A battery soc estimation method and system, electronic device and storage medium

Abstract

The application provides a battery SOC estimation method and system, an electronic device and a storage medium. When the cloud end detects that the current time reaches a parameter update period, the model parameters used for estimating the SOC are updated based on the historical operation data of the battery received from the vehicle end, and the updated model parameters are sent to the vehicle end. The vehicle end determines the battery SOC according to the updated model parameters received from the cloud end. The application improves the battery SOC estimation accuracy.

Description

Technical Field

[0001] This application relates to the field of battery technology, and in particular to a method, system, electronic device, and storage medium for estimating battery SOC. Background Technology

[0002] As the primary energy source for electric vehicles, the State of Charge (SOC) of the battery is one of the most important and fundamental parameters in the energy management system. Currently, the estimation of the battery SOC of electric vehicles mainly uses the ampere-hour integral method as its basis. This method calculates the change in SOC by accumulating the charging and discharging currents in real time, and uses an offline calibrated open-circuit voltage (OCV) lookup table and a family of Kalman filter algorithms to correct the ampere-hour integral results online, in order to suppress integral drift and initial errors.

[0003] However, the model parameters are mostly fixed values ​​and cannot be adaptively adjusted as the battery ages during use, causing the SOC estimation accuracy to continuously decrease with the number of charge-discharge cycles. In addition, the limited computing power on the vehicle side restricts the application of higher-precision estimation methods in real vehicles. Summary of the Invention

[0004] In view of this, embodiments of this application provide a method, system, electronic device, and storage medium for estimating battery SOC, thereby improving the accuracy of battery SOC estimation.

[0005] This application mainly includes the following aspects: In a first aspect, embodiments of this application provide a method for estimating battery SOC, the estimation method being applied to a battery SOC estimation system; the estimation system includes: a vehicle-side endpoint and a cloud-based endpoint; the estimation method includes: When the cloud detects that the current time has reached the parameter update cycle, it updates the model parameters used to estimate SOC based on the historical operating data of the battery received from the vehicle, and sends the updated model parameters to the vehicle. The vehicle determines the battery SOC based on the updated model parameters received from the cloud.

[0006] Furthermore, the model parameters include: polarization resistance and open-circuit voltage in the battery equivalent circuit model; updating the model parameters used to estimate SOC based on historical battery operating data received from the vehicle includes: Based on the equivalent circuit model of the battery, the differential model of the battery is determined. The identification coefficients in the difference model are iteratively updated using the recursive least squares method with an introduced forgetting factor. Based on the iteratively updated identification coefficients, new polarization resistance and open-circuit voltage are obtained; Replace the original polarization resistor and open-circuit voltage with the new polarization resistor and open-circuit voltage, respectively.

[0007] Furthermore, the model parameters also include: the hyperparameters of the Kalman filter; the updating of the model parameters for estimating SOC based on historical battery operating data received from the vehicle includes: Based on the historical operation data, a fitness function is constructed; The hyperparameters are iteratively searched using the tuna optimization algorithm and fitness function to determine the hyperparameters corresponding to the minimum fitness. Replace the original hyperparameters with the hyperparameters corresponding to the minimum fitness.

[0008] Furthermore, the model parameters also include: the battery's State of Health (SOH); the updating of the model parameters for estimating the State of Charge (SOC) based on historical battery operating data received from the vehicle includes: Based on the historical operation data, the cloud determines the feature vector and the true SOH value for each charge-discharge cycle; The feature vectors and true SOH values ​​corresponding to each charge-discharge cycle are used to construct a sample dataset. The sample dataset is randomly sampled to obtain multiple sub-sample sets; Construct corresponding regression models based on multiple subsample sets; Each subsample set is input into the corresponding regression model for training, and the initial SOH prediction value corresponding to each subsample set is obtained; The average of the initial SOH predictions for each subset is used to determine the new SOH of the battery. Replace the original SOH of the battery with the new SOH.

[0009] Furthermore, the method for estimating the battery SOC also includes: Based on historical operating data of batteries from multiple production batches of the target model, the cloud uses a clustering algorithm to determine the factory calibration parameters of the target model's batteries.

[0010] Secondly, embodiments of this application also provide a battery SOC estimation system, the estimation system comprising: vehicle-side and cloud-side; When the cloud detects that the current time has reached the parameter update cycle, it updates the model parameters used to estimate SOC based on the historical operating data of the battery received from the vehicle, and sends the updated model parameters to the vehicle. The vehicle determines the battery SOC based on the updated model parameters received from the cloud.

