Electric quantity consistency determination method and device, equipment, storage medium and product

Abstract

The invention relates to an electric quantity consistency determination method and device, equipment, a storage medium and a product, and the method comprises the steps: obtaining the state data of a target battery in a charging process, carrying out the charging working condition compensation of the state data through a preset working condition compensation model, obtaining the compensation state data, and carrying out the calculation of the electric quantity consistency according to the compensation state data. And determining an electric quantity consistency monitoring result of the target battery at the current moment. By adopting the method, the electric quantity consistency of the battery can be accurately determined.

Description

Technical Field

[0001] The present application relates to the technical field of batteries, and in particular, to a method, device, equipment, storage medium, and product for determining the charge consistency. Background Art

[0002] Generally, the power batteries of electric vehicles are combined by multiple battery cells or parallel modules in series-parallel connection to achieve the battery capacity and output power required for the endurance and power of electric vehicles.

[0003] In the related art, the power battery needs to experience multiple charge-discharge cycles. As the number of cycles increases, the charge consistency among the battery cells or parallel modules in the power battery will also change dynamically. When the charge consistency expands to a certain extent, faults such as undervoltage and low capacity will occur in the power battery.

[0004] Therefore, there is an urgent need for a method that can accurately determine the charge consistency of the battery. Summary of the Invention

[0005] Based on this, in view of the above technical problems, it is necessary to provide a method, device, equipment, storage medium, and product for determining the charge consistency, which can accurately determine the charge consistency of the battery.

[0006] In a first aspect, an embodiment of the present application provides a method for determining charge consistency, including:

[0007] Obtaining the state data of the target battery during the charging process;

[0008] Performing charge condition compensation on the state data through a preset working condition compensation model to obtain compensated state data;

[0009] Determining the charge consistency monitoring result of the target battery at the current moment according to the compensated state data.

[0010] In the embodiment of the present application, the state data of the target battery during charging is obtained, and the charging condition compensation is performed on the state data through a preset working condition compensation model to obtain compensated state data. Then, based on the compensated state data, the power consistency monitoring result of the target battery at the current moment is determined. In this embodiment, since the working condition compensation model is a model for compensating the state data of the battery under the charging condition, it can compensate the state data under the charging condition to the standard state. Therefore, by performing working condition compensation on the state data of the target battery through the pre-trained working condition compensation model, the influence brought by collecting the state data of the target battery during charging can be reduced, so that the obtained compensated state data is the actual state data of the target battery. Calculating the power consistency monitoring result of the target battery at the current moment with the actual state data improves the accuracy of the power consistency monitoring result of the target battery; moreover, determining the power consistency monitoring result of the target battery through the state data of the target battery during charging is equivalent to calculating the power consistency of the battery using the dynamic characteristics of the battery, which can greatly increase the calculation frequency of the power consistency of the battery, thereby achieving accurate monitoring of the power consistency of the target battery.

[0011] In one embodiment, performing charging condition compensation on the state data through a preset working condition compensation model to obtain compensated state data includes:

[0012] Obtain the data to be compensated of the target battery during charging from the state data;

[0013] Perform charging condition compensation on the data to be compensated through the working condition compensation model to obtain compensated voltage data;

[0014] Determine the compensated data and other state data in the state data as the compensated state data.

[0015] In the embodiment of the present application, the voltage data of the target battery during charging is obtained from the state data, and the charging condition compensation is performed on the data to be compensated through the working condition compensation model to obtain compensated data. Then, the compensated data and other state data in the state data are determined as the compensated state data. In this embodiment, by compensating the data to be compensated of the target battery during charging through the working condition compensation model, the influence of the charging condition change of the target battery on the data to be compensated can be reduced, making the compensated data more accurate, thereby further improving the accuracy of the battery consistency monitoring result of the target battery.

[0016] In one embodiment, the data to be compensated includes the first voltage and the second voltage of the target battery at each moment during charging; performing charging condition compensation on the data to be compensated through the working condition compensation model to obtain compensated voltage data includes:

[0017] At any moment, obtain the first voltage compensation amount and the second voltage compensation amount at the moment through the working condition compensation model;

[0018] Superimpose the first voltage compensation amount and the first voltage to obtain the compensated first voltage; and superimpose the second voltage compensation amount and the second voltage to obtain the compensated second voltage;

[0019] Determine the compensated first voltage and the compensated second voltage as the compensated voltage data at the moment.

[0020] In the embodiment of the present application, the voltage data includes the first voltage and the second voltage of the target battery at each moment during the charging process; at any moment, obtain the first voltage compensation amount and the second voltage compensation amount at the moment through the working condition compensation model, and superimpose the first voltage compensation amount and the first voltage to obtain the compensated first voltage, and superimpose the second voltage compensation amount and the second voltage to obtain the compensated second voltage. Finally, determine the compensated first voltage and the compensated second voltage as the compensated voltage data at the moment. In this embodiment, during the charging process of the target battery, since the current of the target battery changes and the internal state of the battery changes, this will cause the voltage of the target battery not to be in a stable state, and the collected first voltage and second voltage are not in a standard state. Therefore, voltage compensation is performed by obtaining the voltage compensation amounts corresponding to the first voltage and the second voltage, so that the obtained compensated voltage data is closer to the steady-state value, thereby greatly improving the accuracy of the power consistency monitoring result of the target battery.

[0021] In one embodiment, the working condition compensation model includes a first voltage compensation model and a second voltage compensation model; the state data includes the current of the target battery at each moment during the charging process; obtaining the first voltage compensation amount and the second voltage compensation amount at the moment through the working condition compensation model includes:

[0022] Input the current at the moment into the first voltage compensation model to obtain the first voltage compensation amount;

[0023] Input the current at the moment into the second voltage compensation model to obtain the second voltage compensation amount.

[0024] In the embodiment of the present application, input the current at the moment into the first voltage compensation model to obtain the first voltage compensation amount, and input the current at the moment into the second voltage compensation model to obtain the second voltage compensation amount. In this embodiment, since the current change will cause the voltage change, therefore, the corresponding voltage compensation amount can be determined according to the current, and the first voltage compensation amount and the second voltage compensation amount are determined through the pre-trained first voltage compensation model and the second voltage compensation model, so that the compensated voltage data is closer to the stable value and the accuracy of the compensated voltage data is improved.

[0025] In one embodiment, the obtaining process of the working condition compensation model includes:

[0026] Obtain the historical state data of the historical battery during multiple historical charging processes;

[0027] For any one of the historical charging processes, obtain the absolute value of the current difference at each adjacent moment during the historical charging process;

[0028] Determine the historical state data corresponding to the adjacent moments with the absolute value of the current difference greater than the preset threshold as the reference state data at each adjacent moment during the historical charging process;

[0029] Construct a working condition compensation model according to the reference state data at each adjacent moment.

[0030] In the embodiments of the present application, the historical state data of the historical battery during multiple historical charging processes is obtained. For any one of the historical charging processes, the absolute value of the current difference at each adjacent moment during the historical charging process is obtained, and the historical state data corresponding to the adjacent moments with the absolute value of the current difference greater than the preset threshold is determined as the reference state data at each adjacent moment during the historical charging process. Finally, a working condition compensation model is constructed according to the reference state data at each adjacent moment. In this embodiment, the historical state data corresponding to the adjacent moments with the absolute value of the current difference greater than the preset threshold is determined as the reference state data at each adjacent moment during the historical charging process. In this way, it is possible to obtain the changes in other parameters in the state data brought about by the obvious current change, making the reference state data more reliable, thereby improving the reliability and accuracy of the construction of the working condition compensation model; constructing a working condition compensation model according to the reference state data at each adjacent moment can improve the reliability and accuracy of the working condition compensation model for compensating the state data.

[0031] In one of the embodiments, constructing a working condition compensation model according to the reference state data at each adjacent moment includes:

[0032] Extract features from the reference state data at each adjacent moment to determine the historical charging characteristics at each adjacent moment;

[0033] Construct a working condition compensation model according to each historical charging characteristic.

[0034] In the embodiments of the present application, features are extracted from the reference state data at each adjacent moment to determine the historical charging characteristics at each adjacent moment, and a working condition compensation model is constructed according to each historical charging characteristic. In this embodiment, since the working condition compensation model compensates the state data under the charging working condition, and the historical charging characteristics at each adjacent moment can reflect the change of the state data at the adjacent moment, therefore, constructing a working condition compensation model through the historical charging characteristics of the reference state data at each adjacent moment can enable the working condition compensation model to accurately compensate the state data.

[0035] In one embodiment, the historical charging characteristics include a first voltage difference, a second voltage difference, a current difference, a temperature, and an accumulated charge amount; according to each historical charging characteristic, a working condition compensation model is constructed, including:

[0036] According to the corresponding relationship between each first voltage difference and each current difference, and the corresponding relationship between each second voltage difference and each current difference, outlier filtering is performed on the historical charging characteristics to obtain filtered charging characteristics;

[0037] According to the filtered charging characteristics, a multiple regression model is constructed;

[0038] Substitute a preset reference temperature, a preset reference accumulated charge amount, and a preset reference current into the multiple regression model to obtain the working condition compensation model.

[0039] In the embodiment of the present application, the historical charging characteristics include a first voltage difference, a second voltage difference, a current difference, a temperature, and an accumulated charge amount; according to the corresponding relationship between each first voltage difference and each current difference, and the corresponding relationship between each second voltage difference and each current difference, outlier filtering is performed on the historical charging characteristics to obtain filtered charging characteristics, and according to the filtered charging characteristics, a multiple regression model is constructed, and then a preset reference temperature, a preset reference accumulated charge amount, and a preset reference current are substituted into the multiple regression model to obtain the working condition compensation model. In this embodiment, since the working condition compensation model compensates the voltage under the charging working condition, during the battery charging process, the current, temperature, and SOC of the battery will all affect the voltage. Therefore, a multiple regression model is constructed through the current difference, voltage difference, temperature, and accumulated charge amount, and the reference current, reference temperature, and reference accumulated charge amount in the preset reference state are substituted into the multiple regression model, so that the finally constructed working condition compensation model can accurately obtain the voltage compensation amount required for the voltage to reach the standard state, further improving the accuracy of the power consistency monitoring result of the target battery; and before constructing the multiple regression model, outlier filtering is performed on the historical charging characteristics, reducing the data deviation of the abnormal points and improving the accuracy of the working condition compensation model.

