A battery status evaluation method and related equipment

Abstract

An embodiment of the present application provides a battery status assessment method and related equipment, the method comprising: obtaining a first battery health SOH of the first battery based on first status data of the first battery in a target battery cycle; if a distance that a first data point deviates from a first battery trend model is less than a first preset threshold, and a distance that the first data point deviates from a second battery trend model is less than a second preset threshold, then determining the first SOH as the SOH of the first battery in the target battery cycle, wherein the first data point includes the first status data and the first SOH, the first battery trend model is used to characterize a changing trend of the SOH of the first battery, and the second battery trend model is used to characterize a changing trend of the SOH of a target type of battery. By adopting the embodiment of the present application, a more accurate SOH can be obtained.

Description

Technical Field

[0001] The present application relates to the field of battery technology, and in particular to a battery status assessment method and related equipment. Background Art

[0002] As power batteries age, their ability to store and deliver energy decreases over time. The state of health (SOH) of a power battery is a quantitative indicator used to assess its degree of aging. Accurate SOH assessment can ensure the safe and reliable operation of power batteries, optimize the use of power battery systems, and provide a basis for energy and safety management in products such as automobiles. The most intuitive manifestation of power battery aging is a decrease in the energy it can release and its power level, which is reflected internally as capacity decay and increased internal resistance. Due to the complex electrochemical kinetics and multi-physics coupling characteristics of power batteries, parameters such as capacity and internal resistance exhibit nonlinear and highly time-varying characteristics, and observable parameters are limited. Furthermore, the actual application scenarios of power batteries involve more complex and variable operating conditions, and existing network data for power batteries also suffers from uncontrollable usage scenarios (e.g., not a complete battery cycle) and missing data. All of these factors make highly accurate and robust SOH estimation of power batteries extremely challenging. Currently, common SOH assessment methods include electrochemical models and equivalent circuit models (ECMs).

[0003] The electrochemical model method is divided into the method based on the aging mechanism and the method based on the electrochemical impedance spectroscopy (EIS). Among them, the method based on the aging mechanism mainly simulates the changes in lithium ion concentration, solid electrolyte interphase (SEI) film thickness and electrode conductivity during battery aging, establishes an SOH prediction model, and then predicts the battery's SOH through the SOH prediction model. The EIS-based method evaluates SOH by measuring the battery's alternating current (AC) impedance.

[0004] The ECM method simulates the battery's polarization and self-discharge reactions based on its electrical characteristics using resistors and capacitors, ensuring the model closely resembles the battery's actual conditions. The ECM method includes the following steps: 1. Selecting a suitable ECM; 2. Identifying ECM parameters, typically using methods such as Hybrid Pulse Power Characteristic (HPPC) testing and curve fitting comparison; 3. Determining the SOH using state variables such as internal resistance and maximum state of charge (SOC) through methods such as the Extended Kalman Filter (EKF).

[0005] However, the SOH error obtained by existing SOH evaluation methods is usually large. How to improve the accuracy of SOH is a technical problem that is being studied by those skilled in the art. Summary of the Invention

[0006] The embodiments of the present application disclose a battery state assessment method and related equipment, which can improve the accuracy of SOH.

[0007] In a first aspect, an embodiment of the present application provides a battery status assessment method, the method comprising:

[0008] Obtaining a first battery state of health (SOH) of the first battery according to first state data of the first battery in a target battery cycle; the battery cycle is a charging cycle or a discharging cycle, wherein the charging cycle includes a process from the start of charging to the end of charging of the battery, and the discharging cycle includes a process from the start of discharging to the end of discharging of the battery;

[0009] If the distance that the first data point deviates from the first battery trend model is less than a first preset threshold, and the distance that the first data point deviates from the second battery trend model is less than a second preset threshold, the first SOH is determined as the SOH of the first battery in the target battery cycle, wherein the first data point includes the first status data and the first SOH, the first battery trend model is used to characterize the changing trend of the SOH of the first battery, and the second battery trend model is used to characterize the changing trend of the SOH of batteries of a target class, and the batteries of the target class include the first battery and at least one second battery.

[0010] In the above method, outlier detection is performed on the data points through the first battery trend model and the second battery trend model to determine whether the SOH in the data point is abnormal. Only when the SOH is detected to be normal is the SOH in the data point used as the SOH of the first battery in the target battery cycle, thereby improving the accuracy of the SOH.

[0011] In combination with the first aspect, in an optional solution, the method further includes:

[0012] Adding the first data point to a data set, wherein the data set includes a plurality of data points, each data point including state data of a battery in a battery cycle and the SOH of the battery in the battery cycle;

[0013] The first battery trend model is trained using data points about the first battery in the data set.

[0014] It can be understood that the data points detected to be normal are added to the data set, while the data points detected to be abnormal are eliminated, so that the data points that are finally left in the data set are those with relatively high SOH accuracy. Therefore, when the above-mentioned first battery trend model is subsequently updated based on this data set, a first battery trend model with higher accuracy can be obtained.

[0015] In combination with the first aspect, in another optional solution, the multiple data points in the data set are all data points of batteries of the target class, and the method further includes: training the second battery trend model using the data set.

[0016] It can be understood that the data points detected to be normal are added to the data set, while the data points detected to be abnormal are eliminated, so that the data points that are finally left in the data set are those with relatively high SOH accuracy. Therefore, when the above-mentioned second battery trend model is subsequently updated based on this data set, a second battery trend model with higher accuracy can be obtained.

[0017] In combination with the first aspect, in another optional scheme, after adding the first data point to the data set, it also includes: training the SOH evaluation model through the SOH in each data point in the data set and the status data corresponding to one or more time segments in a battery cycle of each data point, wherein the SOH evaluation model is used to predict the SOH of the battery.

[0018] It can be understood that normal data points are added to the dataset, while abnormal data points are removed, which means that the data points that remain in the dataset are those with relatively high SOH accuracy. Therefore, based on this dataset, a SOH evaluation model with better prediction effect can be trained. In the future, when SOH cannot be determined directly by inference, SOH can be predicted through the SOH evaluation model.

[0019] In combination with the first aspect, in yet another optional solution, after training the SOH evaluation model using the SOH in each data point in the data set and the state data corresponding to one or more time segments in a battery cycle of each data point, the method further includes:

[0020] Acquire second status data of the first battery in a first time period;

[0021] If the first time period does not belong to the first preset period, inputting the second state data into the SOH evaluation model to obtain a second SOH;

[0022] If the distance that the second data point deviates from the second battery trend model is less than the third preset threshold, and the distance that the second data point deviates from the first battery trend model is less than the fourth preset threshold, the second SOH is determined as the SOH of the first battery in the first time period, and the second data point includes the second state data and the second SOH.

[0023] In this approach, the battery's SOH is predicted using the SOH assessment model for status data outside the first preset cycle, resolving the prior art issue of being unable to determine the SOH based on status data outside the first preset cycle. Furthermore, the SOH predicted by the SOH assessment model is further tested for anomalies using the first and second battery trend models, helping to eliminate abnormal SOH and improving SOH accuracy.

[0024] In combination with the first aspect, in yet another optional solution, the method further includes:

[0025] If the first time period belongs to a first preset cycle, obtaining a third SOH of the first battery according to the second state data of the first battery in the first time period;

[0026] If the distance that the third data point deviates from the first battery trend model is less than the first preset threshold, and the distance that the third data point deviates from the second battery trend model is less than the second preset threshold, the third SOH is determined as the SOH of the first battery in the first time period, wherein the third data point includes the second state data and the third SOH.

[0027] In this way, the SOH of the battery is directly calculated for the status data belonging to the first preset cycle, and the data points are detected as outliers through the first battery trend model and the second battery trend model to determine whether the SOH in the data point is abnormal. Only when the SOH is detected to be normal will the SOH in the data point be used as the SOH of the first battery in the target battery cycle, thereby improving the accuracy of the SOH.

[0028] In combination with the first aspect, in yet another optional solution, the method further includes:

[0029] If the distance of the third data point deviating from the first battery trend model is greater than a first preset threshold, or the distance of the third data point deviating from the second battery trend model is greater than a second preset threshold, inputting the second state data into the SOH evaluation model to obtain a fourth SOH;

[0030] If the distance that the fourth data point deviates from the second battery trend model is less than the third preset threshold, and the distance that the second data point deviates from the first battery trend model is less than the fourth preset threshold, the fourth SOH is determined as the SOH of the first battery in the first time period, and the fourth data point includes the second state data and the fourth SOH.

[0031] In this way, the SOH of the battery is directly calculated for the status data belonging to the first preset cycle, and the data points are detected for outliers through the first battery trend model and the second battery trend model to determine whether the SOH in the data points is abnormal. If an abnormal SOH is detected, the SOH evaluation model is further used to predict the SOH, and ultimately an SOH with relatively high accuracy can always be obtained.

[0032] In combination with the first aspect, in another optional solution, the method further includes: merging the SOH of the first battery in the first time period with the historical SOH of the first battery to obtain a change trajectory of the SOH of the first battery.

[0033] Through this method, the variation of the SOH of the first battery with the target parameter (such as mileage) can be obtained, which is convenient for users to judge the entire life cycle of the battery and make improvements based on this.

[0034] In combination with the first aspect, in another optional solution, the change trajectory of the SOH of the first battery is used to characterize the change relationship of the SOH with the target parameter, and the target parameter includes at least one of mileage, time, and number of cycles.

[0035] In combination with the first aspect, in yet another optional solution, updating the first battery trend model using data points about the first battery in the data set includes:

[0036] determining a first state of charge (SOC) difference and a second state of charge (SOC) difference, wherein the first SOC difference is a difference between the charge of the first battery at a start point and an end point of a battery cycle corresponding to a fifth data point, and the second SOC difference is a difference between the charge of the first battery at a start point and an end point of the battery cycle corresponding to a sixth data point, wherein both the fifth data point and the sixth data point are data points related to the first battery in the data set;

[0037] The first battery trend model is updated according to the fifth data point, the first confidence level, the sixth data point, and the second confidence level, wherein the first confidence level is used to constrain the influence of the fifth data point on the first battery trend model, and the second confidence level is used to constrain the influence of the sixth data point on the first battery trend model.

[0038] By adopting this method, the influence of more data points on the first battery trend model (or the second battery trend model) can be taken into account, thereby improving the robustness of the first battery trend model. At the same time, when considering the influence of more data points, some atypical data points will inevitably be introduced. If such atypical data points are given the same influence as typical data points, then the accuracy of the first battery trend model trained based on these data points may not be guaranteed. Therefore, the embodiment of the present application proposes to set different confidence levels for different data points. By setting the confidence level, the influence of different data points can be differentiated, which can improve the accuracy of the first battery trend model finally trained.

[0039] In combination with the first aspect, in another optional solution, if the first charge SOC difference is greater than the second charge SOC difference, the first confidence level is greater than the second confidence level.