[0011] Furthermore, the model parameters include: polarization resistance and open-circuit voltage in the battery equivalent circuit model; the cloud is used for: Based on the equivalent circuit model of the battery, the differential model of the battery is determined. The identification coefficients in the difference model are iteratively updated using the recursive least squares method with an introduced forgetting factor. Based on the iteratively updated identification coefficients, new polarization resistance and open-circuit voltage are obtained; Replace the original polarization resistor and open-circuit voltage with the new polarization resistor and open-circuit voltage, respectively.

[0012] Furthermore, the model parameters also include: the hyperparameters of the Kalman filter; the cloud is used for: Based on the historical operation data, a fitness function is constructed; The hyperparameters are iteratively searched using the tuna optimization algorithm and fitness function to determine the hyperparameters corresponding to the minimum fitness. Replace the original hyperparameters with the hyperparameters corresponding to the minimum fitness.

[0013] Thirdly, embodiments of this application also provide an electronic device, including: a processor, a memory, and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the memory through the bus, and the machine-readable instructions are executed by the processor to perform the steps of the battery SOC estimation method described in the first aspect or any possible implementation of the first aspect.

[0014] Fourthly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the battery SOC estimation method described in the first aspect or any possible implementation of the first aspect.

[0015] This application provides a method, system, electronic device, and storage medium for estimating battery SOC, comprising: when the cloud detects that the current time has reached the parameter update cycle, updating the model parameters used to estimate SOC based on the historical operating data of the battery received from the vehicle, and sending the updated model parameters to the vehicle; the vehicle determines the battery SOC based on the updated model parameters received from the cloud.

[0016] This improves the accuracy of battery SOC estimation.

[0017] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating a battery SOC estimation method provided in an embodiment of this application is shown; Figure 2 An equivalent circuit model of the battery provided in the embodiments of this application is shown; Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0021] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0022] The methods, apparatus, electronic devices, or computer-readable storage media described in this application can be applied to any scenario that requires battery SOC estimation. This application does not limit specific application scenarios, and any scheme using the battery SOC estimation method and system provided in this application is within the protection scope of this application.

[0023] It is worth noting that, as the primary energy source for electric vehicles, the State of Charge (SOC) of the battery is one of the most important and fundamental parameters in the energy management system. Current SOC estimation for electric vehicles primarily uses the ampere-hour integral method as its basis. This method calculates the SOC change by accumulating the charging and discharging currents in real time, and uses offline calibrated open-circuit voltage (OCV) lookup tables and Kalman filter family algorithms to correct the ampere-hour integral results online, suppressing integral drift and initial errors. Simultaneously, battery state of health (SOH) estimation typically relies on offline lookup models or correction strategies based on specific characteristic points (such as voltage plateaus in constant current charging curves, peak values ​​in incremental capacity curves, etc.), updating the SOH value through periodic or conditionally triggered calibrations. However, existing methods have the following shortcomings: model parameters (such as the OCV-SOC relationship curve) are mostly fixed values ​​calibrated offline at the factory, unable to adaptively adjust with battery aging during use, leading to a continuous decline in the accuracy of SOC and SOH estimations with increasing charge-discharge cycles. Furthermore, performance differences between individual cells and different production batches cannot be eliminated through unified offline parameters, further reducing the robustness of the estimation. Meanwhile, the limited computing power on the vehicle side makes it difficult to deploy data-driven algorithms at the battery cell level, which restricts the application of higher-precision estimation methods in actual vehicles.

[0024] To address the aforementioned issues, this application proposes a method, system, electronic device, and storage medium for estimating battery SOC, thereby improving the accuracy of battery SOC estimation.

[0025] To facilitate understanding of this application, the technical solutions provided in this application will be described in detail below with reference to specific embodiments.

[0026] Please see Figure 1 , Figure 1 A flowchart illustrating a method for estimating battery SOC provided in an embodiment of this application.

[0027] like Figure 1 As shown in the illustration, the battery SOC estimation method provided in this application embodiment is applied to a battery SOC estimation system. The estimation system includes: a vehicle-side component and a cloud-based component.

