is of great value in improving the overall performance and reliability of electric vehicles and ensuring that users enjoy stable and efficient travel services.

[0004] Related technology CN105116343B proposes a power battery SOC estimation method based on least squares support vector machine. It uses the power battery circuit model and related identification parameters to fit the mapping relationship between the open circuit voltage Uoc and SOC, and then uses the LS-SVM model for data training to correct and compensate the SOC obtained by ampere-hour integration, thereby reducing the fitting error and the cumulative error of ampere-hour integration.

[0005] Related technology CN106154168A proposes a multi-algorithm fusion SOC estimation method for power batteries based on Bayesian probability. This method first constructs the corresponding equivalent circuit model and simultaneously identifies the parameters of the model. Subsequently, multiple observers for predicting the battery terminal voltage and SOC are constructed based on a variety of estimation algorithms. In practical applications, based on the residuals between the predicted and measured terminal voltages at different times, the SOC results output by multiple observers are weighted and accumulated in combination with the weighted values ​​assigned by Bayesian probability to obtain the final fused SOC estimation result.

[0006] However, none of the above-mentioned related technologies fully considers the relationship between the characteristics of the battery and the SOC, making it difficult to accurately evaluate the SOC of the battery. Summary of the invention

[0007] The present application provides a SOC estimation method, a model training method, a device, a vehicle and a storage medium to at least solve the technical problem that the SOC estimation method in the related art is difficult to accurately evaluate the SOC of the battery. The technical solution of the present application is as follows:

[0008] According to the first aspect provided by the present application, a SOC estimation method is provided, including: obtaining charging data of a battery of a target vehicle; determining, based on the charging data, voltage change data and battery characteristic parameters when a current jump occurs during the charging process of the battery; and determining, based on the change data, the battery characteristic parameters and a target SOC estimation model, an SOC estimation result corresponding to the charging data.

[0009] According to the above technical means, the present application can more accurately estimate the SOC of the battery through the SOC estimation model by carefully analyzing the current and voltage changes during the battery charging process, as well as the characteristic parameters of the battery itself, thereby providing the driver with more reliable cruising range information.

[0010] In one possible manner, the change data includes: determining multiple current mutation points in the charging data; for each current mutation point in the multiple current mutation points, determining the change data corresponding to the current mutation point; the change data includes voltage characteristics corresponding to multiple time frames after the voltage mutation point.

[0011] In one possible manner, determining the battery characteristic parameters includes: extracting features from charging data to determine the battery characteristic parameters corresponding to a plurality of time frames in the charging data.

[0012] In one possible manner, the SOC estimation model includes a first model and a second model, and determines the SOC estimation result corresponding to the charging data according to the change data, the battery characteristic parameters and the target SOC estimation model, including: inputting the change data into the first model to obtain a first result, and inputting the battery characteristic parameters into the second model to obtain a second result; and performing weighted integration on the first result and the second result to obtain the SOC estimation result.

[0013] In one possible method, the charging data includes the starting charging temperature during the battery charging process; the SOC estimation result corresponding to the charging data is determined based on the change data, the battery characteristic parameters and the target SOC estimation model, including: determining the temperature range corresponding to the starting charging temperature of the charging data; determining the target SOC estimation model corresponding to the temperature range from multiple preset SOC estimation models; one preset SOC estimation model corresponds to one temperature range; the change data and battery characteristic parameters in the charging data are input into the target SOC estimation model to obtain the SOC estimation result corresponding to the charging data.

[0014] According to the second aspect provided by the present application, a model training method is provided, including: obtaining multiple sample charging data of a battery of a target vehicle; for each sample charging data in the multiple sample charging data, determining first characteristic data and second characteristic data of the sample charging data; the first characteristic data includes voltage change data when a current jump occurs during battery charging, and SOC values ​​corresponding to multiple voltages when the voltage changes; the second characteristic data includes battery characteristic parameters during battery charging, and SOC values ​​corresponding to the voltage characteristic parameters; the first characteristic data and the second characteristic data are trained according to a preset algorithm to obtain an SOC estimation model.

[0015] According to the above technical means, the present application can enable the model to learn the complex relationship between the influence of current fluctuation on the battery SOC by considering the voltage change data and its corresponding SOC value (first characteristic data) when the current jumps during the battery charging process, and the second characteristic data includes a variety of characteristic parameters of the battery during the charging process and their corresponding SOC values. These parameters may include battery temperature, internal resistance, voltage change trend, etc., which can jointly reflect the internal state of the battery. Training the model with these parameters can improve the accuracy of SOC estimation. Therefore, the present application can estimate the SOC of the battery with a high accuracy.

[0016] In a possible manner, determining first characteristic data of sample charging data includes: determining multiple current mutation points in the sample charging data; and performing feature extraction on data corresponding to the multiple current mutation points to obtain the first characteristic data of the sample charging data.

[0017] According to the above technical means, the present application can more accurately understand the dynamic characteristics of the charging process by identifying the current mutation points, so as to support the subsequent training of the model so that the model can learn the complex relationship between the impact of current fluctuations on the battery SOC.

[0018] In one possible approach, feature extraction is performed on data corresponding to multiple current mutation points to obtain first feature data of sample charging data; for each current mutation point among the multiple current mutation points, voltage change data corresponding to the current mutation point is determined; the voltage change data includes voltage features corresponding to multiple time frames after the voltage mutation point; the SOC value corresponding to each of the multiple time frames after the voltage mutation point is determined; and the first feature data is determined based on the voltage change data corresponding to the current mutation point and the SOC values ​​corresponding to each of the multiple time frames after the voltage mutation point.

[0019] According to the above-mentioned technical means, the present application can accurately determine the voltage dynamic characteristics of the charging process by identifying the current mutation point and determining the voltage change data corresponding to the current mutation point for each of multiple current mutation points, so as to support the subsequent training of the model, so that the model can learn the complex relationship between the impact of current fluctuations on the battery SOC.

[0020] In one possible method, determining the second characteristic data of the sample charging data includes: performing feature extraction on the sample charging data to determine a set of battery characteristic parameters corresponding to the sample charging data; the battery characteristic set includes battery characteristic parameters corresponding to each of a plurality of time frames in the sample charging data; determining an SOC value corresponding to each of the plurality of time frames; and determining the second characteristic data based on the battery characteristic set and the SOC value corresponding to each battery characteristic in the battery characteristic set.

[0021] According to the above-mentioned technical means, the present application can comprehensively evaluate the performance of the battery in different charging stages through a set of battery characteristic parameters and corresponding SOC values, and combine the data of multiple time frames to analyze the changing trend of battery performance over time, so that the model can learn the corresponding relationship between battery characteristics and battery SOC.