[0040] In a second aspect, an embodiment of the present application provides a battery status assessment method, the method comprising:

[0041] Determining a first state of charge (SOC) difference and a second state of charge (SOC) difference, wherein the first SOC difference is a difference between the charge of the first battery at a start point and an end point of a battery cycle corresponding to a fifth data point, and the second SOC difference is a difference between the charge of the first battery at a start point and an end point of the battery cycle corresponding to a sixth data point, wherein both the fifth data point and the sixth data point are data points related to the first battery in the data set; the battery cycle is a charging cycle or a discharging cycle, wherein the charging cycle includes a process from the start of charging to the end of charging, and the discharging cycle includes a process from the start of discharging to the end of discharging;

[0042] The first battery trend model is updated according to the fifth data point, the first confidence level, the sixth data point, and the second confidence level, wherein the first confidence level is used to constrain the influence of the fifth data point on the first battery trend model, and the second confidence level is used to constrain the influence of the sixth data point on the first battery trend model.

[0043] By adopting this method, the influence of more data points on the first battery trend model (or the second battery trend model) can be taken into account, thereby improving the robustness of the first battery trend model. At the same time, when considering the influence of more data points, some atypical data points will inevitably be introduced. If such atypical data points are given the same influence as typical data points, then the accuracy of the first battery trend model trained based on these data points may not be guaranteed. Therefore, the embodiment of the present application proposes to set different confidence levels for different data points. By setting the confidence level, the influence of different data points can be differentiated, which can improve the accuracy of the first battery trend model finally trained.

[0044] In a third aspect, an embodiment of the present application provides a battery status assessment device, the device comprising:

[0045] a first acquiring unit, configured to acquire a first state of health (SOH) of the first battery according to first state data of the first battery in a target battery cycle; the battery cycle is a charging cycle or a discharging cycle, wherein the charging cycle includes a process from the start of charging to the end of charging of the battery, and the discharging cycle includes a process from the start of discharging to the end of discharging of the battery;

[0046] A first determination unit is configured to determine the first SOH as the SOH of the first battery in the target battery cycle when a distance that the first data point deviates from the first battery trend model is less than a first preset threshold value and a distance that the first data point deviates from the second battery trend model is less than a second preset threshold value, wherein the first data point includes the first status data and the first SOH, the first battery trend model is used to characterize a changing trend of the SOH of the first battery, and the second battery trend model is used to characterize a changing trend of the SOH of batteries of a target class, and the batteries of the target class include the first battery and at least one second battery.

[0047] In the above method, outlier detection is performed on the data points through the first battery trend model and the second battery trend model to determine whether the SOH in the data point is abnormal. Only when the SOH is detected to be normal is the SOH in the data point used as the SOH of the first battery in the target battery cycle, thereby improving the accuracy of the SOH.

[0048] In conjunction with the second aspect, in an optional solution, the device further includes:

[0049] an adding unit, configured to add the first data point to a data set, wherein the data set includes a plurality of data points, each data point including state data of a battery in a battery cycle and the SOH of the battery in the battery cycle;

[0050] A first training unit is configured to train the first battery trend model using data points about the first battery in the data set.

[0051] It can be understood that the data points detected to be normal are added to the data set, while the data points detected to be abnormal are eliminated, so that the data points that are finally left in the data set are those with relatively high SOH accuracy. Therefore, when the above-mentioned first battery trend model is subsequently updated based on this data set, a first battery trend model with higher accuracy can be obtained.

[0052] In conjunction with the third aspect, in yet another optional solution, the multiple data points in the data set are all data points of batteries of the target class, and the apparatus further includes:

[0053] The second training unit is configured to train the second battery trend model using the data set.

[0054] It can be understood that the data points detected to be normal are added to the data set, while the data points detected to be abnormal are eliminated, so that the data points that are finally left in the data set are those with relatively high SOH accuracy. Therefore, when the above-mentioned second battery trend model is subsequently updated based on this data set, a second battery trend model with higher accuracy can be obtained.

[0055] In conjunction with the third aspect, in yet another optional solution, the method further includes:

[0056] A third training unit is used to train an SOH evaluation model after adding the first data point to the data set through the SOH in each data point in the data set and the status data corresponding to one or more time segments in a battery cycle of each data point, wherein the SOH evaluation model is used to predict the SOH of the battery.

[0057] It can be understood that normal data points are added to the dataset, while abnormal data points are removed, which means that the data points that remain in the dataset are those with relatively high SOH accuracy. Therefore, based on this dataset, a SOH evaluation model with better prediction effect can be trained. In the future, when SOH cannot be determined directly by inference, SOH can be predicted through the SOH evaluation model.

[0058] In conjunction with the third aspect, in yet another optional solution, the method further includes:

[0059] a second acquiring unit, configured to acquire second state data of the first battery in a first time period after training the SOH evaluation model using the SOH in each data point in the data set and state data corresponding to one or more time segments in a battery cycle of each data point;

[0060] a first input unit, configured to input the second state data into the SOH evaluation model to obtain a second SOH when the first time period does not belong to a first preset period;

[0061] A second determination unit is configured to determine the second SOH as the SOH of the first battery in the first time period when a distance that the second data point deviates from the second battery trend model is less than a third preset threshold and a distance that the second data point deviates from the first battery trend model is less than a fourth preset threshold, wherein the second data point includes the second state data and the second SOH.

[0062] In this approach, the battery's SOH is predicted using the SOH assessment model for status data outside the first preset cycle, resolving the prior art issue of being unable to determine the SOH based on status data outside the first preset cycle. Furthermore, the SOH predicted by the SOH assessment model is further tested for anomalies using the first and second battery trend models, helping to eliminate abnormal SOH and improving SOH accuracy.

[0063] In conjunction with the third aspect, in yet another optional solution, the method further includes:

[0064] a third acquiring unit, configured to acquire a third SOH of the first battery according to the second state data of the first battery in the first time period when the first time period belongs to a first preset cycle;

[0065] a third determination unit, configured to determine the third SOH as the SOH of the first battery in the first time period when a distance that the third data point deviates from the first battery trend model is less than a first preset threshold and a distance that the third data point deviates from the second battery trend model is less than a second preset threshold, wherein the third data point includes the second state data and the third SOH.

[0066] In this way, the SOH of the battery is directly calculated for the status data belonging to the first preset cycle, and the data points are detected as outliers through the first battery trend model and the second battery trend model to determine whether the SOH in the data point is abnormal. Only when the SOH is detected to be normal will the SOH in the data point be used as the SOH of the first battery in the target battery cycle, thereby improving the accuracy of the SOH.

[0067] In conjunction with the third aspect, in yet another optional solution, the method further includes:

[0068] a second input unit, configured to input the second state data into the SOH evaluation model to obtain a fourth SOH when a distance that the third data point deviates from the first battery trend model is greater than a first preset threshold, or a distance that the third data point deviates from the second battery trend model is greater than a second preset threshold;

[0069] a fourth determination unit, configured to determine the fourth SOH as the SOH of the first battery in the first time period when the distance a fourth data point deviates from the second battery trend model is less than a third preset threshold and the distance a second data point deviates from the first battery trend model is less than a fourth preset threshold, wherein the fourth data point includes the second state data and the fourth SOH.

[0070] In this way, the SOH of the battery is directly calculated for the status data belonging to the first preset cycle, and the data points are detected for outliers through the first battery trend model and the second battery trend model to determine whether the SOH in the data points is abnormal. If an abnormal SOH is detected, the SOH evaluation model is further used to predict the SOH, and ultimately an SOH with relatively high accuracy can always be obtained.

[0071] In conjunction with the third aspect, in yet another optional solution, the method further includes:

[0072] A processing unit is configured to combine the SOH of the first battery in the first time period with the historical SOH of the first battery to obtain a change trajectory of the SOH of the first battery.

[0073] Through this method, the variation of the SOH of the first battery with the target parameter (such as mileage) can be obtained, which is convenient for users to judge the entire life cycle of the battery and make improvements based on this.

[0074] In combination with the third aspect, in another optional solution, the change trajectory of the SOH of the first battery is used to characterize the change relationship of the SOH with the target parameter, and the target parameter includes at least one of mileage, time, and number of cycles.

[0075] In conjunction with the third aspect, in yet another optional solution, in updating the first battery trend model using data points about the first battery in the data set, the first training unit is specifically configured to:

[0076] determining a first state of charge (SOC) difference and a second state of charge (SOC) difference, wherein the first SOC difference is a difference between the charge of the first battery at a start point and an end point of a battery cycle corresponding to a fifth data point, and the second SOC difference is a difference between the charge of the first battery at a start point and an end point of the battery cycle corresponding to a sixth data point, wherein both the fifth data point and the sixth data point are data points related to the first battery in the data set;

[0077] The first battery trend model is updated according to the fifth data point, the first confidence level, the sixth data point, and the second confidence level, wherein the first confidence level is used to constrain the influence of the fifth data point on the first battery trend model, and the second confidence level is used to constrain the influence of the sixth data point on the first battery trend model.

[0078] By adopting this method, the influence of more data points on the first battery trend model (or the second battery trend model) can be taken into account, thereby improving the robustness of the first battery trend model. At the same time, when considering the influence of more data points, some atypical data points will inevitably be introduced. If such atypical data points are given the same influence as typical data points, then the accuracy of the first battery trend model trained based on these data points may not be guaranteed. Therefore, the embodiment of the present application proposes to set different confidence levels for different data points. By setting the confidence level, the influence of different data points can be differentiated, which can improve the accuracy of the first battery trend model finally trained.

[0079] In combination with the third aspect, in another optional solution, if the first charge SOC difference is greater than the second charge SOC difference, the first confidence level is greater than the second confidence level.

[0080] In a fourth aspect, an embodiment of the present application provides a battery status assessment device, the device comprising:

[0081] a fifth determining unit, configured to determine a first charge state occupation (SOC) difference and a second charge state occupation (SOC) difference, wherein the first SOC difference is a charge difference between a start point and an end point of a battery cycle corresponding to a fifth data point of the first battery, and the second SOC difference is a charge difference between a start point and an end point of the battery cycle corresponding to a sixth data point of the first battery, wherein both the fifth data point and the sixth data point are data points related to the first battery in the data set; the battery cycle is a charging cycle or a discharging cycle, wherein the charging cycle includes a process from the start of charging to the end of charging of the battery, and the discharging cycle includes a process from the start of discharging to the end of discharging of the battery;

[0082] a fitting unit, configured to fit a first battery trend model according to a first confidence level using the fifth data point, and to fit the first battery trend model according to a second confidence level using the sixth data point, wherein the first battery trend model is used to characterize a changing trend of the SOH of the first battery.

[0083] In a fifth aspect, an embodiment of the present application provides a battery status assessment device, comprising a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to call the computer program to implement the method described in the first aspect or any possible implementation of the first aspect.

[0084] In a sixth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When the computer-readable storage medium runs on a processor, it implements the method described in the first aspect or any possible implementation of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] The following is an introduction to the drawings used in the embodiments of this application.

[0086] Figure 1 This is a schematic diagram of the architecture of a model training provided in an embodiment of the present application;

[0087] Figure 2 This is a structural diagram of a model training device provided in an embodiment of the present application;

[0088] Figure 3 This is a flow chart of a battery status evaluation method provided in an embodiment of the present application;

[0089] Figure 4 This is a flow chart of a battery status assessment method provided in an embodiment of the present application;

[0090] Figure 5 This is a flow chart of a battery status evaluation method provided in an embodiment of the present application;

[0091] Figure 6 This is a schematic structural diagram of a battery status assessment device provided in an embodiment of the present application;

[0092] Figure 7 This is a schematic structural diagram of a battery status assessment device provided in an embodiment of the present application;

[0093] Figure 8 This is a schematic diagram of the structure of a battery status assessment device provided in an embodiment of the present application;

[0094] Figure 9 This is a schematic diagram of the structure of a battery status assessment device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0095] The embodiments of the present application are described below in conjunction with the drawings in the embodiments of the present application.