[0028] Estimating battery SOC involves the following steps: Step S101: When the cloud detects that the current time has reached the parameter update cycle, it updates the model parameters used to estimate SOC based on the historical operating data of the battery received from the vehicle, and sends the updated model parameters to the vehicle.

[0029] Here, model parameters may include, but are not limited to: polarization resistance and open-circuit voltage in the battery equivalent circuit model, hyperparameters of the Kalman filter, and battery health status. As an example, the update cycle is 30 seconds. The cloud handles computationally intensive tasks, periodically using historical operational data uploaded from the vehicle to optimize and update model parameters online or offline, ensuring the model's accuracy throughout its entire lifecycle.

[0030] In this embodiment of the application, for updating the model parameters, including the polarization resistance and open-circuit voltage in the battery equivalent circuit model, step S101 includes the following steps: Step S11: Determine the differential model of the battery based on the equivalent circuit model of the battery.

[0031] Here, both the cloud and the vehicle construct the same first-order equivalent circuit model. The equivalent circuit model consists of an ohmic internal resistance and a single RC network connected in series. The RC network is used to simulate the electrochemical polarization process inside the battery. The state-space equation of the equivalent circuit model is shown in Equation (1).

[0032] (1), in, The first time derivative of the polarization voltage. The current flowing through the battery, Polarization voltage, Polarized capacitor, Polarization resistor, This is the open-circuit voltage of the battery under no-load conditions. The internal resistance is ohmic. Wherein, It is positive when charging and negative when discharging.

[0033] It should be noted that the cloud platform uses the current and terminal voltage at adjacent time points uploaded by the vehicle to perform online identification of the equivalent circuit model parameters. To do this, the state-space equations first need to be converted into transfer functions through Laplace transform. The transfer function of the equivalent circuit model is shown in Equation (2).

[0034] (2), in, This is the transfer function. Then, the above continuous domain is converted using the backward difference method. The transfer function is converted to the discrete domain Z-transform. The discretized transfer function G(z) is shown in Equation (3).

[0035] (3), Where ΔT is the data sampling period, representing the time interval between two consecutive current or voltage acquisitions. Let be the difference between the terminal voltage and the open-circuit voltage. Performing an inverse Z-transform on equation (3) and rearranging, we obtain the result with the terminal voltage as the difference. The output difference model is shown in equation (4).

[0036] (4), in, Let be the terminal voltage at time k. Let K be the terminal voltage at time k-1. Let K be the current at time k. Let K be the current at time k-1. , , and The identification coefficients are related to the battery model parameters. The specific representation of the identification coefficients is shown in formula (5).

[0037] (5), pass , , It can be obtained The expression for is shown in formula (6).

[0038] (6).

[0039] Step S12: The identification coefficients in the difference model are iteratively updated using the recursive least squares method with an introduced forgetting factor.

[0040] In this embodiment, the cloud uses a recursive least squares method with a forgetting factor to process the battery data. and Identification. The recursive least squares method introduces a forgetting factor (a constant less than 1) to give higher weight to recent data, gradually weakening the influence of old data, thereby effectively avoiding data saturation and achieving faster convergence. Under complex working conditions such as dynamic stress testing, the recursive least squares method with a forgetting factor can accurately and quickly respond to parameter changes. The model of the recursive least squares method with a forgetting factor is shown in formula (7).

[0041] (7), in, Let be the terminal voltage at time k. For data variables, For coefficient variables, For model error, It follows a Gaussian distribution with zero mean. Wherein, the defined... and The expression is shown in formula (8).

[0042] (8).

[0043] In this embodiment of the application, the following steps are specifically used to obtain the result. and First, initialize the coefficient variables and the error covariance matrix, i.e., give initial values ​​θ0 for the coefficient variables and P0 for the error covariance matrix. Then, calculate the gain vector and the error covariance matrix. The expressions for the gain vector and the error covariance matrix are shown in formula (9). (9), in, Genetic factors, For the gain vector, Let be the error covariance matrix. Let be the identity matrix. Then, calculate the model error and coefficient variables, expressed as in formula (10). (10).

[0044] Finally, let k = k + 1, and return to step two for the next iteration.

[0045] Step S13: Based on the iteratively updated identification coefficients, new polarization resistance and open-circuit voltage are obtained.

[0046] After the iteration process is completed, the final estimated values ​​of the coefficient variables are obtained, i.e., the identification coefficients. , , and Combining formula (6), a new result can be obtained in the cloud. and .