[0022] In one possible method, the first feature data and the second feature data are trained according to a preset algorithm to obtain a SOC estimation model, including: training the first feature data according to a first preset algorithm to obtain a first model, and training the second feature data according to a second preset algorithm to obtain a second model; and performing weighted integration of the first model and the second model to obtain the SOC estimation model.

[0023] According to the above technical means, the present application can improve the accuracy of SOC estimation by weighted integration of the first model and the second model corresponding to the first characteristic data (such as the current mutation point and its voltage change data) and the second characteristic data (such as the battery characteristic parameter set and the corresponding SOC value).

[0024] In one possible approach, for each sample charging data among multiple sample charging data, the first characteristic data and the second characteristic data of the sample charging data are divided into a corresponding temperature set based on the starting charging temperature corresponding to the sample charging data; according to a preset algorithm, the first characteristic data and the second characteristic data in the temperature set are trained to obtain an SOC estimation model corresponding to the temperature set.

[0025] According to the above technical means, the present application divides the data by temperature and can train a dedicated SOC estimation model for each temperature range, thereby more accurately considering the impact of temperature on battery performance.

[0026] According to the third aspect provided by the present application, a SOC estimation device is provided, the device including: an acquisition unit, a determination unit and a processing unit; the acquisition unit is used to acquire the charging data of the battery of the target vehicle; the determination unit is used to determine the voltage change data and battery characteristic parameters when the current jumps during the charging process of the battery based on the charging data; the processing unit is used to determine the SOC estimation result corresponding to the charging data based on the change data, the battery characteristic parameters and the target SOC estimation model.

[0027] In one possible manner, the determination unit is specifically used to: determine multiple current mutation points in the charging data; for each current mutation point in the multiple current mutation points, determine the change data corresponding to the current mutation point; the change data includes voltage characteristics corresponding to multiple time frames after the voltage mutation point.

[0028] In a possible manner, the determination unit is specifically used to: extract features from the charging data, and determine battery feature parameters corresponding to each of a plurality of time frames in the charging data.

[0029] In one possible manner, the determination unit is specifically used to: input the change data into a first model to obtain a first result, and input the battery characteristic parameter into a second model to obtain a second result; and perform weighted integration on the first result and the second result to obtain an SOC estimation result.

[0030] In one possible manner, the determination unit is specifically used to: determine the temperature range corresponding to the starting charging temperature of the charging data; determine the target SOC estimation model corresponding to the temperature range from multiple preset SOC estimation models; one preset SOC estimation model corresponds to one temperature range; input the change data and battery characteristic parameters in the charging data into the target SOC estimation model to obtain the SOC estimation result corresponding to the charging data.

[0031] In one possible manner, the device further includes: a training unit, the training unit being configured to train the target SOC estimation model based on the change data, the battery characteristic parameters and the SOC estimation result to obtain a trained target SOC estimation model.

[0032] According to the fourth aspect provided by the present application, a model training device is provided, including: an acquisition unit, a determination unit and a training unit; the acquisition unit is used to acquire multiple sample charging data of a battery of a target vehicle; the determination unit is used to determine, for each sample charging data in the multiple sample charging data, first characteristic data and second characteristic data of the sample charging data; the first characteristic data includes voltage change data when current jumps during battery charging, and SOC values ​​corresponding to multiple voltages when the voltage changes; the second characteristic data includes battery characteristic parameters during battery charging, and SOC values ​​corresponding to voltage characteristic parameters; the training unit is used to train the first characteristic data and the second characteristic data according to a preset algorithm to obtain an SOC estimation model.

[0033] In a possible manner, the determination unit is specifically used to: determine a plurality of current mutation points in the sample charging data; and perform feature extraction on data corresponding to the plurality of current mutation points to obtain first feature data of the sample charging data.

[0034] In one possible manner, the determination unit is specifically used to: determine, for each current mutation point among multiple current mutation points, voltage change data corresponding to the current mutation point; the voltage change data includes voltage characteristics corresponding to multiple time frames after the voltage mutation point; determine the SOC value corresponding to each of the multiple time frames after the voltage mutation point; and determine the first characteristic data based on the voltage change data corresponding to the current mutation point and the SOC value corresponding to each of the multiple time frames after the voltage mutation point.

[0035] In one possible manner, the determination unit is specifically used to: perform feature extraction on the sample charging data to determine a set of battery feature parameters corresponding to the sample charging data; the battery feature set includes battery feature parameters corresponding to each of multiple time frames in the sample charging data; determine the SOC value corresponding to each of the multiple time frames; and determine the second feature data based on the battery feature set and the SOC value corresponding to each battery feature in the battery feature set.

[0036] In one possible manner, the training unit is specifically used to: train the first feature data according to a first preset algorithm to obtain a first model, and train the second feature data according to a second preset algorithm to obtain a second model; and perform weighted integration of the first model and the second model to obtain an SOC estimation model.

[0037] In one possible manner, the training unit is specifically used for: also for dividing the first characteristic data and the second characteristic data of the sample charging data into corresponding temperature sets for each sample charging data in multiple sample charging data based on the starting charging temperature corresponding to the sample charging data; and training the first characteristic data and the second characteristic data in the temperature set according to a preset algorithm to obtain an SOC estimation model corresponding to the temperature set.

[0038] According to the fifth aspect provided by the present application, a processing device is provided, including: a processor; a memory for storing processor executable instructions; wherein the processor is configured to execute instructions to implement the method of the above-mentioned first aspect and any possible implementation method thereof.

[0039] According to the sixth aspect provided by the present application, a computer-readable storage medium is provided. When the instructions in the computer-readable storage medium are executed by a processor of a processing device, the processing device is enabled to execute the method in the above-mentioned first aspect and any possible implementation method thereof.

[0040] According to the seventh aspect provided by the present application, a computer program product is provided, the computer program product comprising computer instructions, and when the computer instructions are executed on a processing device, the processing device executes the method of the above-mentioned first aspect and any possible implementation manner thereof.

[0041] Therefore, the above technical features of the present application have the following beneficial effects:

[0042] (1) By considering the voltage change data and its corresponding SOC value (first characteristic data) when the current jumps during the battery charging process, the model can learn the complex relationship between the current fluctuation and the battery SOC, and the second characteristic data includes a variety of characteristic parameters of the battery during the charging process and their corresponding SOC values. These parameters may include battery temperature, internal resistance, voltage change trend, etc., which can jointly reflect the internal state of the battery. Training the model with these parameters can improve the accuracy of SOC estimation. Therefore, the present application can estimate the SOC of the battery with a high accuracy.