[0096] See Figure 1 , Figure 1 This is a schematic diagram of the architecture of a model training provided by an embodiment of the present application. The architecture 10 includes a cloud platform 100, a model training device 101, and one or more model-using devices 102. The cloud platform 100 and the model training device 101, as well as the model training device 101 and the model-using device 102, can communicate via wired or wireless means. Therefore, the model training device 101 can obtain the battery data required for training from the cloud platform 100. The model training device 101 can perform corresponding operations based on the trained model related to the battery state of health (SOH), such as predicting SOH, detecting whether SOH is abnormal, etc. The model training device 101 can send the trained model related to SOH to the model-using device 102 for use.

[0097] Optionally, the model-using device 102 can feed back the results predicted by the model to the above-mentioned model training device 101, so that the model training device 101 can further train the model based on the prediction results of the model-using device 102; the retrained model can be sent to the model-using device 102 to update the original model.

[0098] The cloud platform 100 can be a server cluster consisting of one server or multiple servers. The cloud platform 100 can establish a communication connection with multiple battery-using devices (such as vehicles), so it can receive battery data sent by each battery-using device. Optionally, the cloud platform 100 can be a platform produced (or developed, or sold) by an automobile manufacturer for monitoring the battery status of vehicles produced (or developed, or sold) by the manufacturer, or a platform used by a vehicle operating platform to monitor the battery status of a vehicle, or a third-party platform trusted by the car (such as registration, or information authorization, etc.). Therefore, the cloud platform 100 can obtain a lot of battery data.

[0099] The model training device 101 can be a device with strong computing power, for example, a server, or a server cluster composed of multiple servers. The model training device 101 can train a first battery trend model, a second battery trend model, and an SOH evaluation model, wherein the first battery trend model corresponds to a battery and is used to characterize the general trend of the SOH of the battery, and the second battery trend model corresponds to a type of battery and is used to characterize the general trend of the SOH of the type of battery. Therefore, the first battery trend model and the second battery trend model can also be regarded as two fitting curves. In addition, the SOH evaluation model is used to predict the SOH of the battery based on the input battery status data.

[0100] Figure 2 This is a schematic diagram of the architecture of a model training device provided in an embodiment of the present application, which includes a preprocessing and feature engineering module 201, a battery capacity calculation module 202, a battery trend model training module 203, and a SOH assessment model training and application module 204, wherein:

[0101] The preprocessing and feature engineering module 201 is used to preprocess, divide the cycle, and extract cycle features from the raw data of the battery obtained from the cloud platform 100. Preprocessing includes performing null value processing, boundary constraints, data type conversion, etc. on the data. Cycle division includes dividing the data into cycles based on the changes in the amount of electricity described in the data or the charge and discharge signs. This process may involve data segmentation, abnormal data processing, etc. Cycle feature extraction is a specific type of data from the data. The specific type is a data type that appears in both a complete battery cycle and a partial battery cycle, including statistical value features, time series features, etc. Optionally, the raw data may include the battery's key performance indicator (KPI) data, the identifier of the device (such as the vehicle) where the battery is located (such as the vehicle identification number (VIN)), the battery type and model, etc. For ease of description, the data obtained after processing by the preprocessing and feature engineering module 201 can be referred to as state data.

[0102] The battery capacity calculation module 202 is used to analyze the status data processed by the preprocessing and feature engineering module 201, for example, analyzing the range of each indicator therein (for example, whether the starting current described in the data is less than a given threshold), analyzing the cycle type corresponding to the status data (for example, analyzing according to the range of the starting remaining power described in the data), such as whether it is a complete battery cycle or a partial battery cycle. There are two different processing paths for the status data of the complete battery cycle and the partial battery cycle. For example, if it is the status data of the complete battery cycle, then the battery capacity calculation model is used to calculate the SOH corresponding to the status data using the corresponding algorithm (such as the ampere-hour integration method, OCV-SOC calibration, etc.). If it is the status data of a partial battery cycle, the status data is output to the SOH evaluation model training and application module 204, and the SOH corresponding to the status data is predicted by the SOH evaluation model in the SOH evaluation model training and application module 204.

[0103] The battery trend model training module 203 is used to perform anomaly detection on the SOH calculated in the battery capacity calculation module 202 through the first battery trend model and the second battery trend model. If the detection is abnormal (i.e., it deviates too far from the trend model), it is eliminated. If the detection is normal, the SOH and its corresponding status data are retained, and then the data points are output to the SOH evaluation model training and application module 204. The SOH evaluation model training and application module 204 is used to train each input data point to obtain an SOH evaluation model, wherein the data points include the retained SOH and its corresponding status data. The first battery trend model can be obtained based on linear, polynomial, time series analysis, moving average, deep learning and other algorithms, and the second battery trend model can be obtained based on regression algorithms, trend algorithms, etc.; statistics, clustering, moving average, time series analysis, tree model algorithms and other algorithms may be applied in anomaly detection.

[0104] The SOH evaluation model training and application module 204 may also predict the SOH of the battery using the SOH evaluation model, or receive the SOH after abnormality detection by the first battery trend model and the second battery trend model.

[0105] After obtaining the SOH of a battery, the SOH assessment model training and application module 204 may also combine the SOH with the historically recorded SOH of the battery to obtain a trend relationship of the SOH of the battery.

[0106] The model uses device 102 as a device that needs to detect the SOH of the battery, such as a handheld device (e.g., a mobile phone, a tablet computer, a PDA, etc.), an in-vehicle device (e.g., a car, a bicycle, an electric car, an airplane, a ship, etc.), a wearable device (e.g., a smart watch (such as iWatch, etc.), a smart bracelet, a pedometer, etc.), a smart home device (e.g., a refrigerator, a television, an air conditioner, an electric meter, etc.), an intelligent robot, a workshop equipment, etc.

[0107] See Figure 3 , Figure 3 This is a battery status evaluation method provided by an embodiment of the present application. The method can be based on Figure 1 and Figure 2 The method is implemented based on the architecture shown in the figure, and can also be implemented based on other architectures. The method includes but is not limited to the following steps:

[0108] Step S301: The model training device obtains data of the first battery.

[0109] Specifically, the model training device can receive data about the first battery sent by other devices. For example, the other device can be a cloud platform (described above) or other storage devices, such as a USB flash drive. Of course, the other device can also be a product device that uses the first battery, such as an in-vehicle device, a handheld device, a smart home device, a wearable device, a workshop device, etc. The model training device and the other device can establish a communication connection through a network or a data cable.

[0110] In an embodiment of the present application, the model training device can obtain data of one or more batteries. When it is used to obtain data of one battery, the first battery is the one battery. When it is used to obtain data of multiple batteries, the first battery is one of the multiple batteries. In this case, the subsequent description of the first battery is equivalent to an example. The processing method for the data of any one of the multiple batteries is the same as the processing method for the data of the first battery mentioned.

[0111] The data may be the raw data or status data of the first battery, wherein the status data is effective data for training obtained after preprocessing, period division, and period feature extraction (optional) of the raw data. The raw data may include the battery's key performance indicator (KPI) data, the battery type (such as lithium battery type), and the battery model; of course, it may also include information about the product equipment where the battery is located. For example, when the product equipment is a vehicle, the product equipment information may include the vehicle VIN, vehicle mileage, vehicle time, etc.

[0112] The status data includes the following information:

[0113] 1. Changes in the battery KPI.

[0114] 2. The cycle characteristics of the battery KPI are used to indicate whether the battery KPI is a KPI for a complete battery cycle or a KPI for a partial battery cycle. For example, the cycle characteristics can be determined based on information such as the battery charge level at the start of charging and the battery charge level at the end of charging, as reflected in the battery KPI. Alternatively, the cycle characteristics can be determined based on information such as the battery charge level at the start of discharging and the battery charge level at the end of discharging, as reflected in the battery KPI. Generally speaking, the lower the charge level at the start of charging and the higher the charge level at the end of charging, the more likely it is to be a complete battery cycle. Alternatively, the higher the charge level at the start of discharging and the lower the charge level at the end of discharging, the more likely it is to be a complete battery cycle. Before specific calculations are performed, a reference value for the battery charge level is set to measure the battery charge level in order to determine the battery cycle characteristics.

[0115] In the embodiment of the present application, a battery cycle refers to a discharge cycle or a charge cycle. A discharge cycle refers to the process from the start of discharge to the end of discharge, and a charge cycle refers to the process from the start of charging to the end of charging, wherein:

[0116] Battery cycles can be divided into complete battery cycles and partial battery cycles. A complete battery cycle has the following characteristics:

[0117] A complete battery cycle can be a complete charging cycle. In this case, the power at the start of charging is less than the first power threshold, and the power at the end of charging is greater than the second power threshold. Optionally, it can be further limited that the difference between the power at the end of charging and the power at the start of charging is greater than a preset first power difference. These conditions can ensure that the battery is fully charged. The first power threshold, the second power threshold, and the first power difference here can all be static values ​​set by developers based on experience and actual needs. Of course, they may also be dynamic values ​​generated based on some specific scenarios or specific parameters. Usually, the first power threshold is smaller and the second power threshold is larger. For example, the first power threshold is 15% and the second power threshold is 85% and the first power difference is 75%*H, where H is the maximum capacity of the battery.

[0118] A complete battery cycle can be a complete discharge cycle. In this case, the power at the start of discharge is greater than the third power threshold, and the power at the end of discharge is less than the fourth power threshold. Optionally, the difference between the power at the start of discharge and the power at the end of discharge can be further limited to be greater than a preset second power difference. These conditions can ensure that the battery is discharged more fully. The third power threshold, the fourth power threshold, and the second power difference here can all be static values ​​set by developers based on experience and actual needs. Of course, they may also be dynamic values ​​generated based on some specific scenarios or specific parameters. Usually, the third power threshold is larger and the fourth power threshold is smaller. For example, the third power threshold is 86*%, the fourth power threshold is 16*%, and the second power difference is 74%*H, where H is the maximum capacity of the battery.

[0119] Accordingly, if a battery cycle is not the complete battery cycle mentioned above, the battery cycle is a partial battery cycle.

[0120] 3. The starting state of the battery's equipment at the start of recording the battery KPI. For example, if the equipment is a vehicle, this starting state could include the vehicle's mileage, time, cycle count, and other parameters at the start of recording the KPI. Optionally, the ending state of the battery's equipment at the end of recording the battery KPI could also be included.

[0121] It can be understood that if the model training device obtains the original data of the battery, it will process the original data to obtain the status data; if the model training device obtains the status data of the battery, no corresponding processing is required.

[0122] Step S302: The model training device obtains a first battery health SOH of the first battery according to first state data of the first battery in a target battery cycle.

