Parameter updating method of power battery model, server, vehicle and storage medium

Abstract

The application relates to the technical field of power batteries, in particular to a power battery model parameter updating method, a server, a vehicle and a storage medium, wherein the method comprises the following steps: acquiring running data of a power battery uploaded by a vehicle and an estimated error table of model parameters; determining model parameters of the power battery according to the running data and a model parameter table of the power battery, constructing a mapping function between running parameters in the running data and the model parameters, and acquiring a confidence interval of the mapping function; determining an evaluation interval to be updated in the model parameter table according to the confidence interval, updating the model parameters of the evaluation interval to be updated according to the mapping function and the estimated error table, and delivering the updated model parameter table to the vehicle, so that the vehicle updates the model parameters of the power battery according to the updated model parameter table. Therefore, the problems of low precision of pure offline parameter identification of a battery equivalent model and excessive resource consumption of online identification at a vehicle end in the prior art are solved.

Description

Technical Field

[0001] This invention relates to the field of power battery technology, specifically to a parameter update method for a power battery model, a server, a vehicle, and a storage medium. Background Technology

[0002] Battery management systems for electric vehicles (EVs) include various state estimation algorithms and control strategies. The battery equivalent model is the foundation of most of these algorithms. The accuracy of the battery equivalent model is closely related to its model parameters. If the model parameters deviate significantly from the actual operating conditions of the electric vehicle, the equivalent model itself will be unable to effectively represent the actual electrical characteristics of the power battery, reducing the estimation accuracy of the relevant state estimation algorithms and thus negatively impacting the use of the electric vehicle.

[0003] Currently, battery equivalent model technologies can be broadly categorized into two types: the first uses offline parameter calibration, and the second uses online identification algorithms. One related technology discloses a rapid estimation method for the internal resistance of lithium-ion batteries. This method analyzes the battery charge-discharge pulses and establishes a first-order RC equivalent circuit model. Based on this model, the charge-discharge pulse curve is divided into two segments influenced by ohmic resistance and polarization resistance. The ohmic internal resistance and polarization internal resistance are obtained based on the voltage and current changes in these two segments, respectively. The time constant and polarization capacitance are obtained by fitting the charge-discharge pulse curve with a first-order step response curve. However, this method can only identify the parameters of the corresponding pulses. Actual operating conditions are time-varying, and the charge-discharge rate is unpredictable, making the selection of the experimental rate extremely difficult. Another related technology discloses an online parameter identification method, system, and storage medium for a power battery model. This method proposes first establishing an equivalent battery model on the vehicle side. At each moment, it simultaneously acquires the terminal voltage output by the equivalent model and the battery terminal voltage collected by sensors. Based on the error between the two, an adaptive online identification algorithm is used to update the parameters of the equivalent model in real time. While this method can improve the problem of excessive model estimation error caused by offline calibration, it also introduces a new algorithm, increasing the computational cost on the vehicle side. Furthermore, compared to offline calibration methods, the algorithm has a certain degree of divergence risk. Summary of the Invention

[0004] One objective of this invention is to provide a parameter update method for a power model, in order to solve the problems of low accuracy in pure offline parameter identification of battery equivalent models and excessive resource consumption in online identification on the vehicle side in the prior art; a second objective is to provide another parameter update method for a power model; a third objective is to provide a server; a fourth objective is to provide a vehicle; and a fifth objective is to provide a computer-readable storage medium.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A method for updating parameters of a power battery model, the method being applied to a server, comprising the following steps: acquiring operational data of the power battery uploaded by a vehicle and an estimation error table of model parameters; determining the model parameters of the power battery based on the operational data and the model parameter table of the power battery, constructing a mapping function between the operational parameters in the operational data and the model parameters, and obtaining a confidence interval of the mapping function; determining an evaluation interval to be updated in the model parameter table based on the confidence interval, updating the model parameters of the evaluation interval to be updated based on the mapping function and the estimation error table, and sending the updated model parameter table to the vehicle, so that the vehicle updates the model parameters of the power battery based on the updated model parameter table.

[0007] Based on the above technical means, the embodiments of this application can determine the corresponding model parameters for the data uploaded by the vehicle, construct the mapping relationship between the operating parameters and the model parameters, determine the evaluation interval to be updated using the confidence interval obtained by the mapping function, and after updating the model parameters, send the model parameter table to the vehicle, enabling the vehicle to update the power battery model parameters. This realizes the interaction of model parameters between the vehicle and the server, fully utilizes the advantages of server-side computing, and realizes the updating of model parameters with poor matching. To a certain extent, this solves the problem of low accuracy of offline parameter calibration and realizes the optimization of the power battery model parameters throughout the entire life cycle.

[0008] Furthermore, updating the model parameters of the evaluation interval to be updated according to the mapping function and the estimation error table includes: identifying the parameters to be updated in the evaluation interval to be updated and the corresponding operating parameters; calculating the model parameters of the operating parameters corresponding to the parameters to be updated using the mapping function, and identifying the estimation errors of the parameters to be updated in the estimation error table; and updating the parameters to be updated according to the model parameters calculated by the mapping function and the estimation errors.

[0009] Based on the above technical means, the embodiments of this application can use mapping functions to calculate the model parameters of the running parameters corresponding to the parameters to be updated, identify the estimation errors of the parameters to be updated in the estimation error table, and further update the parameters to be updated in the evaluation interval using model parameters and estimation errors.

[0010] Furthermore, before updating the model parameters within the evaluation interval to be updated according to the mapping function and the estimation error table, the method further includes: identifying the parameter update request of the vehicle; after identifying the parameter update request, determining whether the evaluation interval to be updated meets the update conditions; if the evaluation interval to be updated meets the update conditions, updating the model parameters within the evaluation interval to be updated; otherwise, not updating the model parameters within the evaluation interval to be updated.

[0011] Based on the above technical means, in this embodiment of the application, after the vehicle sends a parameter update request, it is determined that the evaluation interval to be updated meets the update conditions. Only when the evaluation interval to be updated meets the update conditions will the model parameters be updated, thus avoiding resource waste.

[0012] Further, determining whether the evaluation interval to be updated meets the update conditions includes: obtaining the current ratio, average voltage, and current ratio distribution of the estimation error table for the evaluation interval to be updated; calculating the probability of occurrence of the current ratio in the evaluation interval to be updated based on the current ratio distribution; if the probability of occurrence is less than a probability threshold and the average voltage is less than a first error threshold, then it is determined that the evaluation interval to be updated does not meet the update conditions; otherwise, it is determined that the evaluation interval to be updated meets the update conditions.

[0013] Based on the above technical means, the embodiments of this application can determine whether the evaluation interval to be updated meets the update conditions by judging whether the probability of occurrence of the current multiplier and the average voltage of the interval to be evaluated are within a preset range.

[0014] Furthermore, determining the model parameters of the power battery based on the operating data and the model parameter table of the power battery includes: initializing a first database and a second database; sampling data at a preset distribution density for each evaluation interval in the model parameter table, updating the first database based on the sampled data, and identifying the model parameters in the first database; constructing data pairs based on the model parameters and the operating parameters, updating the data pairs to the second database, and determining the model parameters of the power battery based on the second database.

[0015] Based on the above technical means, the embodiments of this application can update the first database and the second database to determine the model parameters of the power battery.

[0016] Furthermore, updating the data pair to the second database includes: identifying whether the data pair has a data index in the second database; if the data index exists, updating the second database using exponential smoothing; if the data index does not exist, creating a data index for the data pair and updating the data pair and the data index to the second database.

[0017] Based on the above technical means, embodiments of this application can update the second database according to whether the corresponding data index exists and by using different update methods than the second database.

[0018] Furthermore, constructing the mapping function between the running parameters and the model parameters in the running data includes: identifying a first quantity of data in the second database and a second quantity of model parameters in the model parameter table; determining the error function of the fully connected neural network based on the first quantity and the second quantity, and determining the fully connected neural network based on the nonlinear activation function and the error function; determining the nonlinear mapping relationship between the running parameters and the model parameters based on the fully connected neural network, and constructing the mapping function between the running parameters and the model parameters based on the nonlinear mapping relationship.

[0019] Based on the above technical means, the embodiments of this application can construct a mapping function between operating parameters and model parameters through a nonlinear approximation method, so as to determine the confidence interval according to the mapping function and realize the update of model parameters within the evaluation interval.

[0020] Furthermore, obtaining the confidence interval of the mapping function includes: counting the amount of data sampled in each evaluation interval; calculating the confidence level of the evaluation interval based on the amount of data and the target amount of data in each evaluation interval; and determining the confidence interval of the mapping function based on the evaluation intervals with confidence levels greater than a preset value.