[0047] Step S14: Replace the original polarization resistor and polarization voltage with the new polarization resistor and open-circuit voltage, respectively.

[0048] In this embodiment of the application, for updating the hyperparameters of the model parameters, including the Kalman filter, step S101 includes the following steps: Step S21: Construct a fitness function based on the historical running data.

[0049] Using a cloud-based tuna optimization algorithm, a set of hyperparameters is automatically found to maximize the estimation accuracy of the extended Kalman filter. These hyperparameters include: , , , and , The five parameters are combined into an optimization vector. .in, The first diagonal element of the error covariance matrix. The second diagonal element of the error covariance matrix. The first diagonal element of the process noise covariance matrix. is the second diagonal element of the process noise covariance matrix, and r is the observation noise covariance matrix.

[0050] The fitness function aims to minimize the SOC estimation error and the terminal voltage fitting error, and its expression is shown in formula (11).

[0051] (11), in, This represents the total number of samples. For the first Measured value of terminal voltage at time; For the first Calculated value of terminal voltage at time step; For the first Real SOC value at any given moment; For the first SOC estimate at time step; These are the weighting coefficients.

[0052] Step S22: Use the tuna optimization algorithm and fitness function to iteratively search the hyperparameters to determine the hyperparameters corresponding to the minimum fitness.

[0053] The Tuna Optimization Algorithm is an intelligent optimization algorithm that simulates the predatory behavior of tuna schools. It generates a set of candidate solutions, simulates the spiral and parabolic foraging behaviors of tuna schools in the parameter space, and iteratively searches for the global optimum. The overall execution flow of the algorithm is as follows: First, set the population size, i.e. the number of candidate solutions participating in the search at the same time, set the maximum number of iterations, i.e. the upper limit of the number of rounds of repeated search of the tuna optimization algorithm, and set the upper and lower limits of the search for each parameter in the optimization vector to ensure the stability of the extended Kalman filter. The constraint range is shown in formula (12).

[0054] (12), Within the constraints, a population is randomly generated; each individual in the population represents a set of hyperparameters of the Kalman filter to be optimized. Then, for each individual, the five parameter values ​​corresponding to that individual are used as the hyperparameter configuration of the extended Kalman filter, the extended Kalman filter is run, and the state of charge and terminal voltage at all times are obtained by using historical running data. Specifically, a. Using the hyperparameters provided by the individual, as shown in formula (13), the initial state, initial covariance matrix, process noise covariance matrix and observation noise variance of the Kalman filter are set. (13) in, a. The initial state value; b. For each time step, based on the state estimate from the previous time step, predict the state at the current time step, i.e. ,in, This is the predicted state value at time k. c. The state prediction value at time k-1; D. Update the uncertainty of the state prediction, i.e., the error covariance matrix, where A is the state transition matrix and B is the input matrix. ,in, d. Calculate the error covariance matrix at time k; ; Calculate the Kalman gain, i.e. e. Using the actual measured terminal voltage at the current moment Compared with the terminal voltage predicted by the model The difference between the two values ​​is used to correct the predicted state, thus obtaining the optimal state estimate, i.e. f. After completing the state correction, update the error covariance matrix, i.e. g. Let k = k + 1, return to step b, and continue processing the next time step until all data has been processed. After the recursion is completed, output the SOC and terminal voltage at all time steps for use in fitness function calculation. Then, substitute the SOC and terminal voltage at all time steps of the individual into the fitness function to calculate the fitness value of the individual. Next, find the individual in the current population that minimizes fitness and record its position as the current optimal solution. After that, each individual updates its position according to the random rule, that is, adjusts the parameter value. The expression for updating the position is shown in formula (14). (14) in, It is the t+1th iteration One element, It is the tth iteration One element, Represents a random reference individual. Represents the current best individual. and These are the weighting coefficients that influence the current individual's tendency to move towards the optimal state and the previous individual. is a constant used to determine the degree to which the best individual and the previous individual are followed in the initial stage, rand is a random number in the interval [0, 1], and t is the current iteration number. The maximum number of iterations is given by τ, where b is a random number uniformly distributed between 0 and 1, and τ is a random number of 1 or -1. For the spiral search amplitude coefficient, lThis is the search range adjustment coefficient. Finally, after the iteration ends, the hyperparameter corresponding to the minimum fitness value during the entire optimization process is taken as the optimal hyperparameter.