[0040] In one embodiment, the multiple regression model includes a first multiple regression model and a second multiple regression model; according to the filtered charging characteristics, constructing a multiple regression model includes:

[0041] According to the first voltage difference, current difference, temperature, and accumulated charge amount in the filtered charging characteristics, a first multiple regression model is constructed;

[0042] According to the second voltage difference, current difference, temperature, and accumulated charge amount in the filtered charging characteristics, a second multiple regression model is constructed.

[0043] In the embodiments of the present application, a first multiple regression model is constructed according to the first voltage difference, current difference, temperature, and accumulated charge in the filtered charging characteristics, and a second multiple regression model is constructed according to the second voltage difference, current difference, temperature, and accumulated charge in the filtered charging characteristics. In this embodiment, since it is necessary to process the first voltage and the second voltage in the state data, constructing the first multiple regression model corresponding to the first voltage and the second multiple regression model corresponding to the second voltage in a targeted manner makes the voltage compensation amounts corresponding to the first voltage and the second voltage more accurate.

[0044] In one of the embodiments, according to the compensated state data, determining the power consistency monitoring result of the target battery at the current moment includes:

[0045] In the case where the compensated voltage data in the compensated state data exceeds the platform voltage range, the compensated state data within the platform voltage range is determined as the target state data;

[0046] In the case where the compensated voltage data in the compensated state data does not exceed the platform voltage range, according to the compensated state data and the historical state data, determining the target state data; the historical state data includes the state data corresponding to the historical compensated voltage data that did not exceed the platform voltage range within the historical time period;

[0047] According to the target state data, determining the power consistency monitoring result of the target battery at the current moment.

[0048] In the embodiments of the present application, in the case where the compensated voltage data in the compensated state data exceeds the platform voltage range, the compensated state data within the platform voltage range is determined as the target state data; in the case where the compensated voltage data in the compensated state data does not exceed the platform voltage range, according to the compensated state data and the historical state data, determining the target state data; the historical state data includes the state data corresponding to the historical compensated voltage data that did not exceed the platform voltage range within the historical time period; finally, according to the target state data, determining the power consistency monitoring result of the target battery at the current moment. In this embodiment, when obtaining the target state data from the compensated state data, first, it is determined whether the compensated voltage data of the compensated state data exceeds the platform voltage range, and according to different situations, the compensated state data is specifically analyzed, so that the target state data is more comprehensive and accurate, and the accuracy of the power consistency monitoring result is improved.

[0049] In one of the embodiments, according to the compensated state data and the historical state data, determining the target state data includes:

[0050] Combining the compensated state data and the historical state data to obtain combined state data;

[0051] According to the compensation status data and the repeated status data corresponding to the historical status data at each power level, determine the repeated status data with the highest priority as the unique status data corresponding to each power level;

[0052] Filter the merged status data according to each unique status data to obtain the integrated status data;

[0053] Obtain the target status data from the integrated status data according to the platform voltage range.

[0054] In the embodiment of the present application, the compensation status data and the historical status data are merged to obtain the merged status data. According to the repeated status data corresponding to the compensation status data and the historical status data at each power level, the repeated status data with the highest priority is determined as the unique status data corresponding to each power level. Then, the merged status data is filtered according to each unique status data to obtain the integrated status data. Finally, the target status data is obtained from the integrated status data according to the platform voltage range. In this embodiment, when merging the compensation status data and the historical status data, the repeated status data is filtered by priority, making the integrated status data more complete and not including repeated status data, improving the integrity and effectiveness of the integrated status data, enabling the determination of the power consistency monitoring result of the target battery through the integrated status data, and enhancing the computing opportunity of power consistency.

[0055] In one of the embodiments, before determining the repeated status data with the highest priority as the unique status data corresponding to each power level, the method further includes:

[0056] Obtain the first power level charged by the compensation status data within the platform voltage range, and the second power level charged by the historical status data within the platform voltage range;

[0057] Determine the priority of each repeated status data according to the first power level and the second power level.

[0058] In the embodiment of the present application, obtain the first power level charged by the compensation status data within the platform voltage range, and the second power level charged by the historical status data within the platform voltage range, and then determine the priority of each repeated status data according to the first power level and the second power level. In this embodiment, the priority of the compensation status data and the historical status data is evaluated based on the magnitude of the power level charged within the platform voltage range. In this way, when selecting the repeated status data, the most complete status information can be retained, making the integrated status data more reliable.

[0059] In one of the embodiments, the target status data includes the compensated first voltage and the compensated second voltage at multiple moments; according to the target status data, determining the power consistency monitoring result of the target battery at the current moment includes:

[0060] Determine the mean and variance of the pressure difference of the target state data based on the compensated first voltage and the compensated second voltage at each moment;

[0061] Determine the charging characteristics of the target battery based on the mean and variance of the pressure difference;

[0062] Input the charging characteristics into the state classification model to obtain the monitoring result of the power consistency of the target battery at the current moment.

[0063] In the embodiment of the present application, based on the compensated first voltage and the compensated second voltage at each moment, the mean and variance of the pressure difference of the target state data are determined, and based on the mean and variance of the pressure difference, the charging characteristics of the target battery are determined, and then the charging characteristics are input into the state classification model to obtain the monitoring result of the power consistency of the target battery at the current moment. In this embodiment, since the mean and variance of the pressure difference can reflect the health state of the battery, therefore, using the mean and variance of the pressure difference as the charging characteristics of the target terminal battery can achieve accurate monitoring of the power consistency of the target battery; and, by determining the monitoring result of the power consistency through the preset state classification model, directly inputting the charging characteristics into the state classification model can obtain the monitoring result of the power consistency of the target battery, which improves the speed of determining the monitoring result of the power consistency.

[0064] In one of the embodiments, the state classification model includes a primary classification model and a secondary classification model; inputting the charging characteristics into the state classification model to obtain the monitoring result of the power consistency of the target battery at the current moment includes:

[0065] Input the charging characteristics into the primary classification model to obtain a preliminary classification result;

[0066] In the case where the preliminary classification result is the first value, determine that the monitoring result of the power consistency at the current moment is good; good means that the power consistency is less than or equal to the first consistency threshold;

[0067] In the case where the preliminary classification result is the second value, input the charging characteristics into the secondary classification model to obtain the monitoring result of the power consistency of the target battery at the current moment.

[0068] In the embodiments of the present application, charging characteristics are input into a primary classification model to obtain a preliminary classification result; when the preliminary classification result is a first value, it is determined that the power consistency monitoring result at the current moment is good; good means that the power consistency is less than or equal to a first consistency threshold; when the preliminary classification result is a second value, the charging characteristics are input into a secondary classification model to obtain the power consistency monitoring result of the target battery at the current moment. In this embodiment, by inputting the charging characteristics into the primary classification model, a preliminary classification result can be quickly obtained. When the initial classification result is the first value, the power consistency monitoring result of the target battery can be directly determined, reducing the workload of calculating the power consistency; when the initial classification result is the second value, the secondary classification model is continued to perform a detailed classification of the power consistency to refine the power consistency monitoring result, improving the accuracy of the power consistency monitoring result.

[0069] In one of the embodiments, inputting the charging characteristics into the secondary classification model to obtain the power consistency monitoring result of the target battery at the current moment includes:

[0070] Inputting the charging characteristics into the secondary classification model to obtain an advanced classification result;

[0071] When the advanced classification result is a third value, it is determined that the power consistency monitoring result at the current moment is average, and average means that the power consistency is in the interval between the first consistency threshold and the second consistency threshold;

[0072] When the advanced classification result is a fourth value, it is determined that the power consistency monitoring result at the current moment is poor; poor means that the power consistency is greater than or equal to the second consistency threshold.

[0073] In the embodiments of the present application, the charging characteristics are input into the secondary classification model to obtain an advanced classification result, and when the advanced classification result is a third value, it is determined that the power consistency monitoring result at the current moment is average, where average means that the power consistency is in the interval between the first consistency threshold and the second consistency threshold; when the advanced classification result is a fourth value, it is determined that the power consistency monitoring result at the current moment is poor; poor means that the power consistency is greater than or equal to the second consistency threshold. In this embodiment, when the power consistency of the target battery is not good, the secondary classification model is used to perform a fine classification of the average level and poor situation of the power consistency, improving the accuracy of the power consistency monitoring result of the target battery.

[0074] In one of the embodiments, the obtaining process of the state classification model includes:

[0075] Obtaining the historical charging characteristics of the historical battery during multiple historical charging processes, and obtaining the power consistency results of the historical battery at a preset power;

[0076] Match each historical charging feature with each power consistency result according to time to obtain a training data set;

[0077] Train the initial state classification model according to the training data set to obtain a state classification model.

[0078] In the embodiment of the present application, the historical charging features of the historical battery in multiple historical charging processes are obtained, and the power consistency results of the historical battery at a preset power are obtained, and each historical charging feature is matched with each power consistency result according to time to obtain a training data set, and then the initial state classification model is trained according to the training data set to obtain a state classification model. In this embodiment, the state classification model is trained with the historical charging feature and the power consistency result as the training data set, so that the power consistency of the battery can be predicted through the charging feature of the battery in the future; and when constructing the training data set, the historical charging feature and the power consistency result are matched according to time, so that the historical charging feature and the power consistency result are more corresponding, improving the accuracy of the training data set, and thus improving the accuracy of the state classification model.

[0079] In one of the embodiments, obtaining the power consistency result of the historical battery at a preset power includes:

[0080] For any preset power, obtain the maximum measured voltage and the minimum measured voltage of the historical battery at the preset power;

[0081] Substitute the maximum measured voltage and the minimum measured voltage into the voltage-power relationship formula respectively to obtain the first power corresponding to the maximum measured voltage and the second power corresponding to the minimum measured voltage;

[0082] Determine the difference between the first power and the second power as the power consistency result of the historical battery at the power.

[0083] In the embodiment of the present application, for any preset power, obtain the maximum measured voltage and the minimum measured voltage of the historical battery at the preset power, substitute the maximum measured voltage and the minimum measured voltage into the voltage-power relationship formula respectively to obtain the first power corresponding to the maximum measured voltage and the second power corresponding to the minimum measured voltage, and finally determine the difference between the first power and the second power as the power consistency result of the historical battery at the power. In this embodiment, the power consistency result at the preset power is obtained by the static voltage method, improving the accuracy of the training data set of the state classification model, and thus making the obtained state classification model more accurate.