[0021] Based on the above technical means, the embodiments of this application can calculate the confidence level of the evaluation interval by the amount of data sampled in each evaluation interval and the target data amount, and determine the confidence interval of the mapping function based on the evaluation interval with a confidence level greater than a preset value.

[0022] Further, the step of calculating the confidence level of the evaluation interval based on the data volume and the target data volume of each evaluation interval includes: dividing each evaluation interval into multiple sub-intervals; calculating the data coverage of each sub-interval based on the data volume of each sub-interval and the target data volume, and determining whether the minimum coverage among all sub-intervals is less than a coverage threshold; if the minimum coverage among all sub-intervals is less than the coverage threshold, then the confidence level of the evaluation interval is determined to be a first confidence level, otherwise the confidence level of the evaluation interval is determined to be a second confidence level, wherein the first confidence level is less than the second confidence level.

[0023] A method for updating parameters of a power battery model, applied to a vehicle, comprising the following steps: acquiring operating data of the power battery and an estimation error table of model parameters; uploading the operating data and the estimation error table to a server, wherein the server determines the model parameters of the power battery based on the operating data and the model parameter table, constructs a mapping function between the operating parameters in the operating data and the model parameters, obtains the confidence interval of the mapping function, determines the evaluation interval to be updated in the model parameter table based on the confidence interval, updates the model parameters in the evaluation interval to be updated based on the mapping function and the estimation error table, and sends the updated model parameter table to the vehicle; and updating the model parameters of the power battery based on the updated model parameter table.

[0024] Based on the above technical means, the embodiments of this application can upload the operating data of the power battery and the estimation error table of model parameters to the server, so that the server can update the model parameter table. The server then sends the updated model parameter table to the vehicle, and the vehicle updates the power battery model parameters according to the model parameter table. The vehicle does not deploy a highly complex algorithm, which consumes less computing resources on the vehicle side compared to the online parameter identification scheme on the vehicle side, and is also more stable.

[0025] A server includes: a first acquisition module, configured to acquire operating data of a power battery uploaded by a vehicle and an estimation error table of model parameters; a construction module, configured to determine the model parameters of the power battery based on the operating data and the model parameter table of the power battery, construct a mapping function between the operating parameters in the operating data and the model parameters, and acquire a confidence interval of the mapping function; and a first update module, configured to determine an evaluation interval to be updated in the model parameter table based on the confidence interval, update the model parameters of the evaluation interval to be updated based on the mapping function and the estimation error table, and send the updated model parameter table to the vehicle, so that the vehicle updates the model parameters of the power battery based on the updated model parameter table.

[0026] Furthermore, the first update module is further configured to: identify the parameters to be updated in the evaluation interval to be updated and the corresponding operating parameters; calculate the model parameters of the operating parameters corresponding to the parameters to be updated using the mapping function, and identify the estimation errors of the parameters to be updated in the estimation error table; and update the parameters to be updated according to the model parameters calculated by the mapping function and the estimation errors.

[0027] Furthermore, the server also includes: a first identification module, configured to identify the vehicle's parameter update request before updating the model parameters within the evaluation interval to be updated according to the mapping function and the estimation error table; after identifying the parameter update request, determine whether the evaluation interval to be updated meets the update conditions; if the evaluation interval to be updated meets the update conditions, update the model parameters within the evaluation interval to be updated, otherwise do not update the model parameters within the evaluation interval to be updated.

[0028] Furthermore, the first identification module is further configured to: acquire the current ratio, average voltage, and current ratio distribution of the estimation error table for the evaluation interval to be updated; calculate the probability of occurrence of the current ratio in the evaluation interval to be updated based on the current ratio distribution; if the probability of occurrence is less than a probability threshold and the average voltage is less than a first error threshold, then determine that the evaluation interval to be updated does not meet the update condition; otherwise, determine that the evaluation interval to be updated meets the update condition.

[0029] Furthermore, the construction module is further configured to: initialize a first database and a second database; sample data with a preset distribution density for each evaluation interval in the model parameter table, update the first database based on the sampled data, and identify the model parameters in the first database; construct data pairs based on the model parameters and the operating parameters, update the data pairs to the second database, and determine the model parameters of the power battery based on the second database.

[0030] Furthermore, the construction module is further configured to: identify whether the data index of the data pair exists in the second database; if the data index exists, update the second database according to exponential smoothing; if the data index does not exist, create the data index of the data pair, and update the data pair and the data index to the second database.

[0031] Furthermore, the construction module is further configured to: identify a first quantity of data in the second database and a second quantity of model parameters in the model parameter table; determine the error function of the fully connected neural network based on the first quantity and the second quantity, and determine the fully connected neural network based on the nonlinear activation function and the error function; determine the nonlinear mapping relationship between the operating parameters and the model parameters based on the fully connected neural network, and construct the mapping function between the operating parameters and the model parameters based on the nonlinear mapping relationship.

[0032] Furthermore, the construction module is further configured to: count the amount of data sampled in each evaluation interval; calculate the confidence level of the evaluation interval based on the amount of data and the target amount of data in each evaluation interval; and determine the confidence interval of the mapping function based on the evaluation intervals with confidence levels greater than a preset value.

[0033] Furthermore, the construction module is further configured to: divide each evaluation interval into multiple sub-intervals; calculate the data coverage of each sub-interval based on the data volume of each sub-interval and the target data volume, and determine whether the minimum coverage among all sub-intervals is less than a coverage threshold; if the minimum coverage among all sub-intervals is less than the coverage threshold, then determine the confidence level of the evaluation interval as a first confidence level, otherwise determine the confidence level of the evaluation interval as a second confidence level, wherein the first confidence level is less than the second confidence level.

[0034] A vehicle includes: a second acquisition module for acquiring operating data of a power battery and an estimation error table of model parameters; an upload module for uploading the operating data and the estimation error table to a server, wherein the server determines the model parameters of the power battery based on the operating data and the model parameter table of the power battery, constructs a mapping function between the operating parameters in the operating data and the model parameters, acquires a confidence interval of the mapping function, determines an evaluation interval to be updated in the model parameter table based on the confidence interval, updates the model parameters in the evaluation interval to be updated based on the mapping function and the estimation error table, and sends the updated model parameter table to the vehicle; and a second update module for updating the model parameters of the power battery based on the updated model parameter table.

[0035] A computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement a parameter update method for a power battery model as described in the above embodiments.

[0036] The beneficial effects of this invention are:

[0037] This application embodiment updates the power battery model parameters based on vehicle-cloud interaction, fully leveraging the data advantages of the vehicle and the computing advantages of the cloud (server). Through vehicle-cloud collaboration, it achieves parameter optimization throughout the entire lifecycle of the power battery model parameters. It enables vehicle-side monitoring of model parameter matching, allowing for updates to model parameters with poor matching, thus addressing the issue of low accuracy in offline parameter calibration to some extent. For the vehicle side, only the error calculation part is added, without deploying highly complex algorithms. Compared to online parameter identification schemes on the vehicle side, it consumes fewer computing resources on the vehicle side and is more stable.

[0038] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0039] Figure 1 This is a flowchart illustrating a parameter update method for a server-side power battery model provided in an embodiment of the present invention.

[0040] Figure 2 This is a flowchart illustrating the parameter update method for a vehicle-side power battery model provided in an embodiment of the present invention.

[0041] Figure 3 A flowchart illustrating the overall process of a parameter update method for a power battery model according to an embodiment of the present invention.

[0042] Figure 4 This is a schematic diagram of a parameter update method for a power battery model that interacts between the vehicle and server sides, according to an embodiment of the present invention.

[0043] Figure 5 A block diagram of a server provided in an embodiment of the present invention;

[0044] Figure 6 A block diagram of a vehicle provided in an embodiment of the present invention. Detailed Implementation

[0045] The embodiments of the present invention will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention and not for limiting the scope of protection of the present invention.

[0046] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0047] The parameter update method for the power battery model described in this application is based on vehicle-cloud collaboration, that is, the parameters of the power battery model are updated simultaneously on both the vehicle side and the server side. Therefore, the following embodiments illustrate the parameter update method for the power battery model in this application from the perspectives of the server side and the vehicle side, respectively.

[0048] Specifically, Figure 1 This is a flowchart illustrating a parameter update method for a power battery model provided in an embodiment of this application.

[0049] like Figure 1 As shown, the parameter update method for this power battery model, applied to the server, includes the following steps:

[0050] In step S101, the operating data of the power battery uploaded by the vehicle and the estimation error table of model parameters are obtained.