[0123] Specifically, the target battery cycle belongs to the complete battery cycle mentioned above, and the status data of the first battery in the target battery cycle can be referred to as the first status data. The method for obtaining the first status data can refer to the description of the acquisition of status data in step S301. In an embodiment of the present application, the first battery health SOH of the first battery can be obtained based on the first status data of the first battery in the target battery cycle. The method of obtaining can be specifically calculation, for example, applying ampere integral for calculation. Optionally, the calculation result can also be corrected by the open circuit voltage (OCV)-SOC curve.

[0124] Step S303: If the distance that the first data point deviates from the first battery trend model is less than a first preset threshold, and the distance that the first data point deviates from the second battery trend model is less than a second preset threshold, the model training device determines the first SOH as the SOH of the first battery in the target battery cycle.

[0125] Specifically, the first data point includes the first state data and the first SOH. A data point is constructed by using the first state data and the first SOH to facilitate subsequent calculations.

[0126] The first battery trend model is used to characterize the changing trend of the SOH of the first battery, such as the changing trend of the SOH with the target parameter, and the target parameter includes one or more of the mileage, time, and number of cycles of the product equipment where the battery is located. Of course, the target parameter can also be other parameters that can reflect the state change of the product equipment where the battery is located. Optionally, the first battery trend model can be obtained by training multiple groups of state data and SOH (equivalent to multiple data points) of the first battery obtained historically based on a corresponding algorithm. The corresponding algorithm can include one or more of linear regression (LR), polynomial regression, time series decomposition, moving average (MA), exponential smoothing (ES), autoregressive integrated moving average model (ARIMA), recurrent neural network (RNN), and long short-term memory neural network (LSTM) algorithms, and of course other algorithms may also be included.

[0127] The second battery trend model is used to characterize the changing trend of the SOH of batteries of the target class, such as the changing trend of the SOH with the target parameter. The target class of batteries includes the first battery and at least one second battery. In other words, the first battery and the second battery belong to the same class of batteries. For the convenience of description, they are referred to as target classes. It can be understood that when batteries are classified, they can be classified into large categories or detailed categories. When they are classified into detailed categories, the accuracy of the second battery trend model is higher. Optionally, the second battery trend model can be obtained by training the historically acquired status data and SOH of the first battery and the status data and SOH of the second battery based on a corresponding algorithm. The corresponding algorithm can be the same as or different from the algorithm used to train the first battery trend model.

[0128] For ease of understanding, the first battery trend model and the second battery trend model can be considered as two-dimensional coordinate systems. The horizontal axis of the coordinate system is the quantitative parameters of the battery status data (or some parameters in the battery trend model, such as mileage, time, etc.), and the vertical axis of the coordinate system is SOH. Therefore, a first data point including first status data and the first SOH can be evaluated in the two-dimensional coordinate system in terms of its distance from the first battery trend model and the distance from the second battery trend model.

[0129] In the embodiment of the present application, the first preset threshold and the second preset threshold may be the same or different. The first preset threshold and the second preset threshold may be set by R&D personnel based on experience, or obtained by statistics and / or analysis of corresponding data (such as the mean square error between the SOH values ​​in multiple data points and the multiple fitting values ​​SOH in the battery trend model). It may be a dynamic parameter or a static parameter. The first data point is compared with the first battery trend model and the second battery trend model to complete the abnormality detection. When the distance that the first data point deviates from the first battery trend model is less than the first preset threshold, it is considered that the first data point is closer to the first battery trend model, so the first data point is judged to be normal. When the distance that the first data point deviates from the first battery trend model (such as the difference between the SOH value in the data point and the fitting value SOH in the battery trend model) is greater than the first preset threshold, it is considered that the first data point is far away from the first battery trend model, so the first data point is abnormal. Similarly, when the first data point deviates from the second battery trend model, When the distance is less than the second preset threshold, it is considered that the first data point is closer to the second battery trend model, and the first data point is therefore judged to be normal. When the distance that the first data point deviates from the second battery trend model is greater than the second preset threshold, it is considered that the first data point is farther from the second battery trend model, and therefore the first data point is abnormal. When judging the distance that the first data point deviates from the first battery trend model or the second battery trend model, specific anomaly detection algorithms such as statistical methods N-sigma criterion N-sigma, box-plot box-plot, clustering method k-means clustering k-means, density-based clustering of applications with noise (DBSCAN), balanced iterative reducing and clustering using hierarchies (BIRCH), time series decomposition methods, and tree model methods isolation forest (iForest) can be used.

[0130] In an embodiment of the present application, when the first data point is determined to be abnormal by any one of the first battery trend model and the second battery trend model, the first data point will ultimately be considered abnormal. Only when both battery trend models determine that the first data point is normal, the first data point will ultimately be considered normal. Therefore, the first SOH in the first data point (that is, the first SOH calculated previously) is determined as the SOH of the first battery in the target battery cycle.

[0131] Optionally, steps S304-S305 may be further included.

[0132] Step S304: The model training device adds the first data point to the data set.

[0133] The data set includes multiple data points, each data point includes status data of a battery in a battery cycle and the SOH of the battery in the battery cycle. The battery cycle mentioned here is a complete battery cycle.

[0134] Case 1: the multiple data points in the data set are all data points of batteries of the target class, that is, the data set includes both data points of the first battery and data points of the second battery of the same class as the first battery.

[0135] In case 2, all the data points in the data set are data points of the first battery. In this case, there may be another data set that includes both the data points of the first battery and the data points of a second battery of the same type as the first battery.

[0136] Adding this first data point to the dataset is equivalent to increasing the training data required to train the model.

[0137] According to the same principle, many data points related to the first battery can be added to the data set. Optionally, many data points related to the second battery can also be added to the data set.

[0138] Step S305: The model training device trains the SOH evaluation model through the SOH in each data point in the data set and the state data corresponding to one or more time segments in a battery cycle of each data point.

[0139] When certain conditions are met, such as when the number of new data points in the data set reaches a preset threshold, or when a preset time period is reached, the training of the SOH evaluation model is triggered.

[0140] Among them, the SOH evaluation model is used to predict the SOH of the battery. The training of the SOH evaluation model this time can be either the initial training or retraining of the SOH evaluation model based on the existing SOH evaluation model. Retraining is equivalent to updating the existing SOH evaluation model. The update process includes but is not limited to the following methods: First, without relying on the existing SOH evaluation model, a new SOH evaluation model is trained directly based on the data points in the data set; second, on the basis of the existing SOH evaluation model, training is performed based on the newly added data points in the data set to obtain a new SOH evaluation model.

[0141] It should be noted that each data point in the original data set includes a SOH and a battery cycle, that is, one SOH corresponds to the status data of one battery cycle. Now, the status data corresponding to one or more time segments are divided from the status data of the battery cycle, and then it is determined that the SOH corresponds to the status data of one or more time segments in the battery cycle. This is equivalent to reconstructing a data point similar to the original data point, which can be called a quasi-data point for ease of description. The one or more time segments contain periodic characteristics, and the status data of a partial battery cycle (i.e., a non-complete battery cycle) usually also includes the periodic characteristics, such as statistical value characteristics, time series characteristics, etc.

[0142] This approach allows for the construction of multiple quasi-data points corresponding to time segments of varying lengths from a single data point. For example, a battery cycle in a data point might be the process of starting charging at 5% remaining charge and continuing until the remaining charge reaches 95%. At each moment in the charging process, the battery would have corresponding data, such as remaining charge (or charge), voltage, and other parameters. These are all considered state data, and the state data corresponding to different moments or time periods may vary.

[0143] Case 1: The status data included in the quasi-data point obtained based on the one data point may include the status data corresponding to the time segment when charging starts when the remaining battery power is 10% and continues to charge until the remaining power is 50%, and the SOH included in the quasi-data point obtained based on the one data point is the SOH calculated based on the battery cycle (the process of starting charging when the remaining battery power is 5% and continuing to charge until the remaining power is 95%).

[0144] Case 2: The status data included in the quasi-data point obtained based on the one data point may include the status data corresponding to the time segment when charging starts when the remaining battery power is 60% and continues to charge until the remaining power is 90%, and the SOH included in the quasi-data point obtained based on the one data point is the SOH calculated based on the battery cycle (the process of starting charging when the remaining battery power is 5% and continuing to charge until the remaining power is 95%).

[0145] Case 1 and Case 2 are both possible quasi-data points, and many other quasi-data points can be obtained based on the same principle.

[0146] In the prior art, it is usually impossible to calculate the SOH based on the state data of a partial battery cycle (i.e., not a complete battery cycle, i.e., a time segment). However, using the methods of the embodiments of the present application, the quasi-data points constructed can reflect the corresponding relationship between the state data of a partial battery cycle and the SOH. Therefore, by training based on the derived quasi-data points, an SOH evaluation model can be obtained that can predict the SOH based on the state data of a partial battery cycle.

[0147] Optionally, the data used to train the SOH evaluation model may include, in addition to the aforementioned quasi-data points, data points for constructing the quasi-data points, so that the trained SOH evaluation model can also predict SOH based on the status data of a complete battery cycle.

[0148] Optionally, steps S306-S307 may be further included.

[0149] Step S306: The model training device trains the first battery trend model using the data points about the first battery in the data set.

[0150] When certain conditions are met, such as when the number of new data points in the data set reaches a preset threshold, or when a preset time period is reached, the training of the first battery trend model is triggered.

[0151] This training retrains the existing first battery trend model. This retraining is equivalent to updating the existing first battery trend model. This update process includes but is not limited to the following methods: First, a new first battery trend model is trained directly based on the data points related to the first battery in the dataset, independent of the existing first trend model. Second, a new first battery trend model is trained based on the newly added data points related to the first battery in the dataset, based on the existing battery trend model. If the first battery trend model is subsequently used, the latest first battery trend model will be used.

[0152] In an embodiment of the present application, multiple data points are required to retrain the first battery trend model. Different data points can be assigned different confidences (or weights) so that different data points can have different effects on the first battery trend model. In one optional solution, the confidences of any two data points among the multiple data points are different. In another optional solution, the confidences of at least two data points among the multiple data points are different. For ease of understanding, examples are given below.

[0153] If the plurality of data points includes the fifth data point and the sixth data point, see Figure 4 The following execution process exists:

[0154] Step 1: Determine the difference in charge between the starting point and the end point of the battery cycle of the first battery corresponding to the fifth data point. The obtained difference can be called a first SOC difference. Similarly, determine the difference in charge between the starting point and the end point of the battery cycle of the first battery corresponding to the sixth data point. The obtained difference can be called a second SOC difference.

[0155] Step 2: determine the SOH of the battery according to the fifth data point, and determine the SOH of the battery according to the sixth data point.

[0156] Step 3: Assign a first confidence level to the fifth data point and a second confidence level to the sixth data point. For example, there is a pre-existing correspondence table between confidence levels and SOC differences, such as Table 1. The confidence level corresponding to the first SOC difference is searched from the relationship table as the first confidence level, and the confidence level corresponding to the second SOC difference is searched from the relationship table as the second confidence level.

[0157] Table 1

[0158] SOC difference x (unit: Ah) Confidence x＜＝30 0.2 30＜x＜＝35 0.4 35＜x＜＝40 0.6 X＞40 0.9

[0159] The specific confidence value corresponding to each SOC difference interval can be set according to experience and needs. The specific values ​​in Table 1 are only examples.