[0043] (2) By identifying the current mutation points, the dynamic characteristics of the charging process can be understood more accurately to support the subsequent training of the model, so that the model can learn the complex relationship between the impact of current fluctuations on battery SOC.

[0044] (3) By identifying the current mutation point and determining the voltage change data corresponding to each current mutation point among multiple current mutation points, the voltage dynamic characteristics of the charging process can be accurately determined to support the subsequent model training so that the model can learn the complex relationship between the current fluctuation and the battery SOC.

[0045] (4) The battery characteristic parameter set and the corresponding SOC value can be used to comprehensively evaluate the performance of the battery at different charging stages. By combining the data of multiple time frames, the changing trend of battery performance over time can be analyzed, so that the model can learn the corresponding relationship between battery characteristics and battery SOC.

[0046] (5) The accuracy of SOC estimation can be improved by weighted integration of the first model and the second model corresponding to the first characteristic data (such as the current mutation point and its voltage change data) and the second characteristic data (such as the battery characteristic parameter set and the corresponding SOC value).

[0047] (6) By partitioning the data by temperature, a dedicated SOC estimation model can be trained for each temperature range, thereby more accurately considering the impact of temperature on battery performance.

[0048] It should be noted that the technical effects brought about by any implementation method in the second to seventh aspects can refer to the technical effects brought about by the corresponding implementation method in the first aspect, and will not be repeated here.

[0049] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present application, and together with the specification are used to explain the principles of the present application, and do not constitute improper limitations on the present application.

[0051] Figure 1 is a structural schematic diagram of a SOC estimation system according to an exemplary embodiment;

[0052] Figure 2 is a flow chart of a model training method according to an exemplary embodiment;

[0053] Figure 3 is a schematic diagram of extracting a voltage response segment according to an exemplary embodiment;

[0054] Figure 4 is a flow chart of a SOC estimation method according to an exemplary embodiment;

[0055] Figure 5 is a schematic diagram of a model training process according to an exemplary embodiment;

[0056] Figure 6 is a schematic diagram showing a SOC estimation process according to an exemplary embodiment;

[0057] Figure 7 is a schematic diagram of a model deployment process according to an exemplary embodiment;

[0058] Figure 8 is a block diagram of a model training device according to an exemplary embodiment;

[0059] Fig. 9 is a block diagram of a SOC estimation device according to an exemplary embodiment;

[0060] Fig.10 It is a block diagram of a processing device according to an exemplary embodiment. DETAILED DESCRIPTION

[0061] In order to enable ordinary persons in the art to better understand the technical solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0062] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims.

[0063] As the background technology, in order to solve the problem of low SOC estimation accuracy in the related art, the present application provides a model training method and an SOC estimation method. The model training method includes obtaining multiple sample charging data of the battery of the target vehicle, and then determining the first characteristic data and the second characteristic data of the sample charging data for each sample charging data in the multiple sample charging data, so as to further train the first characteristic data and the second characteristic data according to a preset algorithm to obtain an SOC estimation model. The SOC estimation method can accurately evaluate the SOC of the battery according to the SOC estimation model.

[0064] like Figure 1 As shown, Figure 1 A SOC estimation system is provided for an embodiment of the present application. The SOC estimation system may include a processing device 101, a model training device 102, and a SOC estimation device 103. The processing device 101 and the model training device 102 may be communicatively connected. The processing device 101 and the SOC estimation device 103 may be communicatively connected. For example, the connection may be made by wired or wireless means. Among them, the wireless means may include Bluetooth, wireless fidelity (wireless fidelity, WIFI), a communication network (such as the fifth generation (5th generation, 5G), etc. The wired means may include a network cable, a system bus, etc.

[0065] The model training device 102 and the SOC estimation device 103 may be integrated into one module of the processing device 101 .

[0066] The model training device 102 can obtain multiple sample charging data of the battery of the target vehicle, and determine the first characteristic data and the second characteristic data of the sample charging data for each sample charging data in the multiple sample charging data, so as to further train the first characteristic data and the second characteristic data according to a preset algorithm to obtain an SOC estimation model.

[0067] The SOC estimation device 103 can obtain the charging data of the battery of the target vehicle and determine the first characteristic data and the second characteristic data of the charging data, so as to further input the first characteristic data and the second characteristic data into the target SOC estimation model to obtain the SOC estimation result corresponding to the charging data.

[0068] For ease of understanding, the model training method and SOC estimation method provided in this application are specifically introduced below with reference to the accompanying drawings.

[0069] Figure 2 is a flow chart of a model training method according to an exemplary embodiment. Figure 2 As shown, the model training method includes the following steps:

[0070] S201. The processing device obtains a plurality of sample charging data of a battery of a target vehicle.

[0071] In one possible manner, the processing device may obtain multiple sample charging data of the battery of the target vehicle, wherein the sample charging data may be complete sample charging data, i.e., sample charging data when the maximum single cell voltage reaches a preset cut-off voltage threshold during charging, or when the SOC displayed by the vehicle instrument reaches 100% when charging is completed.

[0072] Optionally, the sample charging data may be data of other batteries of the same type as the battery of the target vehicle when they are actually working, or may be data of the battery of the target vehicle when they are actually working. This application does not impose any specific restrictions on this.

[0073] Optionally, the number of time frames included in the sample charging data may be set according to actual needs. For example, the number of time frames may be not less than 100, or not less than 200. This application does not impose any specific restrictions on this.

[0074] Optionally, the duration of the sample charging data can be set according to actual needs. The duration can be no less than 20 minutes, or no less than 10 minutes. This application does not make specific restrictions on this.

[0075] In one possible manner, the sample charging data includes a current mutation point, a terminal time, a charging state, a current, a voltage list, a temperature list, and the like.

[0076] In a possible manner, the processing device may divide the sample charging data into a plurality of sample charging data sets according to the charging start temperature. For example, the present application may divide the sample charging data into a set according to 5 degrees Celsius.

[0077] Optionally, the temperature range involved in the sample charging data can be set according to actual needs. For example, the temperature range can be -40 degrees Celsius to 50 degrees Celsius, or -60 degrees Celsius to 60 degrees Celsius. This application does not make specific restrictions on this.

[0078] In one possible approach, the processing device may clean the common jump (or mutation) data in the voltage and current in the sample charging data, thereby ensuring the data quality and improving the accuracy of subsequent analysis.

[0079] Optionally, the processing device may use a Savitzky-Golay filter to clean the sample charging data. The Savitzky-Golay filter is used for data smoothing to remove random noise in the data while retaining important features of the data, such as peak values, valley values, and signal width, to the maximum extent.