[0051] The operating data of the power battery includes, but is not limited to, the extreme value of the single cell voltage, the single cell temperature, the state of charge (SOC), and the battery operating current.

[0052] It is understood that the embodiments of this application can obtain the operating data of the power battery uploaded by the vehicle and the estimation error of the model parameters, so as to subsequently determine the model parameters of the power battery.

[0053] In step S102, the model parameters of the power battery are determined based on the operating data and the model parameter table of the power battery, a mapping function between the operating parameters and the model parameters in the operating data is constructed, and the confidence interval of the mapping function is obtained.

[0054] The operating parameters include temperature, state of charge (SOC), and current ratio.

[0055] It is understood that, in the embodiments of this application, the model parameters of the power battery can be determined based on the operating data uploaded by the vehicle and the model parameter table of the power battery, a mapping function between the operating parameters and the model parameters can be constructed, and the confidence interval of the mapping function can be obtained so as to determine the evaluation interval to be updated in the model parameter table based on the confidence interval.

[0056] In this embodiment of the application, determining the model parameters of the power battery based on the operating data and the model parameter table of the power battery includes: initializing a first database and a second database; sampling data with a preset distribution density for each evaluation interval in the model parameter table, updating the first database based on the sampled data, and identifying the model parameters in the first database; constructing data pairs based on the model parameters and operating parameters, updating the data pairs to the second database, and determining the model parameters of the power battery based on the second database.

[0057] The evaluation interval consists of SOC and battery temperature. The first database can be named D1, and the second database can be named D2.

[0058] It should be noted that the server-side processing of the operational data uploaded by the vehicle requires setting a data statistics period T. kIn the process, the collected data uploaded by the vehicle is preprocessed, data cleaning and extraction are completed, and the database D1 is updated. k-1 Until database D1 is obtained k The current ratio distribution P in different error statistical intervals was sampled and statistically analyzed. k and the average voltage estimation error U within the interval ek Determine the current statistical period T k Is it finished? If so, read database D1. k The data is processed using an online identification algorithm to perform offline backtracking identification.

[0059] It is understood that the steps in this application embodiment to determine the model parameters of the power battery based on operating data and the model parameter table of the power battery include:

[0060] 1. Initialize the first database D1 and the second database D2;

[0061] 2. Sample the model parameter intervals using a preset distribution density. The formula for sampling the distribution density can be:

[0062]

[0063] Where K(.) is the Gaussian density function; h = 0.9, is the bandwidth of the kernel density estimation; c i The average value of the set charge / discharge rate range.

[0064] Update D1 based on the sampled data, and use an online identification algorithm to perform offline backtracking identification of the data in D1.

[0065] 3. Construct data pairs for model parameters and runtime parameters, using SOC and Temp as examples. avg The current ratio is the input variable, and the model parameters are the output variables. A data pair "(SOC, Temp" is constructed. avg The data pair of "current ratio - model parameters" is used in the form of SOC, Temp avg The current ratio is used as an index to update the data to D2, and the model parameters of the power battery are determined based on D2.

[0066] In this embodiment of the application, updating the data pair to the second database includes: identifying whether there is a data index for the data pair in the second database; if there is a data index, updating the second database according to exponential smoothing; if there is no data index, creating a data index for the data pair, and updating the data pair and the data index to the second database.

[0067] It is understood that the embodiments of this application use SOC and Temp. avgThe current multiplier is used as an index to update the second database. If an index exists, the second database is updated exponentially. If no index exists, a data index is created for the data pair, and the database and data index are written to the second database to complete the update of the second database.

[0068] Exponential smoothing is:

[0069] para t =ω*x t +(1-ω)*para t-1 ,

[0070] Where, para t Let ω be the decay weight after t updates, and x be the data after t updates. t The identification value is the value at update t.

[0071] In this embodiment of the application, constructing a mapping function between running parameters and model parameters in the running data includes: identifying a first quantity of data in the second database and a second quantity of model parameters in the model parameter table; determining the error function of the fully connected neural network based on the first and second quantities, and determining the fully connected neural network based on the nonlinear activation function and the error function; determining the nonlinear mapping relationship between the running parameters and model parameters based on the fully connected neural network, and constructing a mapping function between the running parameters and model parameters based on the nonlinear mapping relationship.

[0072] The nonlinear activation function can be any function with nonlinear characteristics, such as the ReLU function or the sigmoid function, without any specific restrictions.

[0073] It is understood that, in the embodiments of this application, the error function of the connected neural network can be determined based on the first quantity of data in the second database and the second quantity of model parameters in the model parameter table. Then, the fully connected neural network is determined using the nonlinear activation function and the error function. The nonlinear mapping relationship between the running parameters and the model parameters is determined based on the fully connected neural network. The mapping function between the running parameters and the model parameters is constructed based on the nonlinear mapping relationship.

[0074] Specifically, in this embodiment, data is read from a second database, and the running parameters (SOC, Temp) are fitted based on a nonlinear mapping function regression. avg The mapping function is obtained by relating the current ratio and model parameters.

[0075] The nonlinear mapping is implemented using a fully connected neural network F(x), and the calculation process for each layer is as follows:

[0076] para m×1 =f σ (w m×n xn×1 +b m×1 ),

[0077] Among them, f σ (·) represents a non-linear activation function, which can be a function with non-linear characteristics such as the ReLU function or the sigmoid function; x n×1 For input variables, corresponding runtime parameters (SOC, Temp) avg and current ratio); para m×1 For output variables, corresponding to model parameters; w m×n and b m×1 is the weight vector to be learned.

[0078] The error function can be calculated using the following formula, which characterizes both the absolute error and the similarity of the vector spaces:

[0079]

[0080] Where N is the number of corresponding data in the second database; m is the number of model parameters; This is the cosine value corresponding to the angle between the parameter vectors.

[0081] In this embodiment of the application, obtaining the confidence interval of the mapping function includes: counting the amount of data sampled in each evaluation interval; calculating the confidence level of the evaluation interval based on the amount of data and the target amount of data in each evaluation interval; and determining the confidence interval of the mapping function based on the evaluation intervals with confidence levels greater than a preset value.

[0082] The preset values ​​can be set according to specific circumstances, and there are no restrictions on them.

[0083] It is understood that this application can calculate the confidence level of the evaluation interval based on the amount of data sampled in each evaluation interval and the target amount of data in each evaluation interval. If the confidence level is greater than a preset value, the confidence interval of the mapping function is determined.

[0084] In this embodiment of the application, the confidence level of the evaluation interval is calculated based on the data volume and the target data volume of each evaluation interval, including: dividing each evaluation interval into multiple sub-intervals; calculating the data coverage of each sub-interval based on the data volume and target data volume of each sub-interval, and determining whether the minimum coverage among all sub-intervals is less than the coverage threshold; if the minimum coverage among all sub-intervals is less than the coverage threshold, the confidence level of the evaluation interval is determined to be the first confidence level, otherwise the confidence level of the evaluation interval is determined to be the second confidence level, wherein the first confidence level is less than the second confidence level.

[0085] The first confidence level can be represented by 0, and the second confidence level can be represented by 1.

[0086] Specifically, in this application embodiment, data coverage is used to evaluate the confidence level of the mapping function in different intervals. Let the error evaluation interval k be SOC∈[soc i ,soc j Temp avg ∈[T i ,T j If the granularity of SOC and temperature is set, the interval is subdivided into M smaller intervals. The data coverage of each smaller interval is defined as the ratio of the number of data points falling within that region in D2 to the target number of data points. The data coverage of interval k is represented by the lowest coverage among all smaller intervals. If the data coverage of interval k is less than the coverage threshold R... c If the confidence level is 0, then the confidence level of the mapping function for that evaluation interval is 0; otherwise, it is 1.

[0087] In step S103, the evaluation interval to be updated in the model parameter table is determined according to the confidence interval. The model parameters of the evaluation interval to be updated are updated according to the mapping function and the estimation error table. The updated model parameter table is then sent to the vehicle so that the vehicle updates the model parameters of the power battery according to the updated model parameter table.

[0088] It is understood that, in the embodiments of this application, the evaluation interval to be updated in the model parameters can be determined according to the confidence interval, the model parameters of the evaluation interval to be updated can be updated according to the mapping function and the estimation error table, and the updated model parameters can be sent to the vehicle through vehicle-cloud interaction, so that the vehicle can update the model parameters of the power battery according to the updated model parameter table.