[0160] Step 4: Set the mileage corresponding to each data point as the model input and the corresponding SOH as the model output. For example, the starting mileage corresponding to the status data in the fifth data point is used as the model input, and the SOH at the fifth data point is used as the model output; and the starting mileage corresponding to the status data in the sixth data point is used as the model input, and the SOH at the sixth data point is used as the model output. The output here can also be other target parameters, such as time, number of cycles, etc.

[0161] Step 5: Train the first battery trend model using multiple data points, and constrain the influence of each data point on the first battery trend model using the confidence level corresponding to each data point. For example, constrain the influence of the fifth data point on the first battery trend model using the first confidence level, and constrain the influence of the sixth data point on the first battery trend model using the second confidence level. Generally speaking, the greater the confidence level corresponding to a data point, the greater the influence of that data point on the first battery trend model during training.

[0162] Optionally, the training of the second battery trend model can also refer to the training method of the first battery trend model, and the specific details are not repeated here. In addition, this method can be used whether it is the first training or updating an existing model.

[0163] use Figure 4 The method shown can take into account the influence of more data points on the first battery trend model (or the second battery trend model), thereby improving the robustness of the first battery trend model. At the same time, when considering the influence of more data points, some atypical data points will inevitably be introduced. If such atypical data points are given the same influence as typical data points, then the accuracy of the first battery trend model trained based on these data points may not be guaranteed. Therefore, the embodiment of the present application proposes setting different confidence levels for different data points. By setting the confidence levels, the influence of different data points can be differentiated, which can improve the accuracy of the first battery trend model finally trained.

[0164] Step S307: The model training device trains the second battery trend model using the data set.

[0165] When certain conditions are met, such as when the number of new data points in the data set reaches a preset threshold, or when a preset time period is reached, the training of the first battery trend model is triggered.

[0166] It should be noted that, in the embodiment of the present application, the conditions for updating the SOH evaluation model, the first battery trend model, and the second battery trend model may be the same or different, depending on the specific settings.

[0167] This training retrains the second battery trend model based on the existing one. Retraining is equivalent to updating the existing second battery trend model. This update process includes but is not limited to the following methods: First, training a new second battery trend model directly based on the data points in the dataset, independent of the existing second battery trend model; Second, training a new second battery trend model based on the data points in the dataset, based on the existing battery trend model. To use the second battery trend model later, the latest second battery trend model will be used.

[0168] In an alternative approach, the model can also be executed using the device Figure 3 In the method shown in FIG. 1 , steps S302 and S303 are described. In this case, the first state data can be generated by the model-using device itself, and subsequent steps S304 to S307 can also be completed on the model-using device.

[0169] exist Figure 3In the method shown, outlier detection is performed on the data points using the first battery trend model and the second battery trend model to determine whether the SOH in the data points is abnormal. Only when the SOH is detected to be normal is the SOH in the data points used as the SOH of the first battery in the target battery cycle, thereby improving the accuracy of the SOH.

[0170] Furthermore, normal data points are added to the dataset, while abnormal data points are removed. This ensures that only data points with relatively high SOH accuracy remain in the dataset. Therefore, when the first battery trend model and the second battery trend model are subsequently updated based on this dataset, a more accurate first battery trend model and a more accurate second battery trend model can be obtained. In addition, a SOH evaluation model with better prediction results can also be trained based on this dataset. In the future, when SOH cannot be directly determined by inference, the SOH can be predicted using the SOH evaluation model.

[0171] In addition, the first battery trend model, the second battery trend model and the SOH evaluation model in this application can all be trained based on existing network data (including status data and determined SOH). Compared with the data obtained in a limited controlled experimental environment, the various models obtained using the embodiments themselves are more in line with actual working conditions in reflecting the SOH characteristics.

[0172] See Figure 5 , Figure 5 This is a battery status evaluation method provided by an embodiment of the present application. The method can be based on Figure 1 and Figure 2 The architecture shown in the figure can also be implemented based on other architectures. Figure 3 Complements to the methods shown, such as Figure 3 The method shown in the figure mentions the first battery trend model, the second battery trend model, and the SOH evaluation model. Then, after obtaining a battery status data, how to determine the SOH based on the status data? It should be noted that the execution Figure 5 The main body of steps S501-S510 in the method shown can be either the above-mentioned model training device or the above-mentioned model use device. Steps S501-S510 are explained in detail below.

[0173] Step S501: Acquire second status data of the first battery in a first time period.

[0174] Specifically, the state data of the first battery in the first time period may be referred to as second state data. The method for acquiring the second state data may refer to the description of acquiring the state data in step S301.

[0175] Step S502: Determine whether the first time period belongs to a first preset period.

[0176] Specifically, the first preset cycle belongs to a complete battery cycle, and what constitutes a complete battery cycle can be preset.

[0177] For example, if "the process of charging the battery from 10% to 85% remaining power" is set as the critical value of a complete cycle, then as long as the charging process covers "the process of charging the battery from 10% to 85% remaining power", it is counted as a complete battery cycle. For example, "the process of charging the battery from 5% to 90% remaining power" covers "the process of charging the battery from 10% to 85% remaining power", so "the process of charging the battery from 5% to 90% remaining power" belongs to a complete battery cycle.

[0178] For example, if the "process of discharging the battery's remaining power from 85% to 10%" is set as the critical value of a complete cycle, then as long as the discharge process covers the "process of charging the battery's remaining power from 85% to 10%", it is counted as a complete battery cycle. For example, the "process of discharging the battery's remaining power from 90% to 5%" covers the "process of discharging the battery's remaining power from 85% to 10%", so the "process of discharging the battery's remaining power from 90% to 10%" belongs to a complete battery cycle.

[0179] Determining whether the first time period belongs to the first preset cycle specifically includes: judging whether the charging or discharging rules within the first time period meet the charging or discharging rules of a complete battery cycle; if so, the first time period belongs to the first preset cycle; otherwise, the first time period does not belong to the first preset cycle.

[0180] Step S503: If the first time period does not belong to the first preset period, the second state data is input into the SOH evaluation model to obtain a second SOH.

[0181] That is to say, when the second state data of the first time period does not belong to the state data of a complete battery cycle, the SOH is predicted by the SOH evaluation model obtained through previous training. The SOH obtained here can be called the second SOH.

[0182] Step S504: If the distance that the second data point deviates from the second battery trend model is less than the third preset threshold, and the distance that the second data point deviates from the first battery trend model is less than the fourth preset threshold, the second SOH is determined as the SOH of the first battery in the first time period.

[0183] The second data point includes the second state data and the second SOH. That is, after the second SOH is predicted based on the second state data in step S503, a data point is constructed based on the second state data and the second SOH. This data point is called the second data point.

[0184] The second data point is compared not only with the first battery trend model, but also with the second battery trend model. If the distance the second data point deviates from the first battery trend model is less than a fourth preset threshold, and the distance the second data point deviates from the second battery trend model is less than a third preset threshold, then the second SOH is determined to be the SOH of the first battery in the first time period. The fourth preset threshold may be the same as or different from the third preset threshold. The method of comparing the second data point with the first battery trend model and the second battery trend model is the same as the method of comparing the first data point with the first battery trend model and the second battery trend model. In this scheme, when the second data point is determined to be abnormal by any one of the first battery trend model and the second battery trend model, the second data point will ultimately be considered abnormal. Only when both battery trend models determine that the second data point is normal, the second data point will ultimately be considered normal. Therefore, the second SOH in the second data point (that is, the second SOH predicted by the SOH evaluation model) is determined to be the SOH of the first battery in the first time period.

[0185] In another alternative optional scheme, when the distance that the second data point deviates from the second battery trend model is less than the third preset threshold, it indicates that the second data point is normal. When the distance that the second data point deviates from the second battery trend model is not less than the third preset threshold, it indicates that the second data point is abnormal. Among them, the third preset threshold can be set by the R&D personnel based on experience, or obtained by statistics and / or analysis of the corresponding data. It can be a dynamic parameter or a static parameter. The third preset threshold can be the same as or different from the second preset threshold. When it is determined that the second data point is normal, the second SOH in the second data point (that is, the second SOH predicted by the SOH evaluation model) is determined as the SOH of the first battery in the first time period.

[0186] Step S505: If the first time period belongs to a first preset cycle, deriving a third SOH of the first battery according to the second state data of the first battery in the first time period.

[0187] Specifically, the method of deriving the third SOH of the first battery based on the second status data of the first battery in the first time period is the same as the method of "obtaining the first battery health SOH of the first battery based on the first status data of the first battery in the target battery cycle" mentioned in step S302, and will not be elaborated here.

[0188] Step S506: If the distance that the third data point deviates from the first battery trend model is less than the first preset threshold, and the distance that the third data point deviates from the second battery trend model is less than the second preset threshold, the third SOH is determined as the SOH of the first battery in the first time period.

[0189] The third data point includes the second state data and the third SOH.

[0190] Specifically, the implementation principle of step S506 is the same as that of step S303, and will not be further described here.

[0191] Step S507: If the distance that the third data point deviates from the first battery trend model is greater than the first preset threshold, or the distance that the third data point deviates from the second battery trend model is greater than the second preset threshold, the second state data is input into the SOH evaluation model to obtain a fourth SOH.

[0192] Step S508: If the distance that the fourth data point deviates from the second battery trend model is less than the third preset threshold, and the distance that the second data point deviates from the first battery trend model is less than the fourth preset threshold, then the fourth SOH is determined as the SOH of the first battery in the first time period.

[0193] The fourth data point includes the second state data and the fourth SOH.

[0194] Specifically, the implementation principle of step S508 is the same as that of step S504. It should be noted that there are other alternatives to the method of determining the SOH of the first battery in the first time period provided by steps S507-S508. For example, if the distance of the third data point deviating from the first battery trend model is greater than the first preset threshold, or the distance of the third data point deviating from the second battery trend model is greater than the second preset threshold, the third data point is directly substituted into the first battery trend model or the second battery trend model to obtain the SOH of the first battery in the first time period. For example, the SOH corresponding to the second state data in the third data point in the first battery trend model or the second battery trend model is searched, and the corresponding SOH is then used as the SOH of the first battery in the first time period.

[0195] Step S509: merging the SOH of the first battery in the first time period with the historical SOH of the first battery to obtain a change trajectory of the SOH of the first battery.

[0196] Specifically, the SOH of the first battery in the first time period can be considered as the SOH corresponding to a moment (such as the starting moment of the first time period), or the SOH corresponding to a mileage (such as the mileage of the vehicle in which the first battery is located at the starting moment of the first time period), or the SOH corresponding to a number of cycles (such as the mileage of the vehicle in which the first battery is located at the starting moment of the first time period). Therefore, by merging (or concatenating) the SOH of the first battery in the first time period with the historical SOH of the first battery, a trajectory of the change of the SOH of the first battery with the target parameter (such as time, mileage, number of cycles, etc.) can be obtained.

[0197] The battery's SOH has a significant impact on its service life. The SOH status is the prerequisite for predicting the battery's remaining useful life (RUL). Accurate SOH status assessment can provide a basis for regular battery maintenance and safety assessment, thereby ensuring the safe and reliable operation of the power battery system and optimizing the use of the power battery system. In the automotive field, accurate SOH can provide a basis for automobile "three guarantees" regulations, energy management and safety management, etc., to ensure the normal use of car owners and the safe and stable operation of vehicles.