[0055] Step S23: Replace the original hyperparameters with the hyperparameters corresponding to the minimum fitness.

[0056] In this embodiment of the application, for updating the model parameters, including the State of Harmony (SOH) of the battery, a random forest model is used. Step S101 includes the following steps: Step S31: Based on the historical operation data, the cloud determines the feature vector and the true SOH value corresponding to each charge-discharge cycle.

[0057] Here, the cloud identifies and segments each complete charge-discharge cycle based on historical operating data such as current, voltage, and temperature uploaded from the vehicle. For each charge-discharge cycle, multi-dimensional deep features reflecting the State of Harm (SOH) are extracted from the historical operating data, including but not limited to capacity increment curves, differential temperature curves, and constant current / constant voltage charging time ratios. These features are constructed into a feature vector. At the same time, based on the ratio of the actual maximum usable capacity to the factory rated capacity for that cycle, the true SOH value at the corresponding moment of that cycle is calculated.

[0058] Step S32: Construct a sample dataset by combining the feature vectors and the true SOH values ​​corresponding to each charge-discharge cycle.

[0059] Step S33: Randomly sample the sample dataset to obtain multiple sub-sample sets.

[0060] Here, the Bootstrap sampling method is used to randomly sample the sample dataset.

[0061] Step S34: Construct the corresponding regression model based on multiple subsample sets.

[0062] Step S35: Input each subsample set into the corresponding regression model for training to obtain the initial SOH prediction value for each subsample set.

[0063] Step S36: The average value of the initial SOH predictions corresponding to each subsample set is determined as the new SOH of the battery.

[0064] Step S37: Replace the original SOH of the battery with the new SOH.

[0065] It should be noted that before the battery is installed and put into operation, the cloud platform uses a clustering algorithm to determine the factory calibration parameters of the battery model based on historical operating data from multiple production batches of the target model. As an example, the historical operating data covers a temperature range of -20℃ to 30℃ and a SOC range of 0% to 100%, and is taken from the operating records of at least three different production batches of batteries within one year. In this embodiment, the K-means clustering algorithm is used, with a cluster size of 3, corresponding to the three batches. The cloud platform extracts the battery equivalent circuit model parameters and the hyperparameters of the extended Kalman filter corresponding to each charge-discharge cycle from the data of each batch, and uses these parameters to construct feature vectors. After clustering, the model parameters of each cluster center are calculated, and the group of parameters with the highest fitness is selected as the factory calibration parameters for the battery of this vehicle model, using the minimum SOC estimation error as the fitness evaluation criterion. These factory calibration parameters are pre-set in the vehicle's battery management system at the time of vehicle delivery, serving as the initial model parameters for SOC estimation.

[0066] The application employs a CAN bus to communicate with the T-BOX, receiving various updated model parameters (including open-circuit voltage, polarization resistance, SOH value, hyperparameters, and factory calibration parameters) from the cloud. Its update cycle matches the cloud's delivery cycle. A network outage redundancy logic is designed: when vehicle-cloud communication is interrupted (e.g., signal loss), the vehicle automatically switches to offline mode and independently performs SOC estimation based on the previously received cloud parameters. When communication is restored, it automatically synchronizes the latest cloud parameters, completes SOC correction, and ensures vehicle control safety. The vehicle uploads data such as current, voltage, temperature, and current SOC estimate every 10 seconds, while the cloud delivers optimized parameters and models every 30 minutes, forming a closed-loop iterative control cycle of "data upload - cloud optimization - parameter delivery - vehicle correction." Through this mechanism, the system can continuously optimize the model using the powerful computing capabilities of the cloud and ensure high-precision real-time SOC estimation even in complex environments.

[0067] In step S102, the vehicle determines the battery SOC based on the updated model parameters received from the cloud.

[0068] Here, after receiving the updated model parameters from the cloud, the vehicle collects the current voltage, current, and temperature in real time. First, it uses the updated SOH to correct the battery's current actual maximum usable capacity. Then, based on the parameters of the updated equivalent circuit model, it obtains the SOC prior state prediction value by looking up the table using ampere-hour integration and open-circuit voltage. Using the Kalman filter hyperparameters, it calculates the optimal Kalman gain to recursively correct the SOC prior state prediction value, and finally outputs the SOC estimate value.