[0089] In this embodiment of the application, updating the model parameters of the evaluation interval to be updated according to the mapping function and the estimation error table includes: identifying the parameters to be updated and the corresponding operating parameters of the evaluation interval to be updated; calculating the model parameters of the operating parameters corresponding to the parameters to be updated using the mapping function, and identifying the estimation errors of the parameters to be updated in the estimation error table; and updating the parameters to be updated according to the model parameters and estimation errors calculated by the mapping function.

[0090] It is understood that the embodiments of this application can use the mapping function to calculate the model parameters of the running parameters corresponding to the parameters to be updated in the evaluation interval to be updated, and update the parameters to be updated based on the model parameters and estimation error calculated by the mapping function.

[0091] The updated calculation formula is as follows:

[0092]

[0093] Where para represents the parameter to be updated. The parameter influence factors are calculated based on the battery equivalent model. but like but u err F(c) represents the estimation error uploaded from the vehicle; F(c) represents the model parameters obtained based on the mapping function.

[0094] In this embodiment of the application, before updating the model parameters in the evaluation interval to be updated according to the mapping function and the estimation error table, the method further includes: identifying the parameter update request of the vehicle; after identifying the parameter update request, determining whether the evaluation interval to be updated meets the update conditions; if the evaluation interval to be updated meets the update conditions, then the model parameters in the evaluation interval to be updated are updated, otherwise the model parameters in the evaluation interval to be updated are not updated.

[0095] It is understood that, before updating the model parameters in the evaluation interval to be updated according to the mapping function and the estimation error table, this application embodiment determines whether the evaluation interval to be updated meets the update condition. Only if the evaluation interval to be updated meets the update condition HU will the model parameters in the evaluation interval to be updated be updated; otherwise, no update will be performed.

[0096] Specifically, determining whether the evaluation interval to be updated meets the update conditions includes: obtaining the current ratio, average voltage, and current ratio distribution of the estimation error table for the evaluation interval to be updated; calculating the probability of occurrence of the current ratio in the evaluation interval to be updated based on the current ratio distribution; if the probability of occurrence is less than the probability threshold and the average voltage is less than the first error threshold, then the evaluation interval to be updated is determined not to meet the update conditions; otherwise, the evaluation interval to be updated is determined to meet the update conditions.

[0097] Among them, probability prediction can be used as P thd It is indicated that the first error threshold can be set according to specific circumstances, and there is no limitation on it.

[0098] It is understood that the embodiments of this application can obtain the current ratio, average voltage, and current ratio distribution of the estimation error table for the evaluation interval to be updated. The probability of occurrence of the current ratio in the evaluation interval to be updated is calculated based on the current ratio distribution. If the probability of occurrence is less than the probability threshold and the average voltage is less than the first error threshold, it is determined that the evaluation interval to be updated does not meet the update conditions, and the ratio is determined to be low probability and no parameter update is required. If the probability of occurrence is greater than or equal to the probability threshold and the average voltage is greater than or equal to the first error threshold, the ratio is determined to be the mainstream ratio and update calculation is performed.

[0099] The first error threshold can be set according to specific circumstances and is not limited thereto. The error threshold on the server side is for the evaluation error of the evaluation interval.

[0100] Specifically, the steps for determining whether the evaluation interval to be updated meets the conditions in this embodiment of the application are as follows:

[0101] After the server detects the parameter update request initiated by the vehicle, it obtains the error exceeding the limit range and the corresponding current ratio. The current ratio is matched within the range [c1, c2]. Based on the probability density function P... k Find the probability of this multiplier occurring:

[0102]

[0103] If P k |c∈[c1,c2] <P thd And the average error U ek If the value is less than the threshold, the multiplier is determined to be a low probability occurrence, and no parameter update is required.

[0104] If P k |c∈[c1,c2]>=P thd Or the average error U ek If the value is greater than or equal to the threshold, the cloud determines that the multiplier belongs to the mainstream usage multiplier and initiates parameter update calculation.

[0105] In this embodiment of the application, the operating parameters include one or more of the following: state of charge (SOC), battery temperature, current rate, and voltage.

[0106] In summary, the parameter update method for a power battery model applied to the server side as described in this application embodiment includes the following steps in the process of executing this update method on the server side (cloud):

[0107] S11: Set the data statistics period T, initialize databases D1 and D2, and the data coverage threshold R. c probability threshold P thd .

[0108] S12: In the statistical period T k In the process, the collected data uploaded by the vehicle is preprocessed, data cleaning and extraction are completed, and the database D1 is updated. k-1 Until database D1 is obtained k The current ratio distribution P in different error statistical intervals was sampled and statistically analyzed. k and the average voltage estimation error U within the interval ek .

[0109] Database D1 k-1 Corresponding to T k-1 All valid data collected periodically, in T k During the cycle, newly acquired valid data replaces D1. k-1 Record the earliest data in the middle, up to T. k The cycle has ended. In cycle T... k-1 After completion, for each error estimation interval, the distribution density is sampled using the following formula:

[0110]

[0111] Where K(.) is the kernel function, which is non-negative, has an integral of 1, a mean of 0, and conforms to probability density properties. Commonly chosen density functions include Gaussian, rectangular, and Epanechnikov functions; h is a smoothing parameter greater than 0, representing the bandwidth of the kernel density estimate; c i For the set charge / discharge rate, the battery pack can be divided into n intervals based on its designed charge / discharge rate, c. i You can take the mean or boundary value of each interval.

[0112] set up and They correspond to the same error statistical interval, but the former is based on T. k-1 D1 of the period k-1 The calculation shows that the latter is based on T. k D1 of the period k Calculations yielded a value T. k Periodic current ratio distribution P k as follows:

[0113]

[0114] Where k is T k The period index, which takes the value of an integer.

[0115] S13: Determine the current statistical period T k Is it finished? If so, read database D1. k The data is processed using online identification algorithms to perform offline backtracking identification, constructing a system (SOC, Temp) avg Update database D2 with the data pair “(current ratio) - model parameters”; otherwise return to S12.

[0116] The online identification algorithms used include, but are not limited to, recursive least squares methods and recursive least squares methods with forgetting factors, for D1 k From the data, model parameters can be identified, such as SOC and Temp. avg The current multiplier is the input variable, and the model parameters are the output variables. Data pairs are constructed using SOC and Temp. avg The current multiplier is used as an index; update D2.

[0117] If the index already exists in D2, then the parameters are updated using exponential smoothing, i.e.:

[0118] para t =ω*x t +(1-ω)*para t-1 ,

[0119] Where, para t Let x be the data after t updates, ω be the decay weight, and x be the value of the decay weight. t The identification value at update t;

[0120] If the index does not exist in D2, then write directly to D2 to complete the update.

[0121] S14: Read data from database D2 and perform regression fitting based on nonlinear mapping (SOC, Temp). avg The relationship between the current ratio and model parameters is used to obtain the mapping function and its confidence interval.

[0122] The nonlinear mapping is implemented using a fully connected neural network F(x), and the calculation process for each layer is as follows:

[0123] para m×1 =f σ (w m×n x n×1 +b m×1 ),

[0124] Among them, f σ (·) represents a non-linear activation function, which can be a function with non-linear characteristics such as the ReLU function or the sigmoid function; x n×1 For input variables, corresponding to SOC and Temp avg and current ratio; para m×1 For output variables, corresponding to model parameters; w m×n and b m×1 is the weight vector to be learned.

[0125] The network's error function is characterized by both absolute error and vector space similarity:

[0126]

[0127] Where N is the number of corresponding data in D2; m is the number of model parameters; The cosine value corresponding to the angle between the parameter vectors;

[0128] Data coverage is used to evaluate the confidence level of the mapping function in different intervals. Let the error evaluation interval k be SOC∈[soc] i ,soc j Temp avg ∈[T i ,T jThen, set the granularity of SOC and temperature, subdividing the interval into M smaller intervals. The data coverage of each smaller interval is defined as the ratio of the number of data points falling within that region in D2 to the target number of data points. The data coverage of interval k is characterized by the lowest coverage among all smaller intervals. If the data coverage of interval k is less than the threshold R in S11... c If the confidence level is 0, then the confidence level of the mapping function for that evaluation interval is 0; otherwise, it is 1.

[0129] S15: Evaluation: Condition 1 (vehicle initiates parameter update request), Condition 2 (does U exist) ek If the condition (exceeding the limit) is met, the parameter update logic is triggered.

[0130] After the cloud identifies the parameter update request initiated by the vehicle, it obtains the error exceeding the limit range and the corresponding current ratio. The current ratio is matched within the range [c1, c2], and then the probability density function P is used to determine the appropriate current ratio. k To determine the probability of this multiplier occurring, the parameter update logic is as follows:

[0131]

[0132] If P k |c∈[c1,c2] <P thd And the average error U ek If the value is less than the threshold, the multiplier is determined to be a low probability occurrence, and no parameter update is required.