[0080] S202: The processing device determines, for each sample charging data among a plurality of sample charging data, first characteristic data and second characteristic data of the sample charging data.

[0081] The first characteristic data includes voltage change data when the current jumps during the battery charging process, and SOC values ​​corresponding to multiple voltages when the voltage changes. The second characteristic data includes battery characteristic parameters during the battery charging process.

[0082] In one possible manner, the processing device may determine multiple current mutation points in the sample charging data. The processing device may perform feature extraction on the data corresponding to the multiple current mutation points to obtain first feature data of the sample charging data. The specific implementation method of the processing device determining the first feature data may refer to S301-S302 below. No further description is given here.

[0083] In another possible manner, the processing device may perform feature extraction on the sample charging data to determine a set of battery feature parameters corresponding to the sample charging data. The processing device may determine the SOC values ​​corresponding to each of the batteries in multiple time frames. The processing device may determine the second feature data based on the battery feature set and the SOC value corresponding to each battery feature in the battery feature set.

[0084] Among them, the second characteristic data may include but is not limited to the following items: maximum value of single cell voltage, minimum value of single cell voltage, average value of single cell voltage, single cell voltage difference, current, current difference, charging capacity, and charging capacity difference.

[0085] In a possible manner, the processing device may divide the first characteristic data and the second characteristic data of the sample charging data into corresponding temperature sets for each sample charging data among the plurality of sample charging data based on the starting charging temperature corresponding to the sample charging data.

[0086] S203: The processing device trains the first characteristic data and the second characteristic data according to a preset algorithm to obtain an SOC estimation model.

[0087] In a possible manner, the processing device may train the first characteristic data and the second characteristic data in the temperature set according to a preset algorithm to obtain an SOC estimation model corresponding to the temperature set.

[0088] In one example, sample charging data of 5 degrees Celsius to 10 degrees Celsius may be in a set. The processing device may train the first characteristic data and the second characteristic data in the temperature set of 5 degrees Celsius to 10 degrees Celsius according to a preset algorithm to obtain an SOC estimation model corresponding to the temperature set of 5 degrees Celsius to 10 degrees Celsius.

[0089] In one possible way, the processing device can split the multiple sample charging data into training sample data and test sample data according to preset notes. The processing device can perform model training based on the training sample data, and the model can learn the characteristics and patterns of the first feature data and the second feature data corresponding to different SOC values. In order to improve the performance of the model and avoid overfitting, the processing device uses cross-validation (such as K-fold cross-validation) and grid search strategies to find the optimal parameter configuration of each algorithm.

[0090] Optionally, the preset algorithm can be set according to actual needs. For example, the preset algorithm can be one or more of a neural network algorithm, a random forest regression algorithm, a support vector machine regression algorithm, an XGBoost regression algorithm, and an LSTM algorithm. This application does not make specific restrictions on this.

[0091] In one possible manner, the processing device may input the test set data into the trained SOC estimation model, and output the SOC estimation value according to the input first feature data and second feature data. The processing device may use the root mean square error (RMSE) as an evaluation index to calculate the deviation between the model predicted SOC and the actual SOC label. RMSE satisfies the following formula 1:

[0092]

[0093] in, is the estimated SOC value calculated by the model for the i-th time. SOC label is the SOC value corresponding to the characteristic data input for the i-th time. With SOC label The RMSE between .

[0094] In one possible manner, the processing device may train the first feature data according to a first preset algorithm to obtain a first model, and train the second feature data according to a second preset algorithm to obtain a second model. The processing device performs weighted integration on the first model and the second model to obtain the SOC estimation model.

[0095] Optionally, the first preset algorithm may be one or more of a neural network algorithm, a random forest regression algorithm, a support vector machine regression algorithm, an XGBoost regression algorithm, and an LSTM algorithm. The second preset algorithm may be one or more of a neural network algorithm, a random forest regression algorithm, a support vector machine regression algorithm, an XGBoost regression algorithm, and an LSTM algorithm. This application does not make specific restrictions on this.

[0096] Optionally, the weighted integration method can be set according to actual needs. For example, the weighted integration method can be mean weighting, weight weighting, or linear regression weighting. This application does not make specific restrictions on this.

[0097] In another possible embodiment, the processing device may perform weighted integration on the first model and the second model to obtain an SOC estimation model, or the processing device may perform weighted summation on the result output by the first model and the result output by the second model based on set weights to obtain an SOC estimation value.

[0098] based on Figure 2 In the technical solution, this application can consider the voltage change data and its corresponding SOC value (first characteristic data) when the current jumps during the battery charging process, so that the model can learn the complex relationship between the current fluctuation and the battery SOC, and the second characteristic data includes a variety of characteristic parameters of the battery during the charging process and their corresponding SOC values. These parameters may include battery temperature, internal resistance, voltage change trend, etc., which can jointly reflect the internal state of the battery. Training the model with these parameters can improve the accuracy of SOC estimation. Therefore, this application can estimate the SOC of the battery with a high accuracy.

[0099] In some embodiments, in order to determine the first characteristic data of the sample charging data, the model training method provided in the embodiment of the present application also includes the following steps: S301-S302.

[0100] S301: The processing device determines a plurality of current mutation points in sample charging data.

[0101] In a possible manner, the processing device may detect the sample charging data to determine a plurality of current mutation points in the sample charging data.

[0102] In one example, the sample charging data may be an ordered current sequence I = [I0, I1, I2, I3, I4, I5, ...], where In represents the current value measured at time point n. In order to analyze the change in current value, the processing device may calculate the difference between adjacent current values ​​to obtain a differential sequence Idiff = [Idiff0, Idiff1, Idiff2, Idiff3, Idiff4, ...], where Idiffk = Ik+1-Ik. In order to exclude non-significant changes caused by measurement errors or slight fluctuations in the system, the processing device may set an upper threshold and a lower threshold. By comparing the differential result Idiffk with these two thresholds, the differential results whose absolute values ​​are less than the lower threshold or greater than the upper threshold are deleted to obtain a filtered differential sequence Idiff-i.

[0103] In one possible manner, the processing device may calculate the standard deviation Idiff-std of the absolute value of the filtered differential sequence Idiff-i. The processing device may traverse the filtered differential sequence Idiff-i, and if the absolute value of a differential value |Idiff-ik| is greater than the calculated standard deviation Idiff-std, it is determined that a current mutation occurs at time point k+1 (note the corresponding relationship between the time points of the differential sequence and the original current sequence), and the point is marked as a current mutation point.

[0104] S302: The processing device extracts features from data corresponding to a plurality of current mutation points to obtain first feature data of sample charging data.