[0069] This application solves the problem of offline model parameter solidification by establishing an equivalent circuit model and using the forgetting factor recursive least squares method to dynamically identify battery parameters online. It combines the tuna optimization algorithm to adaptively tune the hyperparameters of the extended Kalman filter and uses the random forest model to fuse multi-dimensional features to achieve high-precision health state estimation, providing an accurate calibration benchmark for state of charge estimation. At the same time, a clustering algorithm is introduced to perform global calibration on fleet big data to eliminate individual and batch differences in batteries and output highly robust factory parameters. In addition, a parameter receiving module added to the vehicle enables regular synchronization of cloud parameters, and a network disconnection redundancy fault tolerance mechanism ensures control safety. Finally, a computing power decoupling closed-loop iterative architecture of "cloud complex computing + vehicle real-time control" is constructed, migrating all heavy computing tasks such as parameter identification and AI modeling to the cloud, achieving a comprehensive improvement in the accuracy, robustness and full life cycle adaptability of state of charge estimation without increasing the cost of vehicle hardware.

[0070] This application provides a method for estimating battery SOC, which improves the accuracy of battery SOC estimation.

[0071] Based on the same application concept, this application also provides a battery SOC estimation device corresponding to the battery SOC estimation method provided in the above embodiments. Since the principle of the device in this application is similar to the battery SOC estimation method in the above embodiments of this application, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be described again.

[0072] The battery SOC estimation system provided in this application includes: The estimation system includes: vehicle-side and cloud-side components; When the cloud detects that the current time has reached the parameter update cycle, it updates the model parameters used to estimate SOC based on the historical operating data of the battery received from the vehicle, and sends the updated model parameters to the vehicle. The vehicle determines the battery SOC based on the updated model parameters received from the cloud.

[0073] Furthermore, the model parameters include: polarization resistance and open-circuit voltage in the battery equivalent circuit model; the cloud platform is specifically used for: Based on the equivalent circuit model of the battery, the differential model of the battery is determined. The identification coefficients in the difference model are iteratively updated using the recursive least squares method with an introduced forgetting factor. Based on the iteratively updated identification coefficients, new polarization resistance and open-circuit voltage are obtained; Replace the original polarization resistor and open-circuit voltage with the new polarization resistor and open-circuit voltage, respectively.

[0074] Furthermore, the model parameters also include: the hyperparameters of the Kalman filter; the cloud specifically serves to: Based on the historical operation data, a fitness function is constructed; The hyperparameters are iteratively searched using the tuna optimization algorithm and fitness function to determine the hyperparameters corresponding to the minimum fitness. Replace the original hyperparameters with the hyperparameters corresponding to the minimum fitness.

[0075] Furthermore, the model parameters also include: the battery's state of harmonics (SOH); the cloud platform is specifically used for: Based on the historical operation data, the cloud determines the feature vector and the true SOH value for each charge-discharge cycle; The feature vectors and true SOH values ​​corresponding to each charge-discharge cycle are used to construct a sample dataset. The sample dataset is randomly sampled to obtain multiple sub-sample sets; Construct corresponding regression models based on multiple subsample sets; Each subsample set is input into the corresponding regression model for training, and the initial SOH prediction value corresponding to each subsample set is obtained; The average of the initial SOH predictions for each subset is used to determine the new SOH of the battery. Replace the original SOH of the battery with the new SOH.

[0076] Furthermore, the cloud is specifically used for: Based on historical operating data of batteries from multiple production batches of the target model, the cloud uses a clustering algorithm to determine the factory calibration parameters of the target model's batteries.

[0077] This application provides a battery SOC estimation system, which improves the accuracy of battery SOC estimation.

[0078] Please see Figure 3 , Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0079] like Figure 3 As shown, the electronic device 300 includes a processor 310, a memory 320, and a bus 330.

[0080] The memory 320 stores machine-readable instructions executable by the processor 310. When the electronic device 300 is running, the processor 310 and the memory 320 communicate via the bus 330. When the machine-readable instructions are executed by the processor 310, they can perform the operations described above. Figure 1The steps of the battery SOC estimation method in the method embodiment shown are described in detail in the method embodiment, and will not be repeated here.

[0081] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can perform the above-described actions. Figure 1 The steps of the battery SOC estimation method in the method embodiment shown are described in detail in the method embodiment, and will not be repeated here.