[0198] It should be noted that the above mainly mentions how to obtain the SOH of the battery. Whether it is the first battery trend model, the second battery trend model, or the SOH evaluation model, the main consideration indicator is SOH. In fact, there are many characteristic parameters in the battery that are similar to SOH, such as the actual maximum capacity and internal resistance of the battery. The acquisition method or application method of these parameters can refer to the acquisition method or application method of SOH mentioned above. For example, the first battery trend model and the second battery trend model can be specifically the relationship between the actual maximum capacity and internal resistance of the battery and the change trend of the target parameters (such as mileage, time, number of cycles, etc.).

[0199] exist Figure 5In the described method, different methods are used to determine the SOH for status data from a complete battery cycle and a partial battery cycle. For the complete cycle status data, the SOH is first determined by calculation, while for the partial battery cycle status data, the SOH is predicted using an SOH assessment model. Both methods use the first and second battery trend models obtained previously to perform anomaly detection. Under normal circumstances, the obtained SOH is determined as the SOH of the first battery in the first time period. This approach allows the SOH of both the complete battery cycle and the partial battery cycle to be obtained, improving the efficiency and accuracy of battery SOH acquisition.

[0200] The above describes in detail the method of the embodiment of the present application, and the following provides an apparatus of the embodiment of the present application.

[0201] See Figure 6 , Figure 6 is a structural diagram of a battery status evaluation device provided in an embodiment of the present application. The device 60 may include a first acquisition unit 601 and a first determination unit 602, wherein each unit is described in detail as follows.

[0202] A first acquiring unit 601 is configured to acquire a first state of health (SOH) of the first battery according to first state data of the first battery in a target battery cycle; the battery cycle is a charging cycle or a discharging cycle, wherein the charging cycle includes a process from the start of charging to the end of charging of the battery, and the discharging cycle includes a process from the start of discharging to the end of discharging of the battery;

[0203] The first determination unit 602 is configured to determine the first SOH as the SOH of the first battery in the target battery cycle when a distance that the first data point deviates from the first battery trend model is less than a first preset threshold value and a distance that the first data point deviates from the second battery trend model is less than a second preset threshold value, wherein the first data point includes the first status data and the first SOH, the first battery trend model is used to characterize a changing trend of the SOH of the first battery, and the second battery trend model is used to characterize a changing trend of the SOH of batteries of a target class, and the batteries of the target class include the first battery and at least one second battery.

[0204] In the above method, outlier detection is performed on the data points through the first battery trend model and the second battery trend model to determine whether the SOH in the data point is abnormal. Only when the SOH is detected to be normal is the SOH in the data point used as the SOH of the first battery in the target battery cycle, thereby improving the accuracy of the SOH.

[0205] In an optional solution, the device 60 further includes:

[0206] an adding unit, configured to add the first data point to a data set, wherein the data set includes a plurality of data points, each data point including state data of a battery in a battery cycle and the SOH of the battery in the battery cycle;

[0207] A first training unit is configured to train the first battery trend model using data points about the first battery in the data set.

[0208] It can be understood that the data points detected to be normal are added to the data set, while the data points detected to be abnormal are eliminated, so that the data points that are finally left in the data set are those with relatively high SOH accuracy. Therefore, when the above-mentioned first battery trend model is subsequently updated based on this data set, a first battery trend model with higher accuracy can be obtained.

[0209] In yet another optional solution, the plurality of data points in the data set are all data points of batteries of the target class, and the apparatus further comprises:

[0210] The second training unit is configured to train the second battery trend model using the data set.

[0211] It can be understood that the data points detected to be normal are added to the data set, while the data points detected to be abnormal are eliminated, so that the data points that are finally left in the data set are those with relatively high SOH accuracy. Therefore, when the above-mentioned second battery trend model is subsequently updated based on this data set, a second battery trend model with higher accuracy can be obtained.

[0212] In another optional solution, it also includes:

[0213] A third training unit is used to train an SOH evaluation model after adding the first data point to the data set through the SOH in each data point in the data set and the status data corresponding to one or more time segments in a battery cycle of each data point, wherein the SOH evaluation model is used to predict the SOH of the battery.

[0214] It can be understood that normal data points are added to the dataset, while abnormal data points are removed, which means that the data points that remain in the dataset are those with relatively high SOH accuracy. Therefore, based on this dataset, a SOH evaluation model with better prediction effect can be trained. In the future, when SOH cannot be determined directly by inference, SOH can be predicted through the SOH evaluation model.

[0215] In another optional solution, it also includes:

[0216] a second acquiring unit, configured to acquire second state data of the first battery in a first time period after training the SOH evaluation model using the SOH in each data point in the data set and state data corresponding to one or more time segments in a battery cycle of each data point;

[0217] a first input unit, configured to input the second state data into the SOH evaluation model to obtain a second SOH when the first time period does not belong to a first preset period;

[0218] A second determination unit is configured to determine the second SOH as the SOH of the first battery in the first time period when a distance that the second data point deviates from the second battery trend model is less than a third preset threshold and a distance that the second data point deviates from the first battery trend model is less than a fourth preset threshold, wherein the second data point includes the second state data and the second SOH.

[0219] In this approach, the battery's SOH is predicted using the SOH assessment model for status data outside the first preset cycle, resolving the prior art issue of being unable to determine the SOH based on status data outside the first preset cycle. Furthermore, the SOH predicted by the SOH assessment model is further tested for anomalies using the first and second battery trend models, helping to eliminate abnormal SOH and improving SOH accuracy.

[0220] In another optional solution, it also includes:

[0221] a third acquiring unit, configured to acquire a third SOH of the first battery according to the second state data of the first battery in the first time period when the first time period belongs to a first preset cycle;

[0222] a third determination unit, configured to determine the third SOH as the SOH of the first battery in the first time period when a distance that the third data point deviates from the first battery trend model is less than a first preset threshold and a distance that the third data point deviates from the second battery trend model is less than a second preset threshold, wherein the third data point includes the second state data and the third SOH.

[0223] In this way, the SOH of the battery is directly calculated for the status data belonging to the first preset cycle, and the data points are detected as outliers through the first battery trend model and the second battery trend model to determine whether the SOH in the data point is abnormal. Only when the SOH is detected to be normal will the SOH in the data point be used as the SOH of the first battery in the target battery cycle, thereby improving the accuracy of the SOH.

[0224] In another optional solution, it also includes:

[0225] a second input unit, configured to input the second state data into the SOH evaluation model to obtain a fourth SOH when a distance that the third data point deviates from the first battery trend model is greater than a first preset threshold, or a distance that the third data point deviates from the second battery trend model is greater than a second preset threshold;

[0226] a fourth determination unit, configured to determine the fourth SOH as the SOH of the first battery in the first time period when the distance a fourth data point deviates from the second battery trend model is less than a third preset threshold and the distance a second data point deviates from the first battery trend model is less than a fourth preset threshold, wherein the fourth data point includes the second state data and the fourth SOH.

[0227] In this way, the SOH of the battery is directly calculated for the status data belonging to the first preset cycle, and the data points are detected for outliers through the first battery trend model and the second battery trend model to determine whether the SOH in the data points is abnormal. If an abnormal SOH is detected, the SOH evaluation model is further used to predict the SOH, and ultimately an SOH with relatively high accuracy can always be obtained.

[0228] In another optional solution, it also includes:

[0229] A processing unit is configured to combine the SOH of the first battery in the first time period with the historical SOH of the first battery to obtain a change trajectory of the SOH of the first battery.

[0230] Through this method, the variation of the SOH of the first battery with the target parameter (such as mileage) can be obtained, which is convenient for users to judge the entire life cycle of the battery and make improvements based on this.

[0231] In another optional solution, the variation trajectory of the SOH of the first battery is used to characterize the variation relationship of the SOH with a target parameter, where the target parameter includes at least one of mileage, time, and number of cycles.

[0232] In yet another optional solution, in updating the first battery trend model using data points about the first battery in the data set, the first training unit is specifically configured to:

[0233] determining a first state of charge (SOC) difference and a second state of charge (SOC) difference, wherein the first SOC difference is a difference between the charge of the first battery at a start point and an end point of a battery cycle corresponding to a fifth data point, and the second SOC difference is a difference between the charge of the first battery at a start point and an end point of the battery cycle corresponding to a sixth data point, wherein both the fifth data point and the sixth data point are data points related to the first battery in the data set;

[0234] The first battery trend model is updated according to the fifth data point, the first confidence level, the sixth data point, and the second confidence level, wherein the first confidence level is used to constrain the influence of the fifth data point on the first battery trend model, and the second confidence level is used to constrain the influence of the sixth data point on the first battery trend model.

[0235] By adopting this method, the influence of more data points on the first battery trend model (or the second battery trend model) can be taken into account, thereby improving the robustness of the first battery trend model. At the same time, when considering the influence of more data points, some atypical data points will inevitably be introduced. If such atypical data points are given the same influence as typical data points, then the accuracy of the first battery trend model trained based on these data points may not be guaranteed. Therefore, the embodiment of the present application proposes to set different confidence levels for different data points. By setting the confidence level, the influence of different data points can be differentiated, which can improve the accuracy of the first battery trend model finally trained.

[0236] In another optional solution, if the first charge SOC difference is greater than the second charge SOC difference, the first confidence level is greater than the second confidence level.

[0237] It should be noted that the implementation of each unit can also refer to Figure 3-Figure 4 The corresponding description of the method embodiment shown.

[0238] See Figure 7 , Figure 7 is a structural diagram of a battery status evaluation device 70 provided in an embodiment of the present application. The device 70 may include a fifth determination unit 701 and a fitting unit 702, wherein each unit is described in detail as follows.

[0239] a fifth determining unit 701, configured to determine a first charge state-of-charge (SOC) difference and a second charge state-of-charge (SOC) difference, wherein the first SOC difference is a charge difference between a start point and an end point of a battery cycle corresponding to a fifth data point of the first battery, and the second SOC difference is a charge difference between a start point and an end point of the battery cycle corresponding to a sixth data point of the first battery, wherein both the fifth data point and the sixth data point are data points related to the first battery in the data set; and the battery cycle is a charging cycle or a discharging cycle, wherein the charging cycle includes a process from the start of charging to the end of charging of the battery, and the discharging cycle includes a process from the start of discharging to the end of discharging of the battery.

[0240] A fitting unit 702 is configured to update the first battery trend model based on the fifth data point, the first confidence level, the sixth data point, and the second confidence level, wherein the first confidence level is used to constrain the influence of the fifth data point on the first battery trend model, and the second confidence level is used to constrain the influence of the sixth data point on the first battery trend model.

[0241] In an optional solution, if the first charge SOC difference is greater than the second charge SOC difference, the first confidence level is greater than the second confidence level.

[0242] It should be noted that the implementation of each unit can also refer to Figure 3-Figure 4 The corresponding description of the method embodiment shown.

[0243] See Figure 8 , Figure 8 A battery status evaluation device 80 is provided in an embodiment of the present application. The device 80 includes a processor 801 and a memory 802 , and the processor 801 and the memory 802 are interconnected via a bus.

[0244] The memory 802 includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or compact disc read-only memory (CD-ROM), and is used for related computer programs and data.