[0133] If P k |c∈[c1,c2]>=P thd Or the average error U ek If the value is greater than or equal to the threshold, the cloud determines that the multiplier belongs to the mainstream usage multiplier and initiates parameter update calculation in the cloud.

[0134] First, determine whether the parameter mapping function is reliable within this error interval:

[0135] If reliable, update all parameter elements in the corresponding error range in the cloud model parameter table as follows:

[0136]

[0137] Where para represents the parameter to be updated. The parameter influence factors are calculated based on the battery equivalent model. but like but u err F(c) represents the estimation error uploaded from the vehicle; F(c) represents the model parameters obtained based on the mapping function.

[0138] S16: Based on the execution results of logic 1 and 2 in S15, update the cloud parameter table and determine whether data needs to be sent to update the vehicle parameters.

[0139] According to the parameter update method for the power battery model proposed in this application, the corresponding model parameters can be determined from the data uploaded by the vehicle, a mapping relationship between the operating parameters and the model parameters can be constructed, the confidence interval obtained by the mapping function can be used to determine the evaluation interval to be updated, and after the model parameters are updated, the model parameter table is sent to the vehicle, enabling the vehicle to update the power battery model parameters. This realizes the interaction of model parameters between the vehicle and the server, makes full use of the advantages of server-side computing, and realizes the updating of model parameters with poor matching. To a certain extent, it solves the problem of low accuracy of offline parameter calibration and realizes the optimization of the power battery model parameters throughout the entire life cycle.

[0140] The above embodiments focus on introducing the parameter update method of the power battery model from the server side. The following embodiments will explain the parameter update method of the power battery model from the vehicle side.

[0141] like Figure 2 As shown, the parameter update method for this power battery model, applied to a vehicle, includes the following steps:

[0142] In step S201, the operating data of the power battery and the estimation error table of model parameters are obtained.

[0143] The method for obtaining the operating data and model parameter estimation error table of the power battery includes: periodically collecting the operating data of the power battery; calculating the estimated voltage of the power battery based on the operating data, and calculating the absolute error between the estimated voltage and the actual voltage of the power battery; if the absolute error is greater than the voltage error value in the estimation error table, then the estimation error table is updated based on the absolute error, otherwise the estimation error table is not updated.

[0144] In the estimation error table, the indexes of each element are SOC and average temperature, and each element consists of the error and the corresponding current multiplier.

[0145] It is understood that the embodiments of this application can periodically collect the operating data of the power battery, including but not limited to the extreme values ​​of single cell voltage, single cell temperature, SOC and battery operating current, etc. The sampling period is kept uniform. The estimated voltage of the power battery is calculated based on the operating data, and the absolute difference between the estimated voltage and the average single cell voltage is calculated. If the absolute error is greater than the voltage error value in the error table, the estimation error table is updated based on the absolute error; otherwise, the estimation error table is not updated.

[0146] Specifically, the logic for determining whether the estimation error table should be updated in this application embodiment is as follows:

[0147] Based on the current cycle's SOC and average temperature Temp avg Estimate the corresponding element (U) in the index query model error table. err C), compare the estimation error U of the current period. e ′ rr Compare the voltage error values ​​in the error table:

[0148] If U e ′ rr >U err If U is updated, the error in the error estimation table is updated to the estimation error for the current cycle, and the corresponding current multiplier is recorded; e ′ rr ≤U err If not, the error estimation table will not be updated.

[0149] In this embodiment of the application, before obtaining the operating data of the power battery and the estimation error table of the model parameters, the method further includes: obtaining the equivalent model of the power battery; calibrating the model parameter table of the power battery according to the equivalent model; and synchronously updating the calibrated model parameter table to the server.

[0150] It is understood that, before obtaining the operating data of the power battery and the estimation error table of the model parameters, the equivalent model of the vehicle can be determined in the embodiments of this application. The model parameter table of the power battery can be calibrated according to the equivalent model, and the calibrated model parameter table can be synchronously updated to the server.

[0151] It should be noted that the equivalent model for the vehicle can be, but is not limited to, a first-order RC equivalent model, a second-order RC equivalent model, and a fractional-order model. The initial calibration of the model relies on data obtained through HPPC experiments. The experiments need to consider three dimensions: temperature, state of charge (SOC), and current rate. Due to the limitations of the vehicle's storage space, experiments can be conducted at a uniform current rate. However, the SOC needs to cover the actual usage range, and the temperature needs to cover normal temperature, low temperature, and high temperature. After the initial calibration is completed, the original model parameter table is obtained. The parameter tables on both the vehicle and cloud sides need to be updated synchronously to ensure the effectiveness of subsequent parameter updates in the cloud.

[0152] In step S202, the operating data and estimation error table are uploaded to the server. The server determines the model parameters of the power battery based on the operating data and the model parameter table of the power battery, constructs a mapping function between the operating parameters and the model parameters in the operating data, obtains the confidence interval of the mapping function, determines the evaluation interval to be updated in the model parameter table based on the confidence interval, updates the model parameters in the evaluation interval to be updated based on the mapping function and the estimation error table, and sends the updated model parameter table to the vehicle.

[0153] It is understood that, in this embodiment of the application, the operating data of the power battery and the estimation error table of the model parameters obtained in step S201 above are uploaded to the server so that the server can update the model parameters of the power battery. The method of updating the model parameters of the power battery by the server has been described in the above embodiments and will not be repeated here.

[0154] In this embodiment of the application, before uploading the running data and the estimation error table to the server, the method further includes: identifying whether any parameter in the estimation error table has an update flag; if an update flag exists, generating an update request for the model parameter table and uploading the update request to the server.

[0155] It is understood that, before uploading the running data and the estimation error table to the server, this application embodiment determines whether any parameter in the estimation error table has an update flag. If an update flag exists, the vehicle will generate an update request for the model parameter table and upload the update request to the server.

[0156] In this embodiment of the application, identifying whether any model parameter in the estimation error table has an update flag includes: identifying the voltage extreme value of each estimation interval; if the voltage extreme value is greater than the second error threshold, then generating an update flag for the model parameter in the estimation interval.

[0157] The second error threshold can be set according to specific circumstances and is not limited thereto. The voltage error threshold at the vehicle end is the extreme value of the error for the evaluation range.

[0158] It is understood that the embodiments of this application can count whether there are voltage extreme values ​​in the error table that exceed the second error threshold. If so, the corresponding interval parameter update requirement flag position is 1 and maintained; otherwise, the corresponding flag position is 0, so as to be used for subsequent updating of model parameters.

[0159] In step S203, the model parameters of the power battery are updated according to the updated model parameter table.

[0160] In this embodiment, the model parameters of the power battery are updated based on the updated model parameter table issued by the server.

[0161] In summary, the parameter update method for the power battery model applied to the server side as described in this application embodiment, and the process of executing this update method on the vehicle side (vehicle side) specifically include the following steps:

[0162] S21: Determine the equivalent model on the vehicle side, complete the initial calibration of the battery model parameters on the vehicle side, and synchronously update the model parameters on the vehicle side and the cloud.

[0163] Based on the parameter evaluation range, initialize the model estimation error table. The index of each element in the table is SOC and average temperature, and each element consists of the error and the corresponding current multiplier.

[0164] S22: Periodically collect power battery operation data and calculate the absolute difference between the model estimated voltage and the average voltage of individual cells in the current cycle.

[0165] S23: Update the vehicle-side model estimation error table based on the average temperature and state of charge of the current cycle.

[0166] The logic for updating the vehicle-side model estimation error table is as follows:

[0167] Based on the current cycle's SOC and average temperature Temp avg Estimate the corresponding element (U) in the index query model error table. err C), compare the estimation error U of the current period. e ′ rr The magnitude of the error corresponding to the element:

[0168] If U e ′ rr >U err If so, update the error in the element to the estimated error of the current period and record the corresponding current multiplier;

[0169] If U e ′ rr ≤U err If not, the element will not be updated.

[0170] S24: Are there any elements in the statistical error table whose errors exceed the threshold? If so, update the corresponding interval parameter to the required flag position 1 and keep it; otherwise, the corresponding flag position is 0.

[0171] S25: Upload vehicle-side data to the cloud.

[0172] S26: The cloud effectively cleans the data uploaded from the vehicle and calculates the corresponding parameters of the battery pack based on the cleaned data. It constructs the mapping relationship between SOC, temperature, and rate of change and model parameters through a nonlinear approximation method, comprehensively evaluates the necessity of parameter updates in each segment of the vehicle, and sends data to the vehicle to complete the parameter update when the update requirements are met.