[0082] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and devices described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division; in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection may be through some communication interfaces; the indirect coupling or communication connection of devices or units may be electrical, mechanical, or other forms.

[0083] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0084] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0085] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0086] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for estimating the state of charge (SOC) of a battery, characterized in that, The estimation method is applied to a battery SOC estimation system; The estimation system includes: a vehicle-side component and a cloud-based component; the estimation method includes: When the cloud detects that the current time has reached the parameter update cycle, it updates the model parameters used to estimate SOC based on the historical operating data of the battery received from the vehicle, and sends the updated model parameters to the vehicle. The vehicle determines the battery SOC based on the updated model parameters received from the cloud.

2. The method for estimating battery SOC according to claim 1, characterized in that, The model parameters include: polarization resistance and open-circuit voltage in the battery equivalent circuit model; the updating of model parameters for estimating SOC based on historical battery operating data received from the vehicle includes: Based on the equivalent circuit model of the battery, the differential model of the battery is determined. The identification coefficients in the difference model are iteratively updated using the recursive least squares method with an introduced forgetting factor. Based on the iteratively updated identification coefficients, new polarization resistance and open-circuit voltage are obtained; Replace the original polarization resistor and open-circuit voltage with the new polarization resistor and open-circuit voltage, respectively.

3. The method for estimating battery SOC according to claim 2, characterized in that, The model parameters also include: the hyperparameters of the Kalman filter; the updating of the model parameters for estimating SOC based on historical battery operating data received from the vehicle includes: Based on the historical operation data, a fitness function is constructed; The hyperparameters are iteratively searched using the tuna optimization algorithm and fitness function to determine the hyperparameters corresponding to the minimum fitness. Replace the original hyperparameters with the hyperparameters corresponding to the minimum fitness.

4. The method for estimating battery SOC according to claim 3, characterized in that, The model parameters also include: the battery's State of Health (SOH); the updating of the model parameters for estimating SOC based on historical battery operating data received from the vehicle includes: Based on the historical operation data, the cloud determines the feature vector and the true SOH value for each charge-discharge cycle; The feature vectors and true SOH values ​​corresponding to each charge-discharge cycle are used to construct a sample dataset. The sample dataset is randomly sampled to obtain multiple sub-sample sets; Construct corresponding regression models based on multiple subsample sets; Each subsample set is input into the corresponding regression model for training, and the initial SOH prediction value corresponding to each subsample set is obtained; The average of the initial SOH predictions for each subset is used to determine the new SOH of the battery. Replace the original SOH of the battery with the new SOH.

5. The method for estimating battery SOC according to claim 1, characterized in that, The method for estimating the battery's state of charge (SOC) also includes: Based on historical operating data of batteries from multiple production batches of the target model, the cloud uses a clustering algorithm to determine the factory calibration parameters of the target model's batteries.

6. A battery SOC estimation system, characterized in that, The estimation system includes: vehicle-side and cloud-side components; When the cloud detects that the current time has reached the parameter update cycle, it updates the model parameters used to estimate SOC based on the historical operating data of the battery received from the vehicle, and sends the updated model parameters to the vehicle. The vehicle determines the battery SOC based on the updated model parameters received from the cloud.

7. The battery SOC estimation system according to claim 6, characterized in that, The model parameters include: polarization resistance and open-circuit voltage in the battery equivalent circuit model; the cloud platform is used for: Based on the equivalent circuit model of the battery, the differential model of the battery is determined. The identification coefficients in the difference model are iteratively updated using the recursive least squares method with an introduced forgetting factor. Based on the iteratively updated identification coefficients, new polarization resistance and open-circuit voltage are obtained; Replace the original polarization resistor and open-circuit voltage with the new polarization resistor and open-circuit voltage, respectively.

8. The battery SOC estimation system according to claim 7, characterized in that, The model parameters also include: the hyperparameters of the Kalman filter; the cloud is used for: Based on the historical operation data, a fitness function is constructed; The hyperparameters are iteratively searched using the tuna optimization algorithm and fitness function to determine the hyperparameters corresponding to the minimum fitness. Replace the original hyperparameters with the hyperparameters corresponding to the minimum fitness.

9. An electronic device, characterized in that, include: The device includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus. The machine-readable instructions are executed by the processor to perform the steps of the battery SOC estimation method as described in any one of claims 1 to 5.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the battery SOC estimation method as described in any one of claims 1 to 5.

Images