[0084] In one of the embodiments, determining the power consistency monitoring result of the target battery at the current moment according to the target state data includes:

[0085] Extract features from the target status data to obtain the charging features of the target battery; the charging features include the bulging signal status and the average charging rate;

[0086] When the bulging signal status is that there is a bulging signal and the average charging rate is less than the preset charging threshold, input the charging features into the regression quantization model to obtain the prediction result of the power consistency of the target battery;

[0087] Determine the power consistency monitoring result of the target battery at the current moment according to the prediction result of the power consistency.

[0088] In the embodiment of the present application, features are extracted from the target status data to obtain the charging features of the target battery; the charging features include the bulging signal status and the average charging rate; when the bulging signal status is that there is a bulging signal and the average charging rate is less than the preset charging threshold, input the charging features into the regression quantization model to obtain the prediction result of the power consistency of the target battery; and determine the power consistency monitoring result of the target battery at the current moment according to the prediction result of the power consistency. In this embodiment, under specific working conditions of the battery, the power consistency of the battery has a strong correlation with the charging features of the battery. Therefore, when the bulging signal status is that there is a bulging signal and the average charging rate is less than the preset charging threshold, the power consistency monitoring result of the target battery can be determined through the charging features, thereby improving the accuracy of the power consistency monitoring result; and, the power consistency of the target battery is predicted through the regression quantization model, and then the true power consistency monitoring result of the target battery is determined based on the prediction result of the power consistency, improving the efficiency and accuracy of determining the power consistency monitoring result.

[0089] In one of the embodiments, extracting features from the target status data to obtain the charging features of the target battery includes:

[0090] Determine the pressure difference at each moment according to the compensated first voltage and the compensated second voltage at multiple moments in the target status data;

[0091] Determine the segmentation threshold and the pressure difference time series curve according to the pressure difference at each moment;

[0092] When the segmentation threshold divides the pressure difference time series curve into three continuous regions and the maximum pressure difference of the pressure difference time series curve is in the second region of the three continuous regions, determine that the bulging signal status of the target status data is that there is a bulging signal.

[0093] In the embodiments of the present application, according to the compensated first voltage and the compensated second voltage at multiple moments in the target state data, the pressure difference at each moment is determined. Then, according to the pressure difference at each moment, the segmentation threshold and the pressure difference time sequence curve are determined. Subsequently, when the pressure difference time sequence curve is segmented into three consecutive regions at the segmentation threshold and the maximum pressure difference of the pressure difference time sequence curve is in the second region among the three consecutive regions, it is determined that the bulging signal state of the target state data is that there is a bulging signal. In this embodiment, since the segmentation threshold is determined according to the pressure difference at each moment in the target state data, by segmenting the pressure difference time sequence curve with the segmentation threshold, it is possible to accurately determine whether there is a bulging signal in the target state data, improving the accuracy of determining the bulging signal state, and thus enhancing the reliability of the bulging signal state of the target state data.

[0094] In one of the embodiments, determining the segmentation threshold according to the pressure difference at each moment includes:

[0095] Downsample the pressure difference at each moment to obtain sampled pressure difference data;

[0096] Perform filtering processing on the sampled pressure difference data to obtain filtered pressure difference data;

[0097] Obtain the mean value of the pressure difference of the filtered pressure difference data;

[0098] Determine the segmentation threshold according to the mean value of the pressure difference and the maximum pressure difference in the filtered pressure difference data.

[0099] In the embodiments of the present application, the pressure difference at each moment is downsampled to obtain sampled pressure difference data, and the sampled pressure difference data is subjected to filtering processing to obtain filtered pressure difference data. Then, the mean value of the pressure difference of the filtered pressure difference data is obtained. Finally, the segmentation threshold is determined according to the mean value of the pressure difference and the maximum pressure difference in the filtered pressure difference data. In this embodiment, the segmentation threshold is the segmentation threshold of the pressure difference time sequence curve. Before calculating the segmentation threshold from the pressure difference at each moment, the pressure difference at each moment is downsampled and filtered, so that the data volume for calculating the segmentation threshold is smaller and more accurate, and the filtered pressure difference data after sampling and filtering is used to calculate the segmentation threshold, thereby improving the accuracy of the segmentation threshold.

[0100] In one of the embodiments, determining the power consistency monitoring result of the target battery at the current moment according to the power consistency prediction result includes:

[0101] Obtain the prediction error range of the regression quantization model at a preset consistency confidence level;

[0102] Determine the confidence interval of the current power consistency of the target battery according to the power consistency prediction result and the prediction error range; different confidence intervals represent different degrees of power consistency;

[0103] Determine the confidence interval of the current state of charge consistency as the monitoring result of the state of charge consistency of the target battery at the current moment.

[0104] In an embodiment of the present application, obtain the prediction error range of the regression quantization model at a preset consistency confidence level, and determine the confidence interval of the current state of charge consistency of the target battery according to the state of charge consistency prediction result and the prediction error range; wherein, different confidence intervals represent different degrees of state of charge consistency; then determine the confidence interval of the current state of charge consistency as the monitoring result of the state of charge consistency of the target battery at the current moment. In this embodiment, the confidence interval in which the state of charge consistency of the target battery is located is determined through the prediction error range of the regression quantization model at the consistency confidence level, so as to determine the distribution interval of the state of charge consistency of the target battery at the preset consistency confidence level, effectively measuring the accuracy of the monitoring result of the state of charge consistency of the target battery at the current moment.

[0105] In one embodiment, obtaining the prediction error range of the regression quantization model at a preset consistency confidence level includes:

[0106] Input the preset test data set into the regression quantization model to obtain multiple state of charge consistency prediction results output by the regression quantization model;

[0107] Obtain the difference between each state of charge consistency prediction result and the corresponding sample state of charge consistency result;

[0108] Determine the prediction error range of the regression quantization model at the consistency confidence level according to each difference.

[0109] In an embodiment of the present application, input the preset test data set into the regression quantization model to obtain multiple state of charge consistency prediction results output by the regression quantization model, then obtain the difference between each state of charge consistency prediction result and the corresponding sample state of charge consistency result, and finally determine the prediction error range of the regression quantization model at the consistency confidence level according to each difference. In this embodiment, since the state of charge consistency prediction result is predicted by the regression quantization model, and the sample state of charge consistency result is the true state of charge consistency result in the test data set, therefore, according to the difference between each state of charge consistency prediction result and the corresponding sample state of charge consistency result, the prediction error range of the regression quantization model at the consistency confidence level can be accurately measured.

[0110] In a second aspect, an embodiment of the present application further provides a state of charge consistency determination device, including:

[0111] A first acquisition module, configured to acquire the state data of the target battery during the charging process;

[0112] A compensation module, configured to perform charging condition compensation on the state data through a preset working condition compensation model to obtain compensated state data;

[0113] The first determination module is configured to determine the power consistency monitoring result of the target battery at the current moment according to the compensation status data.

[0114] In a third aspect, an embodiment of the present application provides a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of the method provided in any one of the embodiments in the first aspect are implemented.

[0115] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method provided in any one of the embodiments in the first aspect are implemented.

[0116] In a fifth aspect, an embodiment of the present application further provides a computer program product, including a computer program. When the computer program is executed by a processor, the steps of the method provided in any one of the embodiments in the first aspect are implemented.

[0117] The above description is only an overview of the technical solution of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the specific embodiments of the present application are specifically given below. BRIEF DESCRIPTION OF THE DRAWINGS

[0118] By reading the detailed description of the preferred embodiments below, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of the present application. And in all the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0119] Figure 1 is the internal structure diagram of a computer device in an embodiment;

[0120] Figure 2 is the flowchart of the power consistency determination method in an embodiment;

[0121] Figure 3 is the flowchart of the power consistency determination method in another embodiment;

[0122] Figure 4 is the flowchart of the power consistency determination method in another embodiment;

[0123] Figure 5 is the flowchart of the power consistency determination method in another embodiment;

[0124] Figure 6Schematic flowchart of the method for determining the battery capacity consistency in another embodiment;

[0125] Figure 7 Schematic flowchart of the method for determining the battery capacity consistency in another embodiment;

[0126] Figure 8 Schematic flowchart of the method for determining the battery capacity consistency in another embodiment;

[0127] Figure 9 Schematic flowchart of the method for determining the battery capacity consistency in another embodiment;

[0128] Figure 10 Schematic flowchart of the method for determining the battery capacity consistency in another embodiment;

[0129] Figure 11 Schematic flowchart of the method for determining the battery capacity consistency in another embodiment;

[0130] Figure 12 Schematic flowchart of the method for determining the battery capacity consistency in another embodiment;

[0131] Figure 13 Schematic flowchart of the method for determining the battery capacity consistency in another embodiment;

[0132] Figure 14 Schematic flowchart of the method for determining the battery capacity consistency in another embodiment;

[0133] Figure 15 Schematic flowchart of the method for determining the battery capacity consistency in another embodiment;

[0134] Figure 16 Schematic flowchart of the method for determining the battery capacity consistency in another embodiment;

[0135] Figure 17 Schematic flowchart of the method for determining the battery capacity consistency in another embodiment;

[0136] Figure 18 Schematic flowchart of the method for determining the battery capacity consistency in another embodiment;

[0137] Figure 19 Schematic flowchart of the method for determining the battery capacity consistency in another embodiment;

[0138] Figure 20 Schematic flowchart of the method for determining the battery capacity consistency in another embodiment;

[0139] Figure 21 Schematic flowchart of the method for determining the battery capacity consistency in another embodiment;

[0140] Figure 22 Schematic flow chart of the method for determining power quantity consistency in another embodiment;

[0141] Figure 23 Schematic flow chart of the method for determining power quantity consistency in another embodiment;

[0142] Figure 24 Schematic diagram of the change of the differential pressure time series curve in one embodiment;

[0143] Figure 25 Schematic diagram of the segmentation of the differential pressure time series curve after segmentation in one embodiment;

[0144] Figure 26 Schematic flow chart of the method for determining power quantity consistency in another embodiment;

[0145] Figure 27 Schematic diagram of the distribution of the predicted results of power quantity consistency and the corresponding sample power quantity consistency results in one embodiment;

[0146] Figure 28 Schematic diagram of the difference distribution of the predicted results of power quantity consistency and the corresponding sample power quantity consistency results in one embodiment;

[0147] Figure 29 Schematic diagram of the confidence interval of the sample power quantity consistency result in one embodiment;