[0173] According to the parameter update method for the power battery model proposed in the embodiments of this application, the operating data of the power battery and the estimation error table of the model parameters can be uploaded to the server, enabling the server to update the model parameter table. The server then distributes the updated model parameter table to the vehicle, and the vehicle updates the power battery model parameters according to the model parameter table. The vehicle does not deploy highly complex algorithms, which consumes less computing resources on the vehicle side compared to the online parameter identification scheme on the vehicle side, while also being more stable.

[0174] The parameter update methods for the power battery model described in the above embodiments are applied to both the server (cloud) and the vehicle (vehicle), respectively, to update the parameters of the power battery model through vehicle-cloud collaboration. The following specific embodiment illustrates the parameter update method for the power battery model in the vehicle-cloud system of this application. The overall execution flowchart of the method is as follows: Figure 3 As shown, the vehicle-cloud collaborative interaction process is as follows: Figure 4 As shown, this can be specifically applied to the battery management system of electric vehicles, and its implementation includes the following steps:

[0175] S1: Determine the equivalent model on the vehicle side, complete the initial calibration of the battery model parameters on the vehicle side, and synchronously update the model parameters on the vehicle side and the cloud.

[0176] For example, a first-order RC model is selected as the equivalent circuit model at the vehicle end. Through HPPC experiments, the Ro, Rp, and Cp parameters of the RC model are obtained through offline calibration. A two-dimensional parameter table is constructed with SOC and Temp, where SOC is divided into: [0 5 10 20 30 40 50 60 70 80 90 95 100] and Temp is divided into [-20 -10 0 10 25 40]. The parameter table is updated synchronously at the vehicle end and the cloud.

[0177] S2: Divide the parameter evaluation interval and determine the second error threshold of vehicle-end voltage and the first error threshold of cloud voltage.

[0178] The exemplary parameter evaluation interval is consistent with the granularity of the parameter table, that is, each minimum interval in the parameter table is a parameter evaluation interval. The vehicle-end voltage error threshold is set as u1, and the cloud-end voltage error threshold is u2.

[0179] S3: Vehicle-side: Periodically collects power battery operation data, evaluates model estimation errors in different ranges, and uploads vehicle-side data to the cloud.

[0180] S31: Initialize the model estimation error table based on the parameter evaluation interval.

[0181] S32: Periodically collect power battery operating data and calculate the difference between the model estimated voltage and the average voltage of individual cells in the current cycle.

[0182] S33: Update the vehicle-side model estimation error table based on the average temperature and state of charge of the current cycle.

[0183] Using the current period's SOC and average temperature Tempavg as indexes, query the corresponding element (U) in the model estimation error table. err C), compare the estimation error U of the current period. e ′ rr The numerical magnitude of the error corresponding to the element:

[0184] If U e ′ rr >U err If so, update the error in the element to the estimated error of the current period and record the corresponding current multiplier;

[0185] If U e ′ rr ≤U err If not, the element will not be updated.

[0186] S34: Check if there are any elements in the error table whose errors exceed the threshold. If so, update the parameter requirement flag to 1 and keep it; otherwise, keep the flag at 0.

[0187] S35: Upload vehicle-side data to the cloud.

[0188] The uploaded data includes regular data collected from the vehicle and parameter update request messages. The messages specifically include update request flags, current error over-limit range, and current ratio.

[0189] S4: Cloud: Preprocess vehicle-mounted data, identify model parameters, update and maintain the model parameter table, and comprehensively evaluate the necessity of updating vehicle-mounted parameters.

[0190] S41: Set the data statistics period T, initialize databases D1 and D2, and set the data coverage threshold R. c probability threshold P thd .

[0191] S42: In the statistical period T k In the process, the collected data uploaded by the vehicle is preprocessed, data cleaning and extraction are completed, and the database D1 is updated. k-1 Until database D1 is obtained k The current ratio distribution P in different error statistical intervals was sampled and statistically analyzed. k and the average voltage estimation error U within the interval ek .

[0192] For example, the current ratio c is divided into 8 intervals: (-∞,-2C], (-2C,-1C], (-1,-0.3C], (-0.3C,0], (0,0.3C], (0.3,1C], (1,2C], and (2C,+∞].

[0193] Database D1 k-1 For all valid data collected in period Tk-1, in period Tk, newly collected valid data replaces D1. k-1 The earliest data is recorded until the end of period Tk. After period Tk-1 ends, for each error estimation interval, the distribution density is sampled according to the following formula:

[0194]

[0195] Where K(.) is the Gaussian density function; h = 0.9, is the bandwidth of the kernel density estimation; c i The average value of the set charge / discharge rate range.

[0196] set up and They correspond to the same error statistical interval, but the former is based on T. k-1 D1 of the period k-1 The calculation shows that the latter is based on T. k D1 of the period k The calculated current ratio distribution P for period Tk is then obtained. k for:

[0197]

[0198] Where k is T k The period index, which takes the value of an integer.

[0199] S43: Determine the current statistical period T k Is it finished? If so, read database D. k The data is processed using online identification algorithms to perform offline backtracking identification, constructing a system (SOC, Temp) avg Update database D2 with the data pair "(current ratio) - model parameters"; otherwise return to S42.

[0200] An exemplary use of the recursive least squares method with a forgetting factor, for D1 k From the data, identify model parameters, using SOC and Temp. avg The current multiplier is the input variable, and the model parameters are the output variables. Data pairs are constructed using SOC and Temp. avg The current multiplier is the index; update D2:

[0201] If the index already exists in D2, then the parameters are updated using exponential smoothing, i.e.:

[0202] para t =ω*x t +(1-ω)*para t-1 ,

[0203] If the index does not exist in D2, then write directly to D2 to complete the update.

[0204] S44: Read data from database D2 and perform regression fitting based on nonlinear mapping (SOC, Temp). avg The relationship between the current ratio and model parameters is used to obtain the mapping function and its confidence interval;

[0205] The nonlinear mapping is implemented using a fully connected neural network F(x), and the calculation process for each layer is as follows:

[0206] para 3×1 =f σ (w 3×3 x 3×1 +b 3×1 ),

[0207] Among them, f σ (·) Select the ReLU function; x n×1 For input variables, corresponding to SOC, Tempavg, and current ratio; para m×1 For output variables, in this embodiment they correspond to [Ro, Rp, Cp]; w 3×3 and b 3×1 is the weight vector to be learned.

[0208] Based on the data in D2, the parameters of the fully connected neural network are optimized by backpropagating the error using the following formula:

[0209]

[0210] Where N is the number of corresponding data in D2; This is the cosine value corresponding to the angle between the parameter vectors.

[0211] Data coverage is used to evaluate the confidence level of the mapping function in different intervals. For example, if the error evaluation interval is SOC∈[10,20], Temp avg ∈[10,25], set SOC granularity to 1%, Temp avgThe granularity is 1℃, and the interval is subdivided into 176 smaller intervals. Each smaller interval is required to contain 100 data points. Data coverage is defined as the ratio of the number of data points falling within the specified region in D2 to the target number of data points. The data coverage of the entire interval is characterized by the lowest coverage among all smaller intervals. If the data coverage of an interval is less than the threshold R in S41... c If the confidence level is 0, then the confidence level of the mapping function for that evaluation interval is 0; otherwise, it is 1.

[0212] S45: Evaluation: Condition 1 (vehicle initiates parameter update request), Condition 2 (does U exist) ek If the condition (exceeding the limit) is met, the parameter update logic is triggered.

[0213] After the cloud identifies the parameter update request initiated by the vehicle, it obtains the error exceeding the limit range and the corresponding current ratio. The current ratio is matched within the range [c1, c2], and then the probability density function P is used to determine the appropriate current ratio. k Find the probability of this multiplier occurring:

[0214]

[0215] If P k |c∈[c1,c2] <P thd And the average error U ek If the value is less than the threshold, the multiplier is determined to be a low probability occurrence, and no parameter update is required.

[0216] If P k |c∈[1,c2]>=P thd Or the average error U ek If the value is greater than or equal to the threshold, the cloud determines that the multiplier belongs to the mainstream usage multiplier and initiates parameter update calculation.

[0217] Determine whether the parameter mapping function is reliable within the error interval.

[0218] If reliable, update all parameter elements in the corresponding error range in the cloud model parameter table as follows:

[0219]

[0220] Where para represents the parameter to be updated. The parameter influence factors are calculated based on the battery equivalent model. but like but u err F(c) represents the estimation error uploaded from the vehicle; F(c) represents the model parameters obtained based on the mapping function.