[0105] In one possible manner, the processing device may determine the voltage change data corresponding to the current mutation point for each of the multiple current mutation points. The processing device may determine the SOC values ​​corresponding to each of the multiple time frames after the voltage mutation point. The processing device may determine the first characteristic data based on the voltage change data corresponding to the current mutation point and the SOC values ​​corresponding to each of the multiple time frames after the voltage mutation point.

[0106] The voltage change data includes voltage characteristics corresponding to multiple time frames after the voltage mutation point.

[0107] In one possible manner, the processing device may detect the charging data to determine a voltage response segment after a sudden current change occurs, that is, a segment in which the voltage continues to change.

[0108] In one example, Figure 3As shown, Figure 3 It is a schematic diagram of extracting a voltage response segment. The processing device can determine the time t0 of the current mutation point and the current values ​​I0 and I1 before and after the mutation, and record the voltage Usum0 before the mutation. The processing device can continuously measure the voltage Usum(t) starting from time t0, and calculate the voltage difference Usum-diff-i=Usum(ti+1)-Usum(ti), where ti+1 is the next measurement time point after ti. The processing device can determine that the voltage response ends when the sign of the total voltage difference changes, and record the time interval of the voltage response as [t0,tm].

[0109] In another example, the processing device can perform curve fitting and data reconstruction on the voltage response segment to obtain a complete voltage matrix [U0, U1, U2, ..., Um]. Where m represents the time, and Um represents the voltage vector at time tm. The processing device can perform feature extraction on the voltage matrix [U0, U1, U2, ..., Um] to determine the features corresponding to each voltage vector.

[0110] Optionally, the features corresponding to the voltage vector include but are not limited to the following: current value before response, response current value, response current change value, total voltage change trend, total voltage change amount, maximum voltage change trend, and maximum voltage change amount.

[0111] In a possible manner, the processing device may determine the SOC values ​​corresponding to each of a plurality of time frames after the voltage mutation point, that is, the SOC values ​​corresponding to a plurality of voltage vectors after the voltage mutation point.

[0112] In one example, the processing device can calculate the SOC values ​​corresponding to multiple voltage vectors based on the ampere-hour integration. The processing device can use the time before the full charge cutoff as the starting point for the ampere-hour integration calculation, the battery rated capacity as Q, and the current sequence [I0, I1, I2]. When the current is I2, the battery reaches the full charge cutoff point, and the corresponding time difference sequence of the calculated current sequence is [tdiff-0-1, tdiff-1-2. The SOC0 at I0 calculated by the processing device can satisfy the following formula 2:

[0113]

[0114] The SOC1 at I1 can satisfy the following formula 3:

[0115]

[0116] Among them, SOC 0 It can be used to characterize the SOC value at I0. 1 It can be used to characterize the SOC value at I1. 2 It can be used to represent the fully charged SOC value, which is 100.

[0117] In a possible manner, the processing device may associate the voltage change data corresponding to the current mutation point with the SOC values ​​corresponding to each of a plurality of time frames after the voltage mutation point based on the time frame to obtain the first characteristic data.

[0118] In one example, the processing device can associate the voltage change data corresponding to the current mutation point and the SOC values ​​corresponding to multiple time frames after the voltage mutation point based on the time frame to form an m×n feature data matrix, where m is the number of sample data, and n is the voltage change data corresponding to n-1 current mutation points and the SOC values ​​corresponding to multiple time frames after the voltage mutation point.

[0119] Based on this, by identifying the current mutation points, the dynamic characteristics of the charging process can be understood more accurately to support the subsequent training of the model so that the model can learn the complex relationship between the impact of current fluctuations on battery SOC.

[0120] Figure 4 is a flow chart of a SOC estimation method according to an exemplary embodiment. Figure 4 As shown, the SOC estimation method includes the following steps: S401-S403.

[0121] S401. The processing device obtains charging data of a battery of a target vehicle.

[0122] In one possible manner, the processing device may continuously monitor the battery of the target vehicle while the target vehicle is running. The processing device may obtain charging data of the battery of the target vehicle while monitoring the battery of the target vehicle.

[0123] Optionally, the charging data may be data of the battery of the target vehicle when it is actually working. This application does not impose any specific limitation on this.

[0124] Optionally, the number of time frames included in the charging data may be set according to actual needs. For example, the number of time frames may be no less than 50, or no less than 30. This application does not impose any specific restrictions on this.

[0125] Optionally, the duration of the charging data can be set according to actual needs. The duration can be no less than 10 minutes, or no less than 5 minutes. This application does not make specific restrictions on this.

[0126] In one possible manner, the charging data includes current mutation point, terminal time, charging state, current, voltage list, temperature list, etc.

[0127] S402: The processing device determines, based on the charging data, voltage change data and battery characteristic parameters when the battery current jumps during the charging process.

[0128] In one possible manner, the processing device may determine multiple current mutation points in the charging data. The processing device may determine the change data corresponding to each current mutation point in the multiple current mutation points. The change data includes voltage characteristics corresponding to each of multiple time frames after the voltage mutation point.

[0129] In yet another possible manner, the processing device may perform feature extraction on the charging data to determine battery feature parameters corresponding to multiple time frames in the charging data.

[0130] It should be noted that the processing device determines the voltage change data and battery characteristic parameters when the current jumps during the charging process of the battery based on the charging data. The method in S202 in which the processing device determines the voltage change data when the current jumps during the charging process in the first characteristic data and the battery characteristic parameters during the battery charging process in the second characteristic data from the sample data will not be repeated here.

[0131] S403: The processing device determines an SOC estimation result corresponding to the charging data according to the change data, the battery characteristic parameters and the target SOC estimation model.

[0132] In one possible manner, the processing device may determine a temperature interval corresponding to the starting charging temperature of the charging data. The processing device may determine a target SOC estimation model corresponding to the temperature interval from a plurality of preset SOC estimation models. One preset SOC estimation model may correspond to one temperature interval.

[0133] In one example, the temperature interval may include 5 degrees Celsius to 10 degrees Celsius and 10 degrees Celsius to 15 degrees Celsius. The starting charging temperature of the charging data is 13 degrees Celsius. The processing device may determine that the charging data is in a temperature interval of 10 degrees Celsius to 15 degrees Celsius. The processing device may determine a target SOC estimation model corresponding to the temperature interval of 10 degrees Celsius to 15 degrees Celsius from a plurality of preset SOC estimation models.

[0134] In one possible manner, the SOC estimation model includes a first model and a second model. The processing device may input the change data into the first model to obtain a first result, and input the battery characteristic parameter into the second model to obtain a second result. The processing device may perform weighted integration on the first result and the second result to obtain an SOC estimation result.