[0148] Figure 30 Schematic flow chart of the method for determining power quantity consistency in another embodiment;

[0149] Figure 31 Schematic block diagram of the device for determining power quantity consistency in one embodiment;

[0150] Figure 32 Schematic block diagram of the device for determining power quantity consistency in another embodiment;

[0151] Figure 33 Schematic block diagram of the device for determining power quantity consistency in another embodiment;

[0152] Figure 34 Schematic block diagram of the device for determining power quantity consistency in another embodiment;

[0153] Figure 35 Schematic block diagram of the device for determining power quantity consistency in another embodiment;

[0154] Figure 36 Schematic block diagram of the device for determining power quantity consistency in another embodiment;

[0155] Figure 37 Schematic block diagram of the device for determining power quantity consistency in another embodiment;

[0156] Figure 38 is a structural block diagram of a power quantity consistency determination device in another embodiment;

[0157] Figure 39 is a structural block diagram of a power quantity consistency determination device in another embodiment;

[0158] Figure 40 is a structural block diagram of a power quantity consistency determination device in another embodiment;

[0159] Figure 41 is a structural block diagram of a power quantity consistency determination device in another embodiment;

[0160] Figure 42 is a structural block diagram of a power quantity consistency determination device in another embodiment;

[0161] Figure 43 is a structural block diagram of a power quantity consistency determination device in another embodiment;

[0162] Figure 44 is a structural block diagram of a power quantity consistency determination device in another embodiment;

[0163] Figure 45 is a structural block diagram of a power quantity consistency determination device in another embodiment;

[0164] Figure 46 is a structural block diagram of a power quantity consistency determination device in another embodiment;

[0165] Figure 47 is a structural block diagram of a power quantity consistency determination device in another embodiment;

[0166] Figure 48 is a structural block diagram of a power quantity consistency determination device in another embodiment;

[0167] Figure 49 is a structural block diagram of a power quantity consistency determination device in another embodiment;

[0168] Figure 50 is a structural block diagram of a power quantity consistency determination device in another embodiment;

[0169] Figure 51 is a structural block diagram of a power quantity consistency determination device in another embodiment. Detailed implementation manners

[0170] In order to make the objectives, technical solutions, and advantages of the present application more clearly understood, the embodiments of the technical solutions of the present application will be described in detail below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present application, and thus are only examples and cannot be used to limit the protection scope of the present application.

[0171] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above description of the drawings are intended to cover non-exclusive inclusion.

[0172] In the description of the embodiments of this application, technical terms such as "first" and "second" are only used to distinguish different objects and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity, specific order, or primary-secondary relationship of the indicated technical features. In the description of the embodiments of this application, the meaning of "a plurality" is more than two, unless otherwise specifically defined. The mention of "embodiment" in this article means that the specific features, structures, or characteristics described in connection with the embodiment can be included in at least one embodiment of this application. The appearance of this phrase in various positions in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0173] In the description of the embodiments of this application, the term "a plurality" refers to more than two (including two).

[0174] In practical applications, the power battery of an electric vehicle is usually composed of multiple battery cells or parallel modules combined in a series-parallel manner to achieve the battery capacity and output power required for the endurance and power of the electric vehicle.

[0175] However, due to the non - complete consistency of the physical and chemical properties and the environmental conditions among individual battery cells, as the number of charge - discharge cycles of the power battery increases, there will also be a dynamic change process in the state - of - charge (SOC) consistency among individual battery cells or parallel modules. The SOC consistency represents the battery charge consistency. Generally, the equalization control strategy of the battery management system (BMS) of an electric vehicle can maintain the charge consistency among battery cells in a good state by actively adjusting the charge between cells. However, under certain conditions, the battery charge consistency will show a gradually deteriorating trend, exceeding the range of the equalization adjustment ability. For example, there are large differences in the physical and chemical properties among cells, the driving conditions do not meet the trigger conditions of the equalization strategy, or there are relevant physical defects in the cells. When the charge consistency expands to a certain extent, early warning faults such as cell under - voltage and low capacity will occur, resulting in vehicle breakdown. Therefore, poor charge consistency will affect the power output of the battery pack, and in the case of large charge consistency, it may lead to vehicle breakdown.

[0176] Taking the lithium - iron phosphate battery as an example of the power battery, in the related art, the charge consistency estimation of the lithium - iron phosphate battery usually adopts the static voltage mapping method. mainly, the static voltage of the battery is mapped through the SOC - open - circuit voltage (OCV) curve of the cell to obtain the SOC of the battery, so as to calculate the charge consistency between cells.

[0177] However, since most of the SOC - OCV curve of the lithium - iron phosphate battery is in the flat region, a small change in OCV causes a large change in SOC. Therefore, the accuracy deviation of voltage acquisition will lead to a huge error in charge consistency. Moreover, this calculation method has less computing opportunity for lithium - iron phosphate batteries and is highly dependent on specific working conditions.

[0178] In another method, in order to increase the computing opportunity for charge consistency, the dynamic voltage characteristics are used to calculate the charge consistency of the battery. However, the voltage in this method is greatly affected by the consistency of cell internal resistance, voltage sampling deviation, and working condition differences such as temperature and current, resulting in a low accuracy rate of the calculated charge consistency.

[0179] Based on the above considerations, in order to improve the accuracy of calculating the charge consistency of the battery, the embodiment of this application proposes a method for determining charge consistency. By compensating the state data of the target battery through a pre - trained working condition compensation model, the influence brought by the state data collected during the charging process of the target battery can be reduced, so that the obtained compensated state data is the actual state data of the target battery. The charge consistency monitoring result of the target battery at the current moment is calculated based on the actual state data, thereby improving the accuracy of the charge consistency monitoring result of the target battery.

[0180] Of course, it should be understood that the technical effects that can be achieved by the power consistency determination method provided in the embodiments of the present application are not limited to this, and other technical effects can also be achieved. For example, determining the power consistency monitoring result of the target battery through the state data of the target battery during the charging process is equivalent to calculating the power consistency of the battery using the dynamic characteristics of the battery, which can improve the computing opportunity of the power consistency of the battery, and so on. For the specific technical effects that can be achieved in the embodiments of the present application, please refer to the following embodiments for details.

[0181] It should be noted that the power consistency determination solution provided in the present application is applicable to all new energy-related fields, including but not limited to batteries in new energy vehicles, energy storage, and quick change modules. The embodiments of the present application do not make any limitations in this regard.

[0182] The battery disclosed in the embodiments of the present application can be used in, but is not limited to, electrical devices such as vehicles, ships, or aircraft.

[0183] The embodiments of the present application provide an electrical device using a battery as a power source. The electrical device can be, but is not limited to, a mobile phone, a tablet computer, a laptop computer, an electric toy, an electric tool, a battery car, an electric vehicle, a ship, a spacecraft, and so on. Among them, the electric toy can include fixed or mobile electric toys, such as game consoles, electric vehicle toys, electric ship toys, and electric aircraft toys, etc. The spacecraft can include airplanes, rockets, space shuttles, and spaceships, etc.

[0184] For the convenience of description, the following embodiments are described with a computer device in an embodiment of the present application as the execution subject. The computer device is used to calculate the power consistency of the target battery in the electrical device. The computer device can be a server, and its internal structure diagram can be as Figure 1 shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store power consistency determination data. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals through a network connection. When the computer program is executed by the processor, it realizes the steps of the power consistency determination method provided in any one of the following embodiments of the present application.

[0185] Those skilled in the art can understand that Figure 1 The structure shown in Figure 1 is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0186] In an exemplary embodiment, as Figure 2 shown, a method for determining charge consistency is provided, including the following steps:

[0187] S201, obtain the status data of the target battery during the charging process.

[0188] Among them, the target battery is the battery that needs to be monitored for charge consistency, and the target battery includes at least two battery cells; the status data may include the status data of the target battery or the status data of the battery cells in the target battery; the status data of the target battery is used to reflect the status of the target battery; the status data may include parameter values such as charge rate, temperature, voltage, current, SOC, internal resistance, etc.

[0189] Therefore, when it is necessary to determine the charge consistency of the target battery, the status data of the target battery during the charging process can be obtained, and the charge consistency monitoring result of the target battery can be determined through the status data of the target battery during the charging process; the charging process may represent a certain charging process of the target battery. For example, the charging process may be the most recent charging process of the target battery from the current moment.

[0190] In one embodiment, since the battery management system collects the status data of the target battery in real time during the charging process of the target battery, the computer device can obtain the status data of the target battery during the charging process from the battery management system corresponding to the target battery.

[0191] Optionally, after the battery management system collects the status data of the target battery during the charging process, it can upload the status data of the target battery during the charging process to the cloud, and the computer device can obtain the status data of the target battery during the charging process from the cloud.

[0192] S202, perform charge condition compensation on the status data through a preset working condition compensation model to obtain compensated status data.

[0193] The status data collected by the target battery during the charging process is not accurate enough due to the mutual influence of battery internal resistance, temperature, voltage, current, etc. Therefore, the status data can be compensated for the working condition, and the status data of the charging working condition can be compensated to the standard state to obtain compensated status data. Among them, the compensated status data may be the status data after working condition compensation.

[0194] Specifically, the status data can be input into a pre-trained working condition compensation model, and the working condition compensation analysis is performed on the status data through the working condition compensation model to obtain the compensated status data output by the working condition compensation model.

[0195] Optionally, it can also be to input the status data into a pre-trained working condition compensation model, perform working condition compensation analysis on the status data through the working condition compensation model to obtain the status data compensation amount output by the working condition compensation model, and then compensate the status data according to the status data compensation amount to obtain the compensated status data; among them, the method of compensating the status data according to the status data compensation amount can be to determine the difference between the status data and the status data compensation amount as the compensated status data; it can also be to determine the sum of the status data and the status data compensation amount as the compensated status data.

[0196] Among them, the working condition compensation model can be pre-constructed through historical data, and the construction process of the working condition compensation model can be: obtaining the historical collected status data of the sample battery in multiple charging processes and the corresponding historical standard status data, and then training the initial working condition compensation model through the historical collected status data of multiple charging processes and the corresponding historical standard status data until the initial working condition compensation model meets the preset iterative convergence condition, then determining the initial working condition compensation model that meets the iterative convergence condition as the working condition compensation model.

[0197] It should be noted that the network architecture adopted in the construction of the working condition compensation model in the embodiments of the present application is not limited.