[0221] If the information is unreliable, no update calculation will be performed.

[0222] S46: Based on the logic execution result in S45, update the cloud parameter table and determine whether data needs to be sent to update the vehicle parameters.

[0223] If the parameter update requirement is met in S45, the parameter table in the cloud is updated according to the execution result, and then sent to the vehicle in a timely manner through vehicle-cloud data interaction.

[0224] S5: Cloud data is distributed, and the vehicle verifies the data validity and updates the model parameters.

[0225] For data sent from the cloud, the vehicle verifies the validity of the data and updates the storage when the vehicle is powered off.

[0226] In summary, the power battery model parameter update method based on vehicle-cloud collaboration described in this embodiment has the following effects:

[0227] (1) The vehicle-side monitoring of the matching of model parameters has been realized, and parameters with poor matching can be updated, which can solve the problem of low accuracy of offline parameter calibration to a certain extent.

[0228] (2) The vehicle-side solution is basically the same as the offline calibration solution, except that the model error is calculated. No complex algorithms are deployed. Compared with the vehicle-side online parameter identification solution, it consumes less vehicle-side computing resources and is more stable.

[0229] (3) It has opened up the interaction logic between the vehicle and the cloud in terms of model parameter identification, giving full play to the advantages of vehicle data and cloud computing. Through vehicle-cloud collaboration, it can achieve full life cycle optimization of electric vehicle model parameters.

[0230] Next, the server proposed according to the embodiments of this application is described with reference to the accompanying drawings.

[0231] Figure 5 This is a block diagram of the server according to an embodiment of this application.

[0232] like Figure 5 As shown, the server 10 includes: a first acquisition module 101, a construction module 102, and a first update module 103.

[0233] The first acquisition module 101 is used to acquire the operating data of the power battery uploaded by the vehicle and the estimation error table of model parameters; the construction module 102 is used to determine the model parameters of the power battery according to the operating data and the model parameter table of the power battery, construct the mapping function between the operating parameters and the model parameters in the operating data, and obtain the confidence interval of the mapping function; the first update module 103 is used to determine the evaluation interval to be updated in the model parameter table according to the confidence interval, update the model parameters of the evaluation interval to be updated according to the mapping function and the estimation error table, and send the updated model parameter table to the vehicle, so that the vehicle updates the model parameters of the power battery according to the updated model parameter table.

[0234] In this embodiment of the application, the first update module 103 is further configured to: identify the parameters to be updated in the evaluation interval to be updated and the corresponding operating parameters; calculate the model parameters of the operating parameters corresponding to the parameters to be updated using a mapping function, and identify the estimation errors of the parameters to be updated in the estimation error table; and update the parameters to be updated according to the model parameters and estimation errors calculated by the mapping function.

[0235] In this embodiment of the application, the server 10 further includes a first identification module.

[0236] The first identification module is used to identify the vehicle's parameter update request before updating the model parameters in the evaluation interval to be updated according to the mapping function and the estimation error table; after identifying the parameter update request, it determines whether the evaluation interval to be updated meets the update conditions; if the evaluation interval to be updated meets the update conditions, the model parameters in the evaluation interval to be updated are updated, otherwise the model parameters in the evaluation interval to be updated are not updated.

[0237] In this embodiment of the application, the first identification module is further configured to: obtain the current ratio, voltage mean, and current ratio distribution of the estimation error table of the evaluation interval to be updated; calculate the probability of occurrence of the current ratio of the evaluation interval to be updated based on the current ratio distribution; if the probability of occurrence is less than the probability threshold and the voltage mean is less than the first error threshold, then determine that the evaluation interval to be updated does not meet the update conditions; otherwise, determine that the evaluation interval to be updated meets the update conditions.

[0238] In this embodiment of the application, the construction module 102 is further configured to: initialize the first database and the second database; perform data sampling with a preset distribution density on each evaluation interval in the model parameter table, update the first database according to the sampled data, and identify the model parameters in the first database; construct data pairs according to the model parameters and operating parameters, update the data pairs to the second database, and determine the model parameters of the power battery according to the second database.

[0239] In this embodiment of the application, the construction module 102 is further configured to: identify whether there is a data index for the data pair in the second database; if there is a data index, update the second database according to exponential smoothing; if there is no data index, create a data index for the data pair, and update the data pair and data index to the second database.

[0240] In this embodiment, the construction module 102 is further configured to: identify a first quantity of data in the second database and a second quantity of model parameters in the model parameter table; determine the error function of the fully connected neural network based on the first and second quantities, and determine the fully connected neural network based on the nonlinear activation function and the error function; determine the nonlinear mapping relationship between the running parameters and the model parameters based on the fully connected neural network, and construct the mapping function between the running parameters and the model parameters based on the nonlinear mapping relationship.

[0241] In this embodiment, the construction module 102 is further configured to: count the amount of data sampled in each evaluation interval; calculate the confidence level of the evaluation interval based on the amount of data and the target amount of data in each evaluation interval; and determine the confidence interval of the mapping function based on the evaluation intervals with confidence levels greater than a preset value.

[0242] In this embodiment of the application, the construction module 102 is further configured to: divide each evaluation interval into multiple sub-intervals; calculate the data coverage of each sub-interval based on the data volume and target data volume of each sub-interval, and determine whether the minimum coverage among all sub-intervals is less than the coverage threshold; if the minimum coverage among all sub-intervals is less than the coverage threshold, then determine the confidence level of the evaluation interval as the first confidence level, otherwise determine the confidence level of the evaluation interval as the second confidence level, wherein the first confidence level is less than the second confidence level.

[0243] It should be noted that the explanation of the parameter update method embodiment for the power battery model described above also applies to the server of this embodiment, and will not be repeated here.

[0244] According to the server proposed in this application embodiment, it can determine the corresponding model parameters for the data uploaded by the vehicle, construct the mapping relationship between the operating parameters and the model parameters, determine the evaluation interval to be updated using the confidence interval obtained by the mapping function, and after updating the model parameters, send the model parameter table to the vehicle, enabling the vehicle to update the power battery model parameters. This realizes the interaction of model parameters between the vehicle and the server, fully utilizes the advantages of server-side computing, and realizes the updating of model parameters with poor matching. To a certain extent, it solves the problem of low accuracy of offline parameter calibration and realizes the optimization of the power battery model parameters throughout the entire life cycle.

[0245] Figure 6 This is a block diagram of a vehicle according to an embodiment of this application.

[0246] like Figure 6 As shown, the vehicle 20 includes: a second acquisition module 201, an upload module 202, and a second update module 203.

[0247] The second acquisition module 201 is used to acquire the operating data of the power battery and the estimation error table of model parameters; the upload module 202 is used to upload the operating data and the estimation error table to the server, wherein the server determines the model parameters of the power battery based on the operating data and the model parameter table of the power battery, constructs a mapping function between the operating parameters and the model parameters in the operating data, obtains the confidence interval of the mapping function, determines the evaluation interval to be updated in the model parameter table based on the confidence interval, updates the model parameters in the evaluation interval to be updated based on the mapping function and the estimation error table, and sends the updated model parameter table to the vehicle; the second update module 203 is used to update the model parameters of the power battery based on the updated model parameter table.

[0248] It should be noted that the foregoing explanation of the parameter update method embodiment for the power battery model also applies to the vehicle in this embodiment, and will not be repeated here.

[0249] According to the vehicle proposed in the embodiments of this application, the operating data of the power battery and the estimation error table of model parameters can be uploaded to the server, enabling the server to update the model parameter table. The server then distributes the updated model parameter table to the vehicle, and the vehicle updates the power battery model parameters according to the model parameter table. The vehicle does not deploy highly complex algorithms, which consumes less computing resources on the vehicle side compared to the online parameter identification scheme on the vehicle side, while also being more stable.

[0250] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the parameter update method for the power battery model described above.

[0251] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0252] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0253] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0254] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (FPGAs), field-programmable gate arrays (FPGAs), etc.