[0245] The processor 801 may be one or more central processing units (CPUs). In the case where the processor 801 is a CPU, the CPU may be a single-core CPU or a multi-core CPU.

[0246] The processor 801 in the device 80 is configured to read the computer program code stored in the memory 802 and perform the following operations:

[0247] Obtaining a first battery state of health (SOH) of the first battery according to first state data of the first battery in a target battery cycle; the battery cycle is a charging cycle or a discharging cycle, wherein the charging cycle includes a process from the start of charging to the end of charging of the battery, and the discharging cycle includes a process from the start of discharging to the end of discharging of the battery;

[0248] If the distance that the first data point deviates from the first battery trend model is less than a first preset threshold, and the distance that the first data point deviates from the second battery trend model is less than a second preset threshold, the first SOH is determined as the SOH of the first battery in the target battery cycle, wherein the first data point includes the first status data and the first SOH, the first battery trend model is used to characterize the changing trend of the SOH of the first battery, and the second battery trend model is used to characterize the changing trend of the SOH of batteries of a target class, and the batteries of the target class include the first battery and at least one second battery.

[0249] In the above method, outlier detection is performed on the data points through the first battery trend model and the second battery trend model to determine whether the SOH in the data point is abnormal. Only when the SOH is detected to be normal is the SOH in the data point used as the SOH of the first battery in the target battery cycle, thereby improving the accuracy of the SOH.

[0250] In an optional solution, the processor 801 is further configured to:

[0251] Adding the first data point to a data set, wherein the data set includes a plurality of data points, each data point including state data of a battery in a battery cycle and the SOH of the battery in the battery cycle;

[0252] The first battery trend model is trained using data points about the first battery in the data set.

[0253] It can be understood that the data points detected to be normal are added to the data set, while the data points detected to be abnormal are eliminated, so that the data points that are finally left in the data set are those with relatively high SOH accuracy. Therefore, when the above-mentioned first battery trend model is subsequently updated based on this data set, a first battery trend model with higher accuracy can be obtained.

[0254] In yet another optional solution, the multiple data points in the data set are all data points of batteries of the target class, and the processor is further configured to: train the second battery trend model using the data set.

[0255] It can be understood that the data points detected to be normal are added to the data set, while the data points detected to be abnormal are eliminated, so that the data points that are finally left in the data set are those with relatively high SOH accuracy. Therefore, when the above-mentioned second battery trend model is subsequently updated based on this data set, a second battery trend model with higher accuracy can be obtained.

[0256] In another optional scheme, after adding the first data point to the data set, the processor is further used to: train an SOH evaluation model through the SOH in each data point in the data set and the status data corresponding to one or more time segments in a battery cycle of each data point, wherein the SOH evaluation model is used to predict the SOH of the battery.

[0257] It can be understood that normal data points are added to the dataset, while abnormal data points are removed, which means that the data points that remain in the dataset are those with relatively high SOH accuracy. Therefore, based on this dataset, a SOH evaluation model with better prediction effect can be trained. In the future, when SOH cannot be determined directly by inference, SOH can be predicted through the SOH evaluation model.

[0258] In yet another optional solution, after training the SOH evaluation model using the SOH in each data point in the data set and the state data corresponding to one or more time segments in a battery cycle of each data point, the processor is further configured to:

[0259] Acquire second status data of the first battery in a first time period;

[0260] If the first time period does not belong to the first preset period, inputting the second state data into the SOH evaluation model to obtain a second SOH;

[0261] If the distance that the second data point deviates from the second battery trend model is less than the third preset threshold, and the distance that the second data point deviates from the first battery trend model is less than the fourth preset threshold, the second SOH is determined as the SOH of the first battery in the first time period, and the second data point includes the second state data and the second SOH.

[0262] In this approach, the battery's SOH is predicted using the SOH assessment model for status data outside the first preset cycle, resolving the prior art issue of being unable to determine the SOH based on status data outside the first preset cycle. Furthermore, the SOH predicted by the SOH assessment model is further tested for anomalies using the first and second battery trend models, helping to eliminate abnormal SOH and improving SOH accuracy.

[0263] In yet another optional solution, the processor is further configured to:

[0264] If the first time period belongs to a first preset cycle, obtaining a third SOH of the first battery according to the second state data of the first battery in the first time period;

[0265] If the distance that the third data point deviates from the first battery trend model is less than the first preset threshold, and the distance that the third data point deviates from the second battery trend model is less than the second preset threshold, the third SOH is determined as the SOH of the first battery in the first time period, wherein the third data point includes the second state data and the third SOH.

[0266] In this way, the SOH of the battery is directly calculated for the status data belonging to the first preset cycle, and the data points are detected as outliers through the first battery trend model and the second battery trend model to determine whether the SOH in the data point is abnormal. Only when the SOH is detected to be normal will the SOH in the data point be used as the SOH of the first battery in the target battery cycle, thereby improving the accuracy of the SOH.

[0267] In yet another optional solution, the processor is further configured to:

[0268] If the distance of the third data point deviating from the first battery trend model is greater than a first preset threshold, or the distance of the third data point deviating from the second battery trend model is greater than a second preset threshold, inputting the second state data into the SOH evaluation model to obtain a fourth SOH;

[0269] If the distance that the fourth data point deviates from the second battery trend model is less than the third preset threshold, and the distance that the second data point deviates from the first battery trend model is less than the fourth preset threshold, the fourth SOH is determined as the SOH of the first battery in the first time period, and the fourth data point includes the second state data and the fourth SOH.

[0270] In this way, the SOH of the battery is directly calculated for the status data belonging to the first preset cycle, and the data points are detected for outliers through the first battery trend model and the second battery trend model to determine whether the SOH in the data points is abnormal. If an abnormal SOH is detected, the SOH evaluation model is further used to predict the SOH, and ultimately an SOH with relatively high accuracy can always be obtained.

[0271] In yet another optional solution, the processor is further configured to: merge the SOH of the first battery in the first time period with the historical SOH of the first battery to obtain a change trajectory of the SOH of the first battery.

[0272] Through this method, the variation of the SOH of the first battery with the target parameter (such as mileage) can be obtained, which is convenient for users to judge the entire life cycle of the battery and make improvements based on this.

[0273] In another optional solution, the variation trajectory of the SOH of the first battery is used to characterize the variation relationship of the SOH with a target parameter, where the target parameter includes at least one of mileage, time, and number of cycles.

[0274] In yet another optional solution, in updating the first battery trend model using data points related to the first battery in the data set, the processor is specifically configured to:

[0275] determining a first state of charge (SOC) difference and a second state of charge (SOC) difference, wherein the first SOC difference is a difference between the charge of the first battery at a start point and an end point of a battery cycle corresponding to a fifth data point, and the second SOC difference is a difference between the charge of the first battery at a start point and an end point of the battery cycle corresponding to a sixth data point, wherein both the fifth data point and the sixth data point are data points related to the first battery in the data set;

[0276] The first battery trend model is updated according to the fifth data point, the first confidence level, the sixth data point, and the second confidence level, wherein the first confidence level is used to constrain the influence of the fifth data point on the first battery trend model, and the second confidence level is used to constrain the influence of the sixth data point on the first battery trend model.

[0277] By adopting this method, the influence of more data points on the first battery trend model (or the second battery trend model) can be taken into account, thereby improving the robustness of the first battery trend model. At the same time, when considering the influence of more data points, some atypical data points will inevitably be introduced. If such atypical data points are given the same influence as typical data points, then the accuracy of the first battery trend model trained based on these data points may not be guaranteed. Therefore, the embodiment of the present application proposes to set different confidence levels for different data points. By setting the confidence level, the influence of different data points can be differentiated, which can improve the accuracy of the first battery trend model finally trained.

[0278] In another optional solution, if the first charge SOC difference is greater than the second charge SOC difference, the first confidence level is greater than the second confidence level.

[0279] It should be noted that the implementation of each operation can also refer to Figure 3-Figure 5 The corresponding description of the method embodiment shown.

[0280] See Figure 9 , Figure 9 A battery status evaluation device 90 is provided in an embodiment of the present application. The device 90 includes a processor 901 and a memory 902. The processor 901 and the memory 902 can be interconnected via a bus.

[0281] The memory 902 includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or compact disc read-only memory (CD-ROM), and is used for related computer programs and data.

[0282] The processor 901 may be one or more central processing units (CPUs). In the case where the processor 901 is a CPU, the CPU may be a single-core CPU or a multi-core CPU.

[0283] The processor 901 in the device 90 is configured to read the computer program code stored in the memory 902 and perform the following operations:

[0284] Determining a first state of charge (SOC) difference and a second state of charge (SOC) difference, wherein the first SOC difference is a difference between the charge of the first battery at a start point and an end point of a battery cycle corresponding to a fifth data point, and the second SOC difference is a difference between the charge of the first battery at a start point and an end point of the battery cycle corresponding to a sixth data point, wherein both the fifth data point and the sixth data point are data points related to the first battery in the data set; the battery cycle is a charging cycle or a discharging cycle, wherein the charging cycle includes a process from the start of charging to the end of charging, and the discharging cycle includes a process from the start of discharging to the end of discharging;

[0285] The first battery trend model is updated according to the fifth data point, the first confidence level, the sixth data point, and the second confidence level, wherein the first confidence level is used to constrain the influence of the fifth data point on the first battery trend model, and the second confidence level is used to constrain the influence of the sixth data point on the first battery trend model.

[0286] By adopting this method, the influence of more data points on the first battery trend model (or the second battery trend model) can be taken into account, thereby improving the robustness of the first battery trend model. At the same time, when considering the influence of more data points, some atypical data points will inevitably be introduced. If such atypical data points are given the same influence as typical data points, then the accuracy of the first battery trend model trained based on these data points may not be guaranteed. Therefore, the embodiment of the present application proposes to set different confidence levels for different data points. By setting the confidence level, the influence of different data points can be differentiated, which can improve the accuracy of the first battery trend model finally trained.

[0287] In another optional solution, if the first charge SOC difference is greater than the second charge SOC difference, the first confidence level is greater than the second confidence level.

[0288] It should be noted that the implementation of each operation can also refer to Figure 3-Figure 5 The corresponding description of the method embodiment shown.

[0289] The embodiment of the present application further provides a computer-readable storage medium in which a computer program is stored, which, when executed on a processor, implements Figure 3 or Figure 4 or Figure 5 The method flow shown.

[0290] The present application also provides a computer program product, which, when executed on a processor, implements Figure 3 or Figure 4 or Figure 5The method flow shown.

[0291] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by a computer program or computer program-related hardware. The computer program can be stored in a computer-readable storage medium. When executed, the computer program can include the processes in the above-described method embodiments. The aforementioned storage medium includes various media capable of storing computer program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A battery status assessment method, characterized in that: include: Obtaining a first battery state of health (SOH) of the first battery according to first state data of the first battery in a target battery cycle; the battery cycle is a charging cycle or a discharging cycle, wherein the charging cycle includes a process from the start of charging to the end of charging of the battery, and the discharging cycle includes a process from the start of discharging to the end of discharging of the battery; If the distance that the first data point deviates from the first battery trend model is less than a first preset threshold, and the distance that the first data point deviates from the second battery trend model is less than a second preset threshold, the first SOH is determined as the SOH of the first battery in the target battery cycle, wherein the first data point includes the first state data and the first SOH, the first battery trend model is used to characterize the fitting curve of the changing trend of the SOH of the first battery, and the second battery trend model is used to characterize the fitting curve of the changing trend of the SOH of the target class of batteries, and the target class of batteries includes the first battery and at least one second battery.