[0135] In one possible manner, the processing device can solve the root mean square error based on the SOC estimation result and the current SOC value of the vehicle to obtain an online verification result, and can optimize the SOC estimation model based on the first characteristic data, the second characteristic data of the charging data, and the SOC estimation result.

[0136] In another possible manner, the processing device may train the target SOC estimation model based on the change data, the battery characteristic parameters and the SOC estimation result to obtain a trained target SOC estimation model, that is, update the target SOC estimation model through the change data, the battery characteristic parameters and the SOC estimation result. The manner in which the processing device trains the target SOC estimation model may refer to the above S203 and will not be described in detail here.

[0137] Based on this, the present application can more accurately estimate the SOC of the battery through the SOC estimation model by carefully analyzing the current and voltage changes during the battery charging process, as well as the characteristic parameters of the battery itself, thereby providing the driver with more reliable cruising range information.

[0138] In some embodiments, Figure 5 As shown, it is a schematic diagram of a model training process provided by this application.

[0139] In one possible manner, the processing device may obtain sample charging data. The processing device may obtain the starting charging temperature of the sample data. The processing device may perform data cleaning on the sample charging data. The processing device may determine the full charge cutoff point of the sample charging data. The processing device may reversely infer the ampere-hour integral based on the full charge cutoff point to obtain SOC label data. The processing device may divide the current into stages to determine the current mutation point. The processing device may determine the voltage response segment corresponding to the current mutation point. The processing device may perform data reconstruction and feature extraction on the voltage response segment to obtain voltage change data when the current jumps during the battery charging process. The processing device may extract the entire process of the sample charging data. The processing device may perform feature extraction on the entire process of the sample charging data to obtain battery characteristic parameters during the battery charging process. The processing device may train the voltage change data and SOC label data when the current jumps during the battery charging process to obtain an SOC verification model. The processing device may train the battery characteristic parameters and SOC label data during the battery charging process to obtain an initial SOC estimation model. The processing device can perform weighted integration on the SOC verification model and the initial SOC estimation model to obtain an SOC estimation model integrating voltage response calibration.

[0140] In some embodiments, Figure 6 As shown, it is a schematic diagram of a SOC estimation process provided by the present application.

[0141] In one possible manner, the processing device can obtain charging data. The processing device can obtain the starting charging temperature of the data. The processing device can clean the charging data. The processing device can divide the current into stages and determine the current mutation point. The processing device can determine the voltage response segment corresponding to the current mutation point. The processing device can reconstruct the voltage response segment and extract features to obtain the voltage change data when the current jumps during the battery charging process. The processing device can extract the entire process of the sample charging data. The processing device can extract features of the entire process of the sample charging data to obtain the battery characteristic parameters during the battery charging process. The processing device can match the target SOC estimation model according to the starting charging temperature. The processing device can input the battery characteristic parameters during the battery charging process and the voltage change data when the current jumps during the battery charging process into the target SOC estimation model to obtain the SOC estimation result.

[0142] In some embodiments, Figure 7 As shown, it is a schematic diagram of a model deployment process provided by this application.

[0143] In one possible approach, the model deployment process may include the following three sub-processes: model initialization, model online update, and SOC online estimation.

[0144] Among them, model initialization includes: obtaining charging data of experimental tests, obtaining label SOC through experimental tests, feature extraction and model training, and unified model initialization.

[0145] Specifically, before the vehicle battery leaves the factory, the entire model training process can be completed based on strict standard experimental test data. This process covers everything from the precise recording of the starting temperature of charging, to the comprehensive extraction of data features, to the calculation of the label SOC, and finally the fine training and optimization of the model parameters. This process ensures the high accuracy and reliability of the initial SOC estimation model, laying a solid foundation for subsequent practical applications. In addition, the standard SOC estimation model obtained based on standard experimental data can be uniformly initialized and installed on actual vehicle batteries produced in the same batch. It ensures that all vehicles leaving the factory are equipped with the same benchmark SOC estimation capability, providing a comparable benchmark for subsequent personalized adjustments and calibrations.

[0146] The online model update includes: obtaining real vehicle charging data, inferring label SOC and feature extraction, and model differentiation update to verify the performance of each model.

[0147] Specifically, as the vehicle battery is actually used, the processing device will continuously collect additional charging data. These data reflect the performance changes of the battery under actual working conditions, including but not limited to changes in key parameters such as temperature, current, and voltage. Based on these real vehicle data, the processing device executes the model training process to fine-tune and optimize the initial model to adapt to the actual usage status of the battery. After collecting a sufficient amount of real vehicle charging data, the processing device can incrementally update the standard SOC estimation model. This process includes adjusting the model parameters using the new data and verifying the performance of the updated model. By comparing the model estimation results with the actual battery status, we ensure that the updated model can more accurately reflect the actual status of the current battery and achieve differentiation and personalization of the algorithm model.

[0148] SOC online estimation includes: obtaining real vehicle charging data, feature extraction, model inference, and SOC estimation.

[0149] Specifically, the updated model can be applied to the charging data of the actual vehicle to obtain the SOC estimation results during the charging process. This process not only provides real-time information on the remaining battery power, but also provides a calibration basis for the SOC display on the vehicle side. By comparing the model estimation results with the original SOC display value of the vehicle.

[0150] The above mainly introduces the solution provided by the embodiment of the present application from the perspective of the method. In order to achieve the above functions, the model training device and the SOC estimation device include hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should easily realize that, in combination with the units and algorithm steps of each example described in the embodiments disclosed herein, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware 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 exceed the scope of this application.

[0151] The embodiments of the present application can divide the functional modules of the model training device, SOC estimation device or processing device according to the above method. For example, the model training device and the SOC estimation device can include functional modules corresponding to the functional divisions, or two or more functions can be integrated into one processing module. The above integrated modules can be implemented in the form of hardware or in the form of software functional modules. It should be noted that the division of modules in the embodiments of the present application is schematic and is only a logical functional division. There may be other division methods in actual implementation.

[0152] Figure 8is a block diagram of a model training device according to an exemplary embodiment. Figure 8 The model training device includes: an acquisition unit 501, a determination unit 502 and a training unit 503.

[0153] In a possible manner, the acquisition unit 501 is used to acquire a plurality of sample charging data of a battery of a target vehicle.

[0154] In a possible manner, the determining unit 502 is configured to determine, for each sample charging data among a plurality of sample charging data, first characteristic data and second characteristic data of the sample charging data.

[0155] In a possible manner, the training unit 503 is used to train the first feature data and the second feature data according to a preset algorithm to obtain an SOC estimation model.