[0255] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0256] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A parameter update method for a power battery model, characterized in that, The method is applied to a server, and the method includes the following steps: Obtain the operational data of the power battery uploaded by the vehicle and the estimation error table of model parameters; The model parameters of the power battery are determined based on the operating data and the model parameter table of the power battery. A mapping function between the operating parameters in the operating data and the model parameters is constructed, and the confidence interval of the mapping function is obtained. The evaluation interval to be updated in the model parameter table is determined based on the confidence interval. The model parameters of the evaluation interval to be updated are updated according to the mapping function and the estimation error table. The updated model parameter table is then sent to the vehicle so that the vehicle updates the model parameters of the power battery according to the updated model parameter table. The step of determining the model parameters of the power battery based on the operating data and the model parameter table of the power battery includes: Initialize the first and second databases; Data sampling with a preset distribution density is performed on each evaluation interval in the model parameter table. The first database is updated based on the sampled data, and the model parameters in the first database are identified. A data pair is constructed based on the model parameters and the operating parameters, the data pair is updated to the second database, and the model parameters of the power battery are determined based on the second database. The construction of the mapping function between the running parameters and the model parameters in the running data includes: Identify the first quantity of data in the second database and the second quantity of model parameters in the model parameter table; The error function of the fully connected neural network is determined based on the first quantity and the second quantity, and the fully connected neural network is determined based on the nonlinear activation function and the error function; The nonlinear mapping relationship between the running parameters and the model parameters is determined based on the fully connected neural network, and a mapping function between the running parameters and the model parameters is constructed based on the nonlinear mapping relationship.

2. The parameter update method for the power battery model according to claim 1, characterized in that, The step of updating the model parameters of the evaluation interval to be updated according to the mapping function and the estimation error table includes: Identify the parameters to be updated in the evaluation interval to be updated and the corresponding operating parameters; The model parameters of the running parameters corresponding to the parameters to be updated are calculated using the mapping function, and the estimation errors of the parameters to be updated in the estimation error table are identified. The parameters to be updated are updated based on the model parameters calculated by the mapping function and the estimation error.

3. The parameter update method for the power battery model according to claim 1 or 2, characterized in that, Before updating the model parameters within the evaluation interval to be updated based on the mapping function and the estimation error table, the method further includes: Identify the vehicle's parameter update request; After the parameter update request is detected, it is determined whether the evaluation interval to be updated meets the update conditions. If the evaluation interval to be updated meets the update condition, then the model parameters in the evaluation interval to be updated are updated; otherwise, the model parameters in the evaluation interval to be updated are not updated.

4. The parameter update method for the power battery model according to claim 3, characterized in that, The determination of whether the evaluation interval to be updated meets the update conditions includes: Obtain the current ratio, average voltage, and current ratio distribution of the estimation error table for the evaluation interval to be updated; The probability of the occurrence of the current ratio in the evaluation interval to be updated is calculated based on the current ratio distribution. If the probability of occurrence is less than the probability threshold and the average voltage is less than the first error threshold, then the evaluation interval to be updated is determined to not meet the update conditions; otherwise, the evaluation interval to be updated is determined to meet the update conditions.

5. The parameter update method for the power battery model according to claim 1, characterized in that, The step of updating the data pair to the second database includes: Identify whether the data index of the data pair exists in the second database; If the data index exists, the second database is updated using exponential smoothing. If the data index does not exist, then a data index is created for the data pair, and the data pair and the data index are updated in the second database.

6. The parameter update method for the power battery model according to claim 1, characterized in that, Obtaining the confidence interval of the mapping function includes: The amount of data sampled in each evaluation interval is counted. The confidence level of the evaluation interval is calculated based on the data volume and the target data volume of each evaluation interval; The confidence interval of the mapping function is determined based on the evaluation interval where the confidence level is greater than a preset value.

7. The parameter update method for the power battery model according to claim 6, characterized in that, The step of calculating the confidence level of the evaluation interval based on the data volume and the target data volume of each evaluation interval includes: Each evaluation interval is divided into multiple sub-intervals; The data coverage of each sub-interval is calculated based on the data volume of each sub-interval and the target data volume, and it is determined whether the minimum coverage of all sub-intervals is less than the coverage threshold. If the minimum coverage among all sub-intervals is less than the coverage threshold, then the confidence level of the evaluation interval is determined to be the first confidence level; otherwise, the confidence level of the evaluation interval is determined to be the second confidence level, wherein the first confidence level is less than the second confidence level.

8. A parameter update method for a power battery model, characterized in that, The method is applied to a vehicle, and the method includes the following steps: Obtain the operating data of the power battery and the estimation error table of model parameters; The server uploads the operating data and the estimation error table to the server. The server determines the model parameters of the power battery based on the operating data and the model parameter table of the power battery, constructs a mapping function between the operating parameters in the operating data and the model parameters, obtains the confidence interval of the mapping function, determines the evaluation interval to be updated in the model parameter table based on the confidence interval, updates the model parameters in the evaluation interval to be updated based on the mapping function and the estimation error table, and sends the updated model parameter table to the vehicle. The model parameters of the power battery are updated according to the updated model parameter table; The step of determining the model parameters of the power battery based on the operating data and the model parameter table of the power battery includes: Initialize the first and second databases; Data sampling with a preset distribution density is performed on each evaluation interval in the model parameter table. The first database is updated based on the sampled data, and the model parameters in the first database are identified. A data pair is constructed based on the model parameters and the operating parameters, the data pair is updated to the second database, and the model parameters of the power battery are determined based on the second database. The construction of the mapping function between the running parameters and the model parameters in the running data includes: Identify the first quantity of data in the second database and the second quantity of model parameters in the model parameter table; The error function of the fully connected neural network is determined based on the first quantity and the second quantity, and the fully connected neural network is determined based on the nonlinear activation function and the error function; The nonlinear mapping relationship between the running parameters and the model parameters is determined based on the fully connected neural network, and a mapping function between the running parameters and the model parameters is constructed based on the nonlinear mapping relationship.

9. A server, characterized in that, include: The first acquisition module is used to acquire the operating data of the power battery uploaded by the vehicle and the estimation error table of model parameters; A construction module is used to determine the model parameters of the power battery based on the operating data and the model parameter table of the power battery, construct a mapping function between the operating parameters in the operating data and the model parameters, and obtain the confidence interval of the mapping function; The first update module is used to determine the evaluation interval to be updated in the model parameter table according to the confidence interval, update the model parameters of the evaluation interval to be updated according to the mapping function and the estimation error table, and send the updated model parameter table to the vehicle, so that the vehicle updates the model parameters of the power battery according to the updated model parameter table. The step of determining the model parameters of the power battery based on the operating data and the model parameter table of the power battery includes: Initialize the first and second databases; Data sampling with a preset distribution density is performed on each evaluation interval in the model parameter table. The first database is updated based on the sampled data, and the model parameters in the first database are identified. A data pair is constructed based on the model parameters and the operating parameters, the data pair is updated to the second database, and the model parameters of the power battery are determined based on the second database. The construction of the mapping function between the running parameters and the model parameters in the running data includes: Identify the first quantity of data in the second database and the second quantity of model parameters in the model parameter table; The error function of the fully connected neural network is determined based on the first quantity and the second quantity, and the fully connected neural network is determined based on the nonlinear activation function and the error function; The nonlinear mapping relationship between the running parameters and the model parameters is determined based on the fully connected neural network, and a mapping function between the running parameters and the model parameters is constructed based on the nonlinear mapping relationship.

10. A vehicle, characterized in that, include: The second acquisition module is used to acquire the operating data of the power battery and the estimation error table of the model parameters; An upload module is used to upload the operating data and the estimation error table to the server. The server determines the model parameters of the power battery based on the operating data and the model parameter table of the power battery, constructs a mapping function between the operating parameters in the operating data and the model parameters, obtains the confidence interval of the mapping function, determines the evaluation interval to be updated in the model parameter table based on the confidence interval, updates the model parameters in the evaluation interval to be updated based on the mapping function and the estimation error table, and sends the updated model parameter table to the vehicle. The second update module is used to update the model parameters of the power battery according to the updated model parameter table; The step of determining the model parameters of the power battery based on the operating data and the model parameter table of the power battery includes: Initialize the first and second databases; Data sampling with a preset distribution density is performed on each evaluation interval in the model parameter table. The first database is updated based on the sampled data, and the model parameters in the first database are identified. A data pair is constructed based on the model parameters and the operating parameters, the data pair is updated to the second database, and the model parameters of the power battery are determined based on the second database. The construction of the mapping function between the running parameters and the model parameters in the running data includes: Identify the first quantity of data in the second database and the second quantity of model parameters in the model parameter table; The error function of the fully connected neural network is determined based on the first quantity and the second quantity, and the fully connected neural network is determined based on the nonlinear activation function and the error function; The nonlinear mapping relationship between the running parameters and the model parameters is determined based on the fully connected neural network, and a mapping function between the running parameters and the model parameters is constructed based on the nonlinear mapping relationship.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the parameter update method for the power battery model as described in any one of claims 1-8.

Images