[0198] S203. Determine the power consistency monitoring result of the target battery at the current moment according to the compensated status data.

[0199] The power consistency monitoring result of the target battery at the current moment can represent the power consistency degree of the target battery at the current moment, and can include that the power consistency of the target battery at the current moment is good or poor, etc.

[0200] The above-mentioned compensated status data can include the status values of the target battery during the charging process. The compensated status data can be compared with the preset status threshold range, and the power consistency monitoring result of the target battery at the current moment can be determined according to the comparison result; for example, if all the compensated status data are within the status threshold range, it means that the power consistency of the target battery is good, and it can be determined that the power consistency monitoring result of the target battery at the current moment is good power consistency; if there is data in the compensated status data that is not within the status threshold range, it means that the power consistency of the target battery is poor, and it can be determined that the power consistency monitoring result of the target battery at the current moment is poor power consistency.

[0201] The degree of good or poor charge consistency of the target battery can also be further determined according to a preset multi-level state threshold range.

[0202] Optionally, the charge consistency monitoring result of the target battery at the current moment can also be determined according to a pre-trained consistency monitoring model; specifically, the compensated state data is input into the consistency monitoring model, and the consistency monitoring model analyzes the compensated state data to obtain the consistency monitoring result output by the consistency monitoring model, and the consistency monitoring result output by the consistency monitoring model is determined as the charge consistency monitoring result of the target battery at the current moment.

[0203] The charge consistency determination method provided by the embodiments of the present application obtains the state data of the target battery during the charging process, compensates the state data for the charging working condition through a preset working condition compensation model to obtain compensated state data, and then determines the charge consistency monitoring result of the target battery at the current moment according to the compensated state data. In this method, since the working condition compensation model is a model for compensating the state data of the battery under the charging working condition and can compensate the state data under the charging working condition to the standard state, therefore, compensating the state data of the target battery through a pre-trained working condition compensation model can reduce the influence brought by collecting state data during the charging process of the target battery, so that the obtained compensated state data is the actual state data of the target battery, and calculating the charge consistency monitoring result of the target battery at the current moment with the actual state data improves the accuracy of the charge consistency monitoring result of the target battery; moreover, determining the charge consistency monitoring result of the target battery through the state data during the charging process of the target battery is equivalent to calculating the charge consistency of the battery using the dynamic characteristics of the battery, which can greatly increase the calculation frequency of the charge consistency of the battery, thereby realizing accurate monitoring of the charge consistency of the target battery.

[0204] During the charging process of the target battery, some state data of the target battery are greatly affected by the charging working condition. Therefore, when calculating the charge consistency monitoring result of the target battery, the data to be compensated during the charging process of the target battery can be compensated to obtain more accurate compensated state data. The following details this through an embodiment. In an exemplary embodiment, as Figure 3 shown, compensating the state data for the charging working condition through a preset working condition compensation model to obtain compensated state data includes the following steps:

[0205] S301, obtain the data to be compensated during the charging process of the target battery from the state data.

[0206] The status data can be directly determined as the data to be compensated during the charging process of the target battery. For example, the status data can include data such as the charging rate, temperature, voltage, current, and cumulative charge of the battery. The charging rate data, temperature data, voltage data, current data, and cumulative charge data of the battery can all be directly determined as the data to be compensated; alternatively, some of the status data can be determined as the data to be compensated. For example, the charging rate data, voltage data, and current data can be determined as the data to be compensated.

[0207] Among them, the charging rate data represents the charging rate of the target battery at each moment during the charging process; the current data represents the current passing through the target battery at each moment during the charging process; the temperature data represents the temperature of the target battery at each moment during the charging process; the cumulative charge data of the battery can represent the cumulative charge of the target battery at each moment during the charging process; the voltage data can include the overall voltage of the target battery at each moment during the charging process, the individual voltages of each battery cell in the target battery, the average voltage of all battery cells in the target battery, and so on.

[0208] S302. Perform charging condition compensation on the data to be compensated through the working condition compensation model to obtain compensated voltage data.

[0209] The data to be compensated can be directly input into the working condition compensation model, and the working condition compensation analysis is performed on the data to be compensated through the working condition compensation model to obtain the compensated data output by the working condition compensation model. The compensated data output by the working condition compensation model is determined as the compensated data corresponding to the data to be compensated.

[0210] Optionally, it can also be to input the data to be compensated into the working condition compensation model, perform working condition compensation analysis on the data to be compensated through the working condition compensation model to obtain the compensation amount output by the working condition compensation model, and then compensate the data to be compensated according to the compensation amount to obtain the compensated data; among them, the sum of the compensation amount and the data to be compensated can be determined as the compensated data, or the difference between the data to be compensated and the compensation amount can be determined as the compensated data.

[0211] S303. Determine the compensated data and other status data in the status data as the compensated status data.

[0212] For example, taking the data to be compensated including voltage data as an example, the status data can also include data such as the charging rate, current, temperature, and cumulative charge of the battery. Therefore, the compensated voltage data, charging rate data, current data, temperature data, and cumulative charge data of the battery are directly determined as the compensated status data.

[0213] It should be noted that in this embodiment, the state data including voltage data, charge rate data, current data, temperature data, and cumulative battery charge data are only examples, and the specific data included in the state data is not limited and can be determined according to actual calculation requirements.

[0214] The battery charge consistency determination method provided by the embodiment of the present application obtains the voltage data of the target battery during the charging process from the state data, compensates the data to be compensated for the charging condition through the working condition compensation model to obtain the compensated data, and then determines the compensated data and other state data in the state data as the compensated state data. In this method, compensating the data to be compensated for the target battery during the charging process through the working condition compensation model can reduce the influence of the charging condition change of the target battery on the data to be compensated, making the compensated data more accurate, thereby further improving the accuracy of the battery consistency monitoring result of the target battery.

[0215] During the charging process of the target battery, the voltage is greatly affected by the charging condition. Therefore, when calculating the battery charge consistency monitoring result of the target battery, the voltage data in the state data can be compensated to obtain more accurate voltage data, that is, the data to be compensated includes voltage data; correspondingly, the data to be compensated includes the first voltage and the second voltage at each moment during the charging process of the target battery.

[0216] In one embodiment, the target battery includes at least two battery cells, and the first voltage and the second voltage at each moment are determined from the voltages of at least two battery cells at each moment.

[0217] During the charging process of the target battery, the voltage of each battery cell can be collected at each moment. For any moment, the maximum voltage among the voltages of all battery cells can be obtained, and the minimum voltage among the voltages of all battery cells can be obtained. The maximum voltage at each moment is used as the first voltage at each moment, and the minimum voltage at each moment is used as the second voltage at each moment; that is, the first voltage represents the maximum voltage at each moment, and the second voltage represents the minimum voltage at each moment.

[0218] The battery charge consistency determination method provided by the embodiment of the present application, the target battery includes at least two battery cells, and the maximum voltage and the minimum voltage at each moment are determined from the voltages of at least two battery cells at each moment. In this method, the state data of the target battery includes the maximum voltage and the minimum voltage among all battery cells, and the battery charge consistency of the target battery is calculated based on the maximum voltage and the minimum voltage in the target battery. In this way, the battery charge consistency of the target battery calculated as a whole can represent the battery charge consistency range among the battery cells in the target battery, thereby making the battery charge consistency monitoring result of the target battery at the current moment more reliable.

[0219] In an exemplary embodiment, such asFigure 4 As shown, the charging condition compensation is performed on the voltage data through the working condition compensation model to obtain the compensated voltage data, including the following steps:

[0220] S401. For any moment, obtain the first voltage compensation amount and the second voltage compensation amount at the moment through the working condition compensation model.

[0221] Among them, the first voltage compensation amount is the voltage compensation amount corresponding to the first voltage, and the second voltage compensation amount is the voltage compensation amount corresponding to the second voltage.

[0222] For any moment, the state data at the moment can be input into the working condition compensation model to obtain the first voltage compensation amount and the second voltage compensation amount output by the working condition compensation model at the moment.

[0223] Optionally, the first voltage compensation amount and the second voltage compensation amount at the corresponding moment can also be determined by the current in the state data; for example, the working condition compensation model can include a first voltage compensation model and a second voltage compensation model; the state data includes the current of the target battery at each moment during the charging process; then in an exemplary embodiment, as Figure 5 shown, obtaining the first voltage compensation amount and the second voltage compensation amount at the moment through the working condition compensation model includes the following steps:

[0224] S501. Input the current at the moment into the first voltage compensation model to obtain the first voltage compensation amount.

[0225] The first voltage compensation amount can represent the voltage amount to be compensated for the first voltage at the corresponding moment.

[0226] For any moment, input the current at the moment into the first voltage compensation model, analyze the current through the first voltage compensation model, obtain the voltage compensation amount output by the first voltage compensation model, and then determine the voltage compensation amount output by the first voltage compensation model as the first voltage compensation amount at the moment.

[0227] S502. Input the current at the moment into the second voltage compensation model to obtain the second voltage compensation amount.

[0228] The second voltage compensation amount can represent the voltage amount to be compensated for the second voltage at the corresponding moment.

[0229] For any moment, input the current at the moment into the second voltage compensation model, analyze the current through the second voltage compensation model, obtain the voltage compensation amount output by the second voltage compensation model, and then determine the voltage compensation amount output by the second voltage compensation model as the second voltage compensation amount at the moment.

[0230] In the embodiment of the present application, the current at a moment is input into the first voltage compensation model to obtain the first voltage compensation amount, and the current at the moment is input into the second voltage compensation model to obtain the second voltage compensation amount. In this embodiment, since the current change will cause the voltage change, therefore, the corresponding voltage compensation amount can be determined according to the current, and the first voltage compensation amount and the second voltage compensation amount are determined through the pre-trained first voltage compensation model and the second voltage compensation model, so that the compensated voltage data is closer to the stable value, and the accuracy of the compensated voltage data is improved.

[0231] S402, superimpose the first voltage compensation amount and the first voltage to obtain the compensated first voltage; and superimpose the second voltage compensation amount and the second voltage to obtain the compensated second voltage.

[0232] Among them, the compensated first voltage can represent the voltage after the working condition compensation of the first voltage at the corresponding moment, and the compensated second voltage can represent the voltage after the working condition compensation of the second voltage at the corresponding moment.