[0156] In a possible manner, the determination unit 502 is specifically configured to: determine a plurality of current mutation points in the sample charging data; and perform feature extraction on data corresponding to the plurality of current mutation points to obtain first feature data of the sample charging data.

[0157] In one possible manner, the determination unit 502 is specifically used to: determine, for each current mutation point among multiple current mutation points, voltage change data corresponding to the current mutation point; the voltage change data includes voltage characteristics corresponding to multiple time frames after the voltage mutation point; determine the SOC value corresponding to each of the multiple time frames after the voltage mutation point; and determine the first characteristic data based on the voltage change data corresponding to the current mutation point and the SOC value corresponding to each of the multiple time frames after the voltage mutation point.

[0158] In one possible manner, the determination unit 502 is specifically used to: perform feature extraction on the sample charging data to determine a set of battery feature parameters corresponding to the sample charging data; the battery feature set includes battery feature parameters corresponding to multiple time frames in the sample charging data; determine the SOC value corresponding to each of the multiple time frames; and determine the second feature data based on the battery feature set and the SOC value corresponding to each battery feature in the battery feature set.

[0159] In one possible manner, the training unit 503 is specifically used to: train the first feature data according to a first preset algorithm to obtain a first model, and train the second feature data according to a second preset algorithm to obtain a second model; and perform weighted integration of the first model and the second model to obtain an SOC estimation model.

[0160] In one possible manner, the training unit 503 is specifically used for: also for dividing the first characteristic data and the second characteristic data of the sample charging data into corresponding temperature sets for each sample charging data in multiple sample charging data based on the starting charging temperature corresponding to the sample charging data; and training the first characteristic data and the second characteristic data in the temperature set according to a preset algorithm to obtain an SOC estimation model corresponding to the temperature set.

[0161] Fig. 9 FIG. 1 is a block diagram of a SOC estimation device according to an exemplary embodiment. Fig. 9 The SOC estimation device includes: an acquisition unit 601, a determination unit 602 and a processing unit 603.

[0162] In a possible manner, the acquisition unit 601 is used to acquire charging data of a battery of a target vehicle.

[0163] In a possible manner, the determination unit 602 is configured to determine first characteristic data and second characteristic data of the charging data.

[0164] In a possible manner, the processing unit 603 is used to input the first characteristic data and the second characteristic data into a target SOC estimation model to obtain an SOC estimation result corresponding to the charging data.

[0165] In one possible manner, the determination unit 602 is specifically used to: determine multiple current mutation points in the charging data; for each current mutation point among the multiple current mutation points, determine the voltage change data corresponding to the current mutation point, and determine the voltage change data as the first characteristic data; the voltage change data includes voltage characteristics corresponding to each of the multiple time frames after the voltage mutation point.

[0166] In one possible manner, the determination unit 602 is specifically used to: extract features from the charging data, determine a battery feature parameter set corresponding to the charging data, and determine the battery feature parameter set as the second feature data; the battery feature set includes battery feature parameters corresponding to multiple time frames in the charging data.

[0167] In one possible manner, the processing unit 603 is specifically used to: input the change data into the first model to obtain a first result, and input the battery characteristic parameter into the second model to obtain a second result; and perform weighted integration on the first result and the second result to obtain an SOC estimation result.

[0168] In one possible manner, the processing unit 603 is further used to: determine a temperature range corresponding to a starting charging temperature of the charging data; determine a target SOC estimation model corresponding to the temperature range from a plurality of preset SOC estimation models; and one preset SOC estimation model corresponds to one temperature range.

[0169] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0170] Fig.10 FIG. 1 is a block diagram of a processing device according to an exemplary embodiment. Fig.10 As shown, the processing device includes but is not limited to: a processor 701 and a memory 702 .

[0171] The memory 702 is used to store executable instructions of the processor 701. It can be understood that the processor 701 is configured to execute instructions to implement the model training method and SOC estimation method in the above embodiment.

[0172] It should be noted that those skilled in the art can understand that Fig.10 The processing device structure shown in the figure does not constitute a limitation on the processing device, and the processing device may include Fig.10 More or fewer components may be shown, or certain components may be combined, or the components may be arranged differently.

[0173] The processor 701 is the control center of the processing device. It uses various interfaces and lines to connect various parts of the entire processing device. By running or executing software programs and / or modules stored in the memory 702, and calling data stored in the memory 702, it performs various functions of the processing device and processes data, thereby monitoring the processing device as a whole. The processor 701 may include one or more processing units. Optionally, the processor 701 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, and the modem processor mainly processes wireless communications. It is understandable that the above-mentioned modem processor may not be integrated into the processor 701.

[0174] The memory 702 may be used to store software programs and various data. The memory 702 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required by at least one functional module (such as a determination unit, a processing unit, etc.), etc. In addition, the memory 702 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0175] In an exemplary embodiment, a computer-readable storage medium including instructions is also provided, such as a memory 702 including instructions, and the above instructions can be executed by a processor 701 of a processing device to implement the method in the above embodiment.

[0176] In actual implementation, Figure 8 The acquisition unit 501, the determination unit 502, the training unit 503 and Fig. 9 The functions in the acquisition unit 601, the determination unit 602, and the processing unit 603 can all be implemented by Fig.10 The processor 701 in the embodiment calls the computer program stored in the memory 702. The specific execution process can refer to the description of the method part in the above embodiment, which will not be repeated here.

[0177] Optionally, the computer-readable storage medium may be a non-temporary computer-readable storage medium, for example, the non-temporary computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0178] In an exemplary embodiment, the present application also provides a computer program product including one or more instructions, and the one or more instructions can be executed by the processor 701 of the processing device to complete the method in the above embodiment.

[0179] It should be noted that when the instructions in the above-mentioned computer-readable storage medium or one or more instructions in the computer program product are executed by the processor of the processing device, the various processes of the above-mentioned method embodiment are implemented, and the same technical effect as the above-mentioned method can be achieved. To avoid repetition, they will not be repeated here.

[0180] Through the description of the above implementation methods, technical personnel in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0181] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of modules or units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0182] The units described as separate components may or may not be physically separated, and the components shown as units may be one physical unit or multiple physical units, that is, they may be located in one place or distributed in multiple different places. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0183] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0184] 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 readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application is essentially or the part that contributes to the prior art or the full classification part or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium, including a number of instructions to enable a device (which can be a single-chip microcomputer, chip, etc.) or a processor (processor) to execute the full classification part or part of the steps of each embodiment method of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, ROM, RAM, disk or CD and other media that can store program code.