2. The method according to claim 1, characterized in that The method further comprises: Adding the first data point to a data set, wherein the data set includes a plurality of data points, each data point including state data of a battery in a battery cycle and the SOH of the battery in the battery cycle; The first battery trend model is trained using data points about the first battery in the data set.

3. The method according to claim 2, characterized in that The plurality of data points in the data set are all data points of batteries of the target class, and the method further includes: The second battery trend model is trained using the data set.

4. The method according to any one of claim 1, characterized in that After adding the first data point to the data set, the method further includes: The SOH evaluation model is trained using the SOH in each data point in the data set and the state data corresponding to one or more time segments in a battery cycle of each data point, wherein the SOH evaluation model is used to predict the SOH of the battery.

5. The method according to claim 4, characterized in that After training the SOH evaluation model using the SOH of each data point in the data set and the state data corresponding to one or more time segments in a battery cycle of each data point, the method further includes: Acquire second status data of the first battery in a first time period; If the first time period does not belong to the first preset period, inputting the second state data into the SOH evaluation model to obtain a second SOH; If the distance that the second data point deviates from the second battery trend model is less than the third preset threshold, and the distance that the second data point deviates from the first battery trend model is less than the fourth preset threshold, the second SOH is determined as the SOH of the first battery in the first time period, and the second data point includes the second state data and the second SOH.

6. The method according to claim 5, characterized in that Also includes: If the first time period belongs to a first preset cycle, obtaining a third SOH of the first battery according to the second state data of the first battery in the first time period; If the distance that the third data point deviates from the first battery trend model is less than the first preset threshold, and the distance that the third data point deviates from the second battery trend model is less than the second preset threshold, the third SOH is determined as the SOH of the first battery in the first time period, wherein the third data point includes the second state data and the third SOH.

7. The method according to claim 6, characterized in that Also includes: If the distance of the third data point deviating from the first battery trend model is greater than a first preset threshold, or the distance of the third data point deviating from the second battery trend model is greater than a second preset threshold, inputting the second state data into the SOH evaluation model to obtain a fourth SOH; If the distance that the fourth data point deviates from the second battery trend model is less than the third preset threshold, and the distance that the second data point deviates from the first battery trend model is less than the fourth preset threshold, the fourth SOH is determined as the SOH of the first battery in the first time period, and the fourth data point includes the second state data and the fourth SOH.

8. The method according to any one of claims 5 to 7, characterized in that: Also includes: The SOH of the first battery in the first time period is combined with the historical SOH of the first battery to obtain a change trajectory of the SOH of the first battery.

9. The method according to claim 8, characterized in that The variation trajectory of the SOH of the first battery is used to characterize the variation relationship of the SOH with a target parameter, where the target parameter includes at least one of mileage, time, and number of cycles.

10. The method according to any one of claim 2, characterized in that Updating the first battery trend model using data points related to the first battery in the data set includes: determining a first state of charge (SOC) difference and a second state of charge (SOC) difference, wherein the first SOC difference is a difference between the charge of the first battery at a start point and an end point of a battery cycle corresponding to a fifth data point, and the second SOC difference is a difference between the charge of the first battery at a start point and an end point of the battery cycle corresponding to a sixth data point, wherein both the fifth data point and the sixth data point are data points related to the first battery in the data set; The first battery trend model is fitted according to a first confidence level using the fifth data point, and the first battery trend model is fitted according to a second confidence level using the sixth data point, wherein the first confidence level is used to constrain the influence of the fifth data point on the first battery trend model, and the second confidence level is used to constrain the influence of the sixth data point on the first battery trend model.

11. The method according to claim 10, characterized in that If the first charge SOC difference is greater than the second charge SOC difference, the first confidence level is greater than the second confidence level.

12. The method according to any one of claim 2, characterized in that: Updating the first battery trend model using data points related to the first battery in the data set includes: Determining a first state of charge (SOC) difference and a second state of charge (SOC) difference, wherein the first SOC difference is a difference between the charge of the first battery at a start point and an end point of a battery cycle corresponding to the fifth data point, and the second SOC difference is a difference between the charge of the first battery at a start point and an end point of the battery cycle corresponding to the sixth data point, wherein both the fifth data point and the sixth data point are data points related to the first battery in the data set; the battery cycle is a charging cycle or a discharging cycle, wherein the charging cycle includes a process from the start of charging to the end of charging of the battery, and the discharging cycle includes a process from the start of discharging to the end of discharging of the battery; The first battery trend model is fitted according to a first confidence level using the fifth data point, and the first battery trend model is fitted according to a second confidence level using the sixth data point, wherein the first confidence level is used to constrain the influence of the fifth data point on the first battery trend model, and the second confidence level is used to constrain the influence of the sixth data point on the first battery trend model.

13. The method according to claim 12, characterized in that If the first charge SOC difference is greater than the second charge SOC difference, the first confidence level is greater than the second confidence level.

14. A battery status evaluation device, characterized in that: include: a first acquiring unit, configured to acquire a first state of health (SOH) of the first battery according to first state data of the first battery in a target battery cycle; the battery cycle is a charging cycle or a discharging cycle, wherein the charging cycle includes a process from the start of charging to the end of charging of the battery, and the discharging cycle includes a process from the start of discharging to the end of discharging of the battery; A first determination unit is configured to determine the first SOH as the SOH of the first battery in the target battery cycle when a distance at which a first data point deviates from a first battery trend model is less than a first preset threshold value and a distance at which the first data point deviates from a second battery trend model is less than a second preset threshold value, wherein the first data point includes the first status data and the first SOH, the first battery trend model is used to characterize a fitting curve of a changing trend of the SOH of the first battery, and the second battery trend model is used to characterize a fitting curve of a changing trend of the SOH of batteries of a target class, and the batteries of the target class include the first battery and at least one second battery.

15. The device according to claim 14, characterized in that The device further comprises: an adding unit, configured to add the first data point to a data set, wherein the data set includes a plurality of data points, each data point including state data of a battery in a battery cycle and the SOH of the battery in the battery cycle; A first training unit is configured to train the first battery trend model using data points about the first battery in the data set.

16. The device according to claim 15, characterized in that The plurality of data points in the data set are all data points of batteries of the target class, and the apparatus further comprises: The second training unit is configured to train the second battery trend model using the data set.

17. The device according to any one of claims 15, characterized in that Also includes: A third training unit is used to train an SOH evaluation model after adding the first data point to the data set through the SOH in each data point in the data set and the status data corresponding to one or more time segments in a battery cycle of each data point, wherein the SOH evaluation model is used to predict the SOH of the battery.

18. The device according to claim 17, characterized in that Also includes: a second acquiring unit, configured to acquire second state data of the first battery in a first time period after training the SOH evaluation model using the SOH in each data point in the data set and state data corresponding to one or more time segments in a battery cycle of each data point; a first input unit, configured to input the second state data into the SOH evaluation model to obtain a second SOH when the first time period does not belong to a first preset period; A second determination unit is configured to determine the second SOH as the SOH of the first battery in the first time period when a distance that the second data point deviates from the second battery trend model is less than a third preset threshold and a distance that the second data point deviates from the first battery trend model is less than a fourth preset threshold, wherein the second data point includes the second state data and the second SOH.

19. The device according to claim 18, characterized in that Also includes: a third acquiring unit, configured to acquire a third SOH of the first battery according to the second state data of the first battery in the first time period when the first time period belongs to a first preset cycle; a third determination unit, configured to determine the third SOH as the SOH of the first battery in the first time period when a distance that the third data point deviates from the first battery trend model is less than a first preset threshold and a distance that the third data point deviates from the second battery trend model is less than a second preset threshold, wherein the third data point includes the second state data and the third SOH.

20. The device according to claim 19, characterized in that Also includes: a second input unit, configured to input the second state data into the SOH evaluation model to obtain a fourth SOH when a distance that the third data point deviates from the first battery trend model is greater than a first preset threshold, or a distance that the third data point deviates from the second battery trend model is greater than a second preset threshold; a fourth determination unit, configured to determine the fourth SOH as the SOH of the first battery in the first time period when the distance a fourth data point deviates from the second battery trend model is less than a third preset threshold and the distance a second data point deviates from the first battery trend model is less than a fourth preset threshold, wherein the fourth data point includes the second state data and the fourth SOH.

21. The device according to any one of claims 18 to 20, characterized in that Also includes: A processing unit is configured to combine the SOH of the first battery in the first time period with the historical SOH of the first battery to obtain a change trajectory of the SOH of the first battery.

22. The device according to claim 21, characterized in that The variation trajectory of the SOH of the first battery is used to characterize the variation relationship of the SOH with a target parameter, where the target parameter includes at least one of mileage, time, and number of cycles.

23. The device according to any one of claims 15, characterized in that In terms of updating the first battery trend model using data points about the first battery in the data set, the first training unit is specifically configured to: determining a first state of charge (SOC) difference and a second state of charge (SOC) difference, wherein the first SOC difference is a difference between the charge of the first battery at a start point and an end point of a battery cycle corresponding to a fifth data point, and the second SOC difference is a difference between the charge of the first battery at a start point and an end point of the battery cycle corresponding to a sixth data point, wherein both the fifth data point and the sixth data point are data points related to the first battery in the data set; The first battery trend model is fitted using the fifth data point according to a first confidence level, and the first battery trend model is fitted using the sixth data point according to a second confidence level.

24. The device according to claim 23, characterized in that If the first charge SOC difference is greater than the second charge SOC difference, the first confidence level is greater than the second confidence level.

25. The device according to any one of claims 15, characterized in that In terms of updating the first battery trend model using data points about the first battery in the data set, the first training unit is specifically configured to: a fifth determining unit, configured to determine a first charge state occupation (SOC) difference and a second charge state occupation (SOC) difference, wherein the first SOC difference is a charge difference between a start point and an end point of a battery cycle corresponding to a fifth data point of the first battery, and the second SOC difference is a charge difference between a start point and an end point of the battery cycle corresponding to a sixth data point of the first battery, wherein both the fifth data point and the sixth data point are data points related to the first battery in the data set; the battery cycle is a charging cycle or a discharging cycle, wherein the charging cycle includes a process from the start of charging to the end of charging of the battery, and the discharging cycle includes a process from the start of discharging to the end of discharging of the battery; The first battery trend model is fitted according to a first confidence level using the fifth data point, and the first battery trend model is fitted according to a second confidence level using the sixth data point, wherein the first confidence level is used to constrain the influence of the fifth data point on the first battery trend model, and the second confidence level is used to constrain the influence of the sixth data point on the first battery trend model.

26. The device according to claim 25, characterized in that If the first charge SOC difference is greater than the second charge SOC difference, the first confidence level is greater than the second confidence level.

27. A battery status evaluation device, characterized in that: The method comprises a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to call the computer program to implement the method according to any one of claims 1 to 13.

28. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed on a processor, implements the method according to any one of claims 1 to 13.

Images