[0233] Determine the sum of the first voltage compensation amount and the first voltage as the compensated first voltage at the corresponding moment, and determine the sum of the second voltage compensation amount and the second voltage as the compensated second voltage at the corresponding moment.

[0234] S403, determine the compensated first voltage and the compensated second voltage as the compensated voltage data at the moment.

[0235] Determine the compensated first voltage and the compensated second voltage as the compensated voltage data at the corresponding moment.

[0236] Based on the above method, the compensated voltage data at each moment during the charging process of the target battery can be determined, and the embodiments of the present application will not be elaborated here.

[0237] It should be noted that if the data to be compensated is other data, the compensation data corresponding to other data can also be determined according to the method of determining the compensated voltage data.

[0238] The method for determining the power consistency provided by the embodiments of the present application, where the voltage data includes the first voltage and the second voltage of the target battery at each moment during the charging process; for any moment, the first voltage compensation amount and the second voltage compensation amount at the moment are obtained through the working condition compensation model, and the first voltage compensation amount is superimposed on the first voltage to obtain the compensated first voltage, and the second voltage compensation amount is superimposed on the second voltage to obtain the compensated second voltage. Finally, the compensated first voltage and the compensated second voltage are determined as the compensated voltage data at the moment. In this method, during the charging process of the target battery, since the current of the target battery changes and the internal state of the battery changes, this will cause the voltage of the target battery not to be in a stable state, and the collected first voltage and second voltage are not in a standard state. Therefore, by obtaining the voltage compensation amounts corresponding to the first voltage and the second voltage for voltage compensation, the obtained compensated voltage data is closer to the steady-state value, thereby greatly improving the accuracy of the power consistency monitoring result of the target battery.

[0239] The above embodiments all illustrate how to determine the compensated state data through the working condition compensation model. The following will detail the acquisition process of the working condition compensation model through an embodiment. In an exemplary embodiment, as Figure 6 shown, the acquisition process of the working condition compensation model includes:

[0240] S601, obtain the historical state data of the historical battery during multiple historical charging processes.

[0241] The computer device can obtain the initial state data of the historical battery during multiple historical charging processes from the battery management system, or the computer device can directly obtain the initial state data of the historical battery during multiple historical charging processes from the cloud.

[0242] The computer device can directly determine the above-obtained initial state data as the historical state data of the historical battery during multiple historical charging processes.

[0243] Optionally, the computer device can also perform abnormal data cleaning on the above-obtained initial state data, and determine the initial state data after abnormal data cleaning as the historical state data of the historical battery during multiple historical charging processes.

[0244] Among them, abnormal data cleaning includes removing invalid values, removing outliers, removing duplicate values, and field unit mapping, etc.; the method of performing abnormal data cleaning on the initial state data can be to input the initial state data into a pre-trained abnormal data screening model, analyze the initial state data through the abnormal data screening model, and the abnormal data screening model outputs the initial state data after abnormal data cleaning.

[0245] S602. For any historical charging process, obtain the absolute value of the current difference at each adjacent moment during the historical charging process.

[0246] The absolute value of the current difference at each adjacent moment during each historical charging process can be obtained separately.

[0247] Specifically, for any historical charging process, the historical state data includes the current at each moment during the historical charging process. Therefore, the absolute value of the current difference at each adjacent moment, that is, the absolute value of the current difference, can be directly obtained.

[0248] S603. Determine the historical state data corresponding to the adjacent moments with the absolute value of the current difference greater than the preset threshold as the reference state data at each adjacent moment during the historical charging process.

[0249] Determine the historical state data corresponding to the adjacent moments with the absolute value of the current difference greater than the preset threshold as the reference state data at each adjacent moment during the historical charging process. That is, the reference state data is the historical state data corresponding to the adjacent moments with the absolute value of the current difference greater than the preset threshold.

[0250] For example, if the historical charging process successively includes moment A, moment B, and moment C, the absolute value of the current difference between moment A and moment B, and the absolute value of the current difference between moment B and moment C can be obtained; if the absolute value of the current difference between moment A and moment B is greater than the preset threshold, then the historical state data corresponding to moment A and moment B is determined as the reference state data.

[0251] In another exemplary embodiment, the reference state data at each adjacent moment during each historical charging process can also be obtained through a data selection model; specifically, for any historical charging process, the historical state data of the historical charging process can be input into the data selection model to obtain the reference state data at each adjacent moment during the historical charging process output by the data selection model.

[0252] S604. Construct a working condition compensation model according to the reference state data at each adjacent moment.

[0253] The reference state data at each adjacent moment can be used as a training set to train and obtain a working condition compensation model. Specifically, the initial working condition compensation model is trained through the reference state data at each adjacent moment until the initial working condition compensation model converges, and the converged initial working condition compensation model is determined as the working condition compensation model. Among them, the condition for the initial working condition compensation model to converge can be that the number of iterations of the initial working condition compensation model reaches the preset number of iterations, or the accuracy rate of the initial working condition compensation model reaches the preset accuracy rate threshold.

[0254] Optionally, in an exemplary embodiment, such as Figure 7As shown, according to the reference state data at adjacent moments, a working condition compensation model is constructed, including the following steps:

[0255] S701, perform feature extraction on the reference state data at adjacent moments to determine the historical charging characteristics at adjacent moments.

[0256] Among them, the historical charging characteristics at adjacent moments may include the first voltage difference, the second voltage difference, the current difference, the cumulative charged power, and so on.

[0257] Among them, the historical charging characteristics at adjacent moments can be calculated according to the reference state data at adjacent moments and the calculation formula of charging characteristics; for example, for any adjacent moment, the current difference can represent the difference between the current at the latter moment and the current at the former moment in adjacent moments, the first voltage difference can represent the difference between the first voltage at the latter moment and the first voltage at the former moment in adjacent moments, the second voltage difference can represent the difference between the second voltage at the latter moment and the second voltage at the former moment in adjacent moments, and the cumulative charged power can represent the power charged into the battery relative to the initial charging moment at the former moment or the latter moment in adjacent moments.

[0258] Optionally, the historical charging characteristics at adjacent moments can also be determined according to a preset feature extraction model; specifically, the reference state data at adjacent moments are input into the feature extraction model, and the feature extraction model performs feature analysis on the reference state data at adjacent moments to obtain the historical charging characteristics at adjacent moments output by the feature extraction model.

[0259] S702, construct a working condition compensation model according to each historical charging characteristic.

[0260] Each historical charging characteristic can be used as a training set to train a working condition compensation model. Specifically, the initial working condition compensation model is trained through each historical charging characteristic until the initial working condition compensation model converges, and the converged initial working condition compensation model is determined as the working condition compensation model. Among them, the condition for the initial working condition compensation model to converge can be that the number of iterations of the initial working condition compensation model reaches a preset iteration number threshold, or the accuracy of the initial working condition compensation model reaches a preset accuracy threshold.

[0261] In the embodiment of the present application, feature extraction is performed on the reference state data at adjacent moments to determine the historical charging characteristics at adjacent moments, and a working condition compensation model is constructed according to each historical charging characteristic. In this embodiment, since the working condition compensation model compensates the state data under the charging working condition, and the historical charging characteristics at adjacent moments can reflect the change of state data at adjacent moments, therefore, constructing a working condition compensation model through the historical charging characteristics of the reference state data at adjacent moments can enable the working condition compensation model to accurately compensate the state data.

[0262] The method for determining charge quantity consistency provided by the embodiments of the present application obtains historical state data of a historical battery during multiple historical charging processes. For any one of the historical charging processes, the absolute value of the current difference at adjacent moments during the historical charging process is obtained, and the historical state data corresponding to the adjacent moments with the absolute value of the current difference greater than a preset threshold is determined as the reference state data at each adjacent moment during the historical charging process. Finally, a working condition compensation model is constructed based on the reference state data at each adjacent moment. In this method, the historical state data corresponding to the adjacent moments with the absolute value of the current difference greater than the preset threshold is determined as the reference state data at each adjacent moment during the historical charging process. In this way, it is possible to obtain changes in other parameters in the state data caused by obvious current changes, making the reference state data more reliable, thereby improving the reliability and accuracy of the construction of the working condition compensation model; constructing the working condition compensation model based on the reference state data at each adjacent moment can improve the reliability and accuracy of the working condition compensation model for compensating the state data.

[0263] It is described with the historical charging characteristics including the first voltage difference, the second voltage difference, the current difference, the temperature, and the cumulative charged quantity; in an exemplary embodiment, as Figure 8 shown, constructing a working condition compensation model according to each historical charging characteristic includes the following steps:

[0264] S801, filter out outliers from the historical charging characteristics according to the corresponding relationship between each first voltage difference and each current difference, and the corresponding relationship between each second voltage difference and each current difference, to obtain filtered charging characteristics.

[0265] Among them, the first voltage difference can represent the difference between the maximum voltages at adjacent moments, and the second voltage difference can represent the difference between the minimum voltages at adjacent moments.

[0266] For all historical charging processes, according to the current switching characteristics and voltage change characteristics, a fitting model of the current difference and the voltage difference is established, the absolute value of the difference between the fitting value and the true value of the fitting model is calculated, and the outliers outside a certain quantile of the difference are filtered out to reduce the influence of working condition differences and data acquisition noise. The quantile can be taken as the 90% quantile.

[0267] Specifically, according to each first voltage difference and the corresponding current difference, a first voltage fitting model is constructed. The first voltage fitting model includes the corresponding relationship between the current difference and the first voltage difference. Then, through the first voltage fitting model, the fitted first voltage difference corresponding to the current difference at each adjacent moment is obtained, and the absolute value of the residual between each fitted first voltage difference and the corresponding true first voltage difference is obtained. Then, the charging characteristics outside the preset quantile of the absolute value of the residual are screened out as outliers, and the screened charging characteristics are determined as the first filtered charging characteristics.

[0268] According to each second voltage difference and the corresponding current difference, a second voltage fitting model is constructed. The second voltage fitting model includes the corresponding relationship between the current difference and the second voltage difference. Then, through the second voltage fitting model, the fitted second voltage difference corresponding to the current difference at each adjacent moment is obtained, and the absolute value of the residual between each fitted second voltage difference and the corresponding true second voltage difference is obtained. Then, the charging characteristics outside the preset quantile of the absolute value of the residual are screened out as outliers, and the screened charging characteristics are determined as the second filtered charging characteristics.

[0269] The first filtered charging characteristics and the second filtered charging characteristics can be determined as the filtered charging characteristics; or the repeated charging characteristics in the first filtered charging characteristics and the second filtered charging characteristics can be determined as the filtered charging characteristics.