[0185] The above are only specific implementations of the present application, but the protection scope of the present application is not limited thereto, and any changes or substitutions within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A SOC estimation method, characterized in that: The method comprises: Obtaining charging data of the battery of the target vehicle; Determine, according to the charging data, voltage change data and battery characteristic parameters when the current of the battery jumps during the charging process; An SOC estimation result corresponding to the charging data is determined according to the change data, the battery characteristic parameters and a target SOC estimation model.

2. The method according to claim 1, characterized in that The determining of voltage change data when a current jump occurs during the charging process of the battery includes: Determining a plurality of current mutation points in the charging data; For each current mutation point among the multiple current mutation points, the change data corresponding to the current mutation point is determined; the change data includes voltage characteristics corresponding to multiple time frames after the voltage mutation point.

3. The method according to claim 2, characterized in that Determine battery characteristic parameters, including: Feature extraction is performed on the charging data to determine battery feature parameters corresponding to each of a plurality of time frames in the charging data.

4. The method according to any one of claims 1 to 3, characterized in that The SOC estimation model includes a first model and a second model, and determining the SOC estimation result corresponding to the charging data according to the change data, the battery characteristic parameter and the target SOC estimation model includes: Inputting the change data into the first model to obtain a first result, and inputting the battery characteristic parameter into the second model to obtain a second result; The first result and the second result are weightedly integrated to obtain the SOC estimation result.

5. The method according to claim 4, characterized in that The charging data includes a starting charging temperature during the battery charging process; and determining the SOC estimation result corresponding to the charging data according to the change data, the battery characteristic parameter and the target SOC estimation model includes: Determine the temperature range corresponding to the starting charging temperature of the charging data; Determining a target SOC estimation model corresponding to the temperature interval from a plurality of preset SOC estimation models; one preset SOC estimation model corresponds to one temperature interval; The change data in the charging data and the battery characteristic parameters are input into the target SOC estimation model to obtain the SOC estimation result corresponding to the charging data.

6. The method according to claim 5, characterized in that The method further comprises: Based on the change data, the battery characteristic parameters and the SOC estimation result, the target SOC estimation model is trained to obtain a trained target SOC estimation model.

7. A model training method, characterized in that: include: Acquire multiple sample charging data of a battery of a target vehicle; For each sample charging data of the plurality of sample charging data, first characteristic data and second characteristic data of the sample charging data are determined; the first characteristic data includes voltage change data when a current jump occurs during the battery charging process, and state of charge (SOC) values ​​corresponding to a plurality of voltages when the voltage changes; the second characteristic data includes battery characteristic parameters during the battery charging process, and SOC values ​​corresponding to the voltage characteristic parameters; The first characteristic data and the second characteristic data are trained according to a preset algorithm to obtain an SOC estimation model.

8. The method according to claim 7, characterized in that Determining first characteristic data of the sample charging data includes: Determining a plurality of current mutation points in the sample charging data; Feature extraction is performed on the data corresponding to the multiple current mutation points to obtain first feature data of the sample charging data.

9. The method according to claim 8, characterized in that said extracting features from the data corresponding to the multiple current mutation points to obtain first feature data of the sample charging data; For each current mutation point among the multiple current mutation points, determine voltage change data corresponding to the current mutation point; the voltage change data includes voltage characteristics corresponding to each of multiple time frames after the voltage mutation point; Determine the SOC values ​​corresponding to the multiple time frames after the voltage mutation point; The first characteristic data is determined based on the voltage change data corresponding to the current mutation point and the SOC values ​​corresponding to a plurality of time frames after the voltage mutation point.

10. The method according to claim 7, characterized in that The determining of the second characteristic data of the sample charging data comprises: Extracting features from the sample charging data to determine a battery feature parameter set corresponding to the sample charging data; the battery feature parameter set includes battery feature parameters corresponding to each of a plurality of time frames in the sample charging data; Determine the SOC value of the battery corresponding to each of the multiple time frames; The second feature data is determined based on the battery feature set and the SOC value corresponding to each battery feature in the battery feature set.

11. The method according to any one of claims 7 to 10, characterized in that: The step of training the first characteristic data and the second characteristic data according to a preset algorithm to obtain an SOC estimation model includes: The first feature data is trained according to a first preset algorithm to obtain a first model, and the second feature data is trained according to a second preset algorithm to obtain a second model; The first model and the second model are weighted integrated to obtain the SOC estimation model.

12. The method according to any one of claims 7 to 10, characterized in that: The step of training the first characteristic data and the second characteristic data according to a preset algorithm to obtain an SOC estimation model includes: For each sample charging data of the plurality of sample charging data, based on a starting charging temperature corresponding to the sample charging data, dividing the first characteristic data and the second characteristic data of the sample charging data into a corresponding temperature set; According to the preset algorithm, the first characteristic data and the second characteristic data in the temperature set are trained to obtain an SOC estimation model corresponding to the temperature set.

13. A SOC estimation device, characterized in that: The device comprises: an acquisition unit, a determination unit and a processing unit; The acquisition unit is used to acquire charging data of a battery of a target vehicle; The determining unit is used to determine, according to the charging data, voltage change data and battery characteristic parameters when the current of the battery jumps during the charging process; The processing unit is used to determine the SOC estimation result corresponding to the charging data according to the change data, the battery characteristic parameters and the target SOC estimation model.

14. A model training device, characterized in that: The device comprises: an acquisition unit, a determination unit and a training unit; The acquisition unit is used to acquire a plurality of sample charging data of the battery of the target vehicle; The determining unit is used to determine, for each sample charging data of the plurality of sample charging data, first characteristic data and second characteristic data of the sample charging data; the first characteristic data includes voltage change data when a current jump occurs during the battery charging process, and SOC values ​​corresponding to a plurality of voltages when the voltage changes; the second characteristic data includes battery characteristic parameters during the battery charging process, and SOC values ​​corresponding to the voltage characteristic parameters; The training unit is used to train the first feature data and the second feature data according to a preset algorithm to obtain an SOC estimation model.

15. A processing device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is configured to execute the instructions to implement the method according to any one of claims 1 to 6, or to implement the method according to any one of claims 7 to 12.

16. A computer-readable storage medium, characterized in that: When the computer-executable instructions stored in the computer-readable storage medium are executed by a processor of a processing device, the processing device can execute the method according to any one of claims 1 to 6, or execute the method according to any one of claims 7 to 12.

17. A vehicle, characterized in that: The vehicle includes the SOC estimating device according to claim 13 and the model training device according to claim 14.

18. A computer program product comprising instructions, characterized in that When the instructions are executed by a computer, the computer is enabled to execute the method according to any one of claims 1 to 6, or execute the method according to any one of claims 7 to 12.

Images