[0270] S802, construct a multiple regression model according to the filtered charging characteristics.

[0271] It is possible to directly perform fitting according to the filtered charging characteristics in the above embodiments, and construct a multiple regression model with the filtered charging characteristics as unknowns through a machine learning algorithm or a linear regression algorithm.

[0272] Optionally, since the working condition compensation model includes a maximum working condition compensation model and a minimum working condition model, the multiple regression model can include a first multiple regression model and a second multiple regression model. Among them, if the above first filtered charging characteristics and second filtered charging characteristics are determined as the filtered charging characteristics, the first filtered charging characteristics can be used as the filtered charging characteristics to construct the first multiple regression model, and the second filtered charging characteristics can be used as the filtered charging characteristics to construct the second multiple regression model; if the repeated charging characteristics in the first filtered charging characteristics and the second filtered charging characteristics are determined as the filtered charging characteristics, the first multiple regression model and the second multiple regression model can be directly constructed according to the filtered charging characteristics.

[0273] Next, taking the filtered charging characteristics as an example, the construction of the first multiple regression model and the second multiple regression model is directly carried out. In an exemplary embodiment, as Figure 9 shown, constructing a multiple regression model according to the filtered charging characteristics includes the following steps:

[0274] S901. Construct a first multiple regression model based on the first voltage difference, current difference, temperature, and cumulative charge input in the filtered charging characteristics.

[0275] Establish a first multiple regression model based on the first voltage difference, current difference, temperature, and cumulative charge input according to a machine learning algorithm or a linear regression algorithm. That is, the first multiple regression model takes the first voltage difference, current difference, temperature, and cumulative charge input as unknowns.

[0276] Specifically, use the first voltage difference, current difference, temperature, and cumulative charge input in the filtered charging characteristics as a training set to train the initial multiple regression model until the initial multiple regression model converges, and then determine the converged initial multiple regression model as the first multiple regression model.

[0277] S902. Construct a second multiple regression model based on the second voltage difference, current difference, temperature, and cumulative charge input in the filtered charging characteristics.

[0278] Establish a second multiple regression model based on the second voltage difference, current difference, temperature, and cumulative charge input according to a machine learning algorithm or a linear regression algorithm. That is, the second multiple regression model takes the second voltage difference, current difference, temperature, and cumulative charge input as unknowns.

[0279] Specifically, use the second voltage difference, current difference, temperature, and cumulative charge input in the filtered charging characteristics as a training set to train the initial multiple regression model until the initial multiple regression model converges, and then determine the converged initial multiple regression model as the second multiple regression model.

[0280] In the embodiments of the present application, a first multiple regression model is constructed based on the first voltage difference, current difference, temperature, and cumulative charge input in the filtered charging characteristics, and a second multiple regression model is constructed based on the second voltage difference, current difference, temperature, and cumulative charge input in the filtered charging characteristics. In this embodiment, since it is necessary to perform working conditions on the first voltage and the second voltage in the state data, constructing the first multiple regression model corresponding to the first voltage and the second multiple regression model corresponding to the second voltage in a targeted manner makes the voltage compensation amount corresponding to the first voltage and the voltage compensation amount corresponding to the second voltage more accurate.

[0281] S803. Substitute the preset reference temperature, preset reference cumulative charge input, and preset reference current into the multiple regression model to obtain a working condition compensation model.

[0282] Substitute the values of the preset reference temperature, preset reference cumulative charge input, and preset reference current into the multiple regression model to obtain a working condition compensation model.

[0283] In the case where the multiple regression model includes a first multiple regression model and a second multiple regression model, the preset reference temperature, the preset reference cumulative charge amount, and the preset reference current are respectively substituted into the first multiple regression model and the second multiple regression model. The first multiple regression model with the reference values substituted is determined as the first voltage compensation model, and the second multiple regression model with the reference values substituted is determined as the second voltage compensation model.

[0284] Among them, the preset reference temperature can be 25 degrees, the preset reference cumulative charge amount can be 0, and the preset reference current can be 0.

[0285] In this way, the first voltage compensation model and the second voltage compensation model include the corresponding relationship between the current and the voltage difference. In practical applications, the current can be input into the first voltage compensation model and the second voltage compensation model, and the voltage difference output by the first voltage compensation model is determined as the first voltage compensation amount at the current moment. The voltage difference output by the second voltage compensation model is determined as the second voltage compensation amount at the current moment.

[0286] The method for determining the charge consistency provided by the embodiments of the present application, the historical charging characteristics include the first voltage difference, the second voltage difference, the current difference, the temperature, and the cumulative charge amount; according to the corresponding relationship between each first voltage difference and each current difference, and the corresponding relationship between each second voltage difference and each current difference, the historical charging characteristics are filtered for outliers to obtain the filtered charging characteristics, and a multiple regression model is constructed based on the filtered charging characteristics. Then, the preset reference temperature, the preset reference cumulative charge amount, and the preset reference current are substituted into the multiple regression model to obtain the working condition compensation model. In this method, since the working condition compensation model compensates the voltage under the charging working condition, during the battery charging process, the current, temperature, and SOC of the battery will all affect the voltage. Therefore, a multiple regression model is constructed through the current difference, voltage difference, temperature, and cumulative charge amount, and the reference current, reference temperature, and reference cumulative charge amount in the preset reference state are substituted into the multiple regression model, so that the finally constructed working condition compensation model can accurately obtain the voltage compensation amount required to reach the standard state, further improving the accuracy of the charge consistency monitoring result of the target battery; and before constructing the multiple regression model, the historical charging characteristics are filtered for outliers, reducing the data deviation of the abnormal points and improving the accuracy of the working condition compensation model.

[0287] The above embodiments all illustrate how to obtain the compensation state data. The following uses an embodiment to illustrate how to determine the charge consistency monitoring result of the target battery at the current moment according to the compensation state data. In an exemplary embodiment, as Figure 10As shown, determining the power consistency monitoring result of the target battery at the current moment according to the compensation status data includes the following steps:

[0288] S1001, in the case where the compensation voltage data in the compensation status data exceeds the platform voltage range, determine the compensation status data within the platform voltage range as the target status data.

[0289] Among them, the platform voltage range can represent the range of battery fluctuations of the target battery during charging; when the compensation voltage data in the compensation status data exceeds the platform voltage range, further obtain the compensation status data within the platform voltage range in the compensation status data.

[0290] In the compensation status data, for any moment, when the compensation first voltage at this moment is less than or equal to the maximum value of the platform voltage range and the compensation second voltage at this moment is greater than or equal to the minimum value of the platform voltage range, determine that the compensation status data corresponding to this moment is within the platform voltage range.

[0291] Based on this, obtain the compensation status data within the platform voltage range in the compensation status data, and determine the compensation status data within the platform voltage range as the target status data, that is, the compensation first voltage and the compensation second voltage at each moment in the target status data are both within the platform voltage range.

[0292] S1002, in the case where the compensation voltage data in the compensation status data does not exceed the platform voltage range, determine the target status data according to the compensation status data and the historical status data.

[0293] Since the compensation voltage data of the target battery does not span the platform area, therefore, historical status data that does not span the platform area and is close to the charging process time can be aggregated to integrate the target status data that meets the platform voltage range conditions.

[0294] Among them, the historical status data includes the status data corresponding to the historical compensation voltage data not exceeding the platform voltage range within the historical time period. The historical status data can include the status data of multiple historical charging processes, and the historical time period can be a preset duration from the charging process; for example, the historical status data is the status data corresponding to the historical compensation voltage data not exceeding the platform voltage range within 5 days.

[0295] It should be noted that the voltage data in the historical status data can be the compensation voltage data obtained by performing working condition compensation on the initial voltage data through the working condition compensation model.

[0296] In one embodiment, the integrated model can be used to analyze the compensation status data and the historical status data to obtain the target status data. Specifically, the compensation status data and the historical status data are input into the integrated model, and the integrated model analyzes the compensation status data and the historical status data to obtain the target status data.

[0297] S1003. Determine the power consistency monitoring result of the target battery at the current moment according to the target status data.

[0298] The target status data can be compared with the preset status threshold range, and the power consistency monitoring result of the target battery at the current moment can be determined according to the comparison result. For example, if all the target status data are within the status threshold range, it means that the power consistency of the target battery is good, and the power consistency monitoring result of the target battery at the current moment can be determined as good power consistency. If there is data in the target status data that is not within the status threshold range, it means that the power consistency of the target battery is poor, and the power consistency monitoring result of the target battery at the current moment can be determined as poor power consistency.

[0299] The degree of good or poor power consistency of the target battery can also be further determined according to the preset multi-level status threshold range.

[0300] Optionally, the power consistency monitoring result of the target battery at the current moment can also be determined according to the pre-trained consistency monitoring model. Specifically, the target status data is input into the consistency monitoring model, and the consistency monitoring model analyzes the target status data to obtain the consistency monitoring result output by the consistency monitoring model, and the consistency monitoring result output by the consistency monitoring model is determined as the power consistency monitoring result of the target battery at the current moment.

[0301] In the power consistency determination method provided by the embodiment of the present application, when the compensation voltage data in the compensation status data exceeds the platform voltage range, the compensation status data within the platform voltage range is determined as the target status data. When the compensation voltage data in the compensation status data does not exceed the platform voltage range, the target status data is determined according to the compensation status data and the historical status data. The historical status data includes the status data corresponding to the historical compensation voltage data that did not exceed the platform voltage range within the historical time period. Finally, the power consistency monitoring result of the target battery at the current moment is determined according to the target status data. In this method, when obtaining the target status data from the compensation status data, it is first judged whether the compensation voltage data of the compensation status data exceeds the platform voltage range, and according to different situations, the compensation status data is specifically analyzed, so that the target status data is more comprehensive and accurate, and the accuracy of the power consistency monitoring result is improved.

[0302] Since the target state data is determined from the compensation state data according to the platform voltage range, by determining the power consistency monitoring result based on the compensation state data that meets the voltage condition, the power consistency monitoring result can be made more accurate and effective.

[0303] Before obtaining the target state data from the compensation state data, it can be first determined whether the compensation state data spans the platform voltage range, that is, whether the compensation state data exceeds the platform voltage range, and the target state data is determined from the compensation state data according to whether the compensation state data exceeds the platform voltage range.

[0304] The following illustrates how to determine whether the compensation state data exceeds the platform voltage range through an embodiment. In an exemplary embodiment, the compensation state data includes the compensated first voltage and the compensated second voltage at multiple moments during the charging process of the target battery; this embodiment includes: obtaining the maximum value of the compensated first voltages at multiple moments and the minimum value of the compensated second voltages at multiple moments; in the case where the maximum value is greater than or equal to the maximum value of the platform voltage range and the minimum value is less than or equal to the minimum value of the platform voltage range, it is determined that the compensated voltage data in the compensation state data exceeds the platform voltage range.

[0305] Since the compensation state data includes the compensated first voltage and the compensated second voltage at multiple moments, therefore, the maximum value of all the compensated first voltages in the compensation state data and the minimum value of all the compensated second voltages can be obtained.

[0306] If the maximum value of the compensated first voltage in the compensation state data is greater than or equal to the maximum value of the platform voltage range and the minimum value of the compensated second voltage in the compensation state data is less than or equal to the minimum value of the platform voltage range, it means that the compensation state data spans the platform area, so it can be determined that the compensated voltage data in the compensation state data exceeds the platform voltage range.

[0307] Otherwise, in the case where the maximum value is less than the maximum value of the platform voltage range and / or the minimum value is greater than the minimum value of the platform voltage range, it means that the compensation state data does not span the platform area, and it is determined that the compensated voltage data in the compensation state data does not exceed the platform voltage range.

[0308] The method for determining the power consistency provided by the embodiment of the present application obtains the maximum value of the compensated first voltage at multiple moments and the minimum value of the compensated second voltage at multiple moments; in the case where the maximum value is greater than or equal to the maximum value of the platform voltage range and the minimum value is less than or equal to the minimum value of the platform voltage range, it is determined that the compensated voltage data in the compensation state data exceeds the platform voltage range. In this method, it is determined whether the corresponding compensation state data exceeds the platform voltage range by whether the compensated voltage data spans the platform area. In this way, the target state data for calculating the power consistency monitoring result is made more comprehensive, thereby improving the power consistency monitoring result of the target battery.

[0309] The following uses an embodiment to elaborate in detail on the method for determining the target state data when the compensation state data does not exceed the platform voltage range. In an exemplary embodiment, as Figure 11 shown, determining the target state data according to the compensation state data and the historical state data includes the following steps:

[0310] S1101, Merge the compensation state data and the historical state data to obtain the merged state data.

[0311] Merge the compensation state data and the historical state data to obtain the merged state data. Among them, the merged state data includes the compensation state data and the historical state data.

[0312] S1102, According to the repeated state data corresponding to the compensation state data and the historical state data at each power, determine the uniquely corresponding state data at each power as the repeated state data with the highest priority.

[0313] Among them, there may be aggregated charging segments in the compensation state data and the historical state data. For example, the compensation state data includes the state data from a battery power of 50% SOC to 90% SOC, and the historical state data includes the state data from a battery power of 20% SOC to 60% SOC. It can be determined that there is repeated state data in the compensation state data and the historical state data, that is, the repeated state data is the state data corresponding to 50% SOC to 60% SOC.

[0314] Therefore, it is necessary to eliminate the repeated state data at each power and retain the uniquely corresponding state data at each power in the repeated state data.

[0315] The unique state data at each power level can be determined by means of priority. For example, when the priority of the compensation state data is higher than that of the historical state data, the repeated state data corresponding to each power level in the compensation state data is determined as the unique state data corresponding to each power level. When the priority of the compensation state data is lower than that of the historical state data, the repeated state data corresponding to each power level in the historical state data is determined as the unique state data corresponding to each power level.

[0316] For example, the repeated state data is the state data corresponding to 50% SOC to 60% SOC. When the priority of the historical state data is higher than that of the compensation state data, the state data corresponding to 50% SOC to 60% SOC in the historical state data is determined as the unique state data corresponding to 50% SOC to 60% SOC.

[0317] Optionally, if the historical state data may include the state data of multiple charging processes, the repeated state data with the highest priority among the compensation state data and each historical state data can be determined as the unique state data at each power level.

[0318] Among them, the priority of the compensation state data and the historical state data can be preset; or the priority of the compensation state data and the historical state data can be set according to the principle that the one closest to the current moment has the highest priority.

[0319] S1103. Screen the merged state data according to each unique state data to obtain the integrated state data.

[0320] Since the merged state data is the state data directly merged from the compensation state data and the historical state data, the merged state data includes all the state data of the compensation state data and the historical state data, including the repeated state data.

[0321] Therefore, other repeated state data corresponding to each unique state data in the merged state data can be deleted, and the merged state data after deletion is the integrated state data.

[0322] For example, the compensation state data includes the state data from the battery power of 50% SOC to 90% SOC, the historical state data includes the state data from the battery power of 20% SOC to 60% SOC, the repeated state data is the state data corresponding to 50% SOC to 60% SOC, and the unique state data is the historical state data corresponding to 50% SOC to 60% SOC. Therefore, the compensation state data corresponding to 50% SOC to 60% SOC in the merged state data can be deleted, and the merged state data after deletion is the integrated state data.

[0323] S1104. Obtain the target state data from the integrated state data according to the platform voltage range.

[0324] After obtaining the integrated status data, the target status data can be obtained from the integrated status data according to the platform voltage range.

[0325] Specifically, continue to determine whether the compensation voltage data in the integrated status data exceeds the platform voltage range. In the case where the compensation voltage data in the integrated status data exceeds the platform voltage range, obtain the compensation status data within the platform voltage range from the integrated status data, and determine the compensation status data within the platform voltage range as the target status data.

[0326] If the compensation voltage data in the integrated status data does not exceed the platform voltage range, continue to expand the time window, and integrate the historical status data and the integrated status data in the time window to obtain the integrated status data until the target status data is obtained from the integrated status data.

[0327] It should be noted that the method for determining whether the compensation voltage data in the integrated status data exceeds the platform voltage range and obtaining the target status data from the integrated status data in the embodiments of the present application is the same as the implementation method for determining whether the compensation voltage data in the compensation status data exceeds the platform voltage range and obtaining the target status data from the compensation status data in the above embodiments, and will not be elaborated herein in the embodiments of the present application.

[0328] The method for determining the charge consistency provided by the embodiments of the present application combines the compensation status data and the historical status data to obtain the combined status data, and determines the duplicate status data with the highest priority as the unique status data corresponding to each charge according to the duplicate status data corresponding to the compensation status data and the historical status data at each charge. Then, the combined status data is screened according to the unique status data to obtain the integrated status data. Finally, the target status data is obtained from the integrated status data according to the platform voltage range. In this method, when combining the compensation status data and the historical status data, the duplicate status data is screened by priority, so that the integrated status data is more complete and does not include duplicate status data, improving the integrity and effectiveness of the integrated status data, enabling the charge consistency monitoring result of the target battery to be determined through the integrated status data, and enhancing the computing opportunity of the charge consistency.

[0329] It is also possible to determine the priority of the compensation status data and the historical status data according to the charge charged into the platform area. The following describes this in detail through an embodiment. In an exemplary embodiment, as Figure 12 shown, before determining the duplicate status data with the highest priority as the unique status data corresponding to each charge, this embodiment includes the following steps:

[0330] S1201, obtain the first amount of electricity charged within the platform voltage range for the compensation status data, and the second amount of electricity charged within the platform voltage range for the historical status data.

[0331] Among them, the first amount of electricity can represent the amount of electricity charged within the platform voltage range for the compensation status data, and the second amount of electricity can represent the amount of electricity charged within the platform voltage range for the historical status data.

[0332] Obtain the compensation status data within the platform voltage range for the compensation status data, and determine the first amount of electricity charged within the platform area based on the compensation status data within the platform voltage range.

[0333] Obtain the historical status data within the platform voltage range for the historical status data, and determine the second amount of electricity charged within the platform area based on the historical status data within the platform voltage range.

[0334] The status data also includes the SOC at each moment. Therefore, the first amount of electricity charged within the platform voltage range for the compensation status data can be determined according to the SOC change of the compensation status data within the platform voltage range, and the second amount of electricity charged within the platform voltage range for the historical status data can be determined according to the SOC change of the historical status data within the platform voltage range. The first amount of electricity and the second amount of electricity can be represented in the form of the SOC of the battery.

[0335] S1202, determine the priority of each repeated status data according to the first amount of electricity and the second amount of electricity.

[0336] According to the principle that the higher the amount of electricity charged within the platform voltage range, the higher the priority, determine the priority of the compensation status data and the historical status data according to the first amount of electricity and the second amount of electricity; for example, if the first amount of electricity is greater than the second amount of electricity, the priority of the compensation status data is greater than the priority of the repeated status data.

[0337] For any repeated status data, determine the priority of the status data to which the repeated status data belongs as the priority of the repeated status data; for example, if the repeated status data belongs to the compensation status data, determine the priority of the compensation status data as the priority of the repeated status data; if the repeated status data belongs to the historical status data, determine the priority of the historical status data as the priority of the repeated status data.

[0338] The method for determining the charge consistency provided by the embodiments of the present application obtains the first charge charged within the platform voltage range for the compensation status data and the second charge charged within the platform voltage range for the historical status data, and then determines the priority levels of the respective repeated status data according to the first charge and the second charge. In this method, the priority levels of the compensation status data and the historical status data are evaluated based on the magnitude of the charge charged within the platform voltage range. In this way, when selecting the repeated status data, the most complete status information can be retained, making the integrated status data more reliable.

[0339] The following uses an embodiment to illustrate how to determine the charge consistency monitoring result of the target battery at the current moment based on the target status data. In an exemplary embodiment, the target status data includes the compensated first voltage and the compensated second voltage at multiple moments; as Figure 13 shown, determining the charge consistency monitoring result of the target battery at the current moment according to the target status data includes the following steps:

[0340] S1301. Determine the mean voltage difference and the variance of the voltage difference of the target status data according to the compensated first voltage and the compensated second voltage at each moment.

[0341] First, respectively obtain the voltage differences between the compensated first voltage and the compensated second voltage at each moment, and then determine the mean voltage difference of the target status data according to the voltage differences at each moment; according to the voltage differences and the mean voltage difference at each moment, determine the variance of the voltage difference of the target status data.

[0342] For any moment, determine the difference between the compensated first voltage and the compensated second voltage at one moment as the voltage difference, so as to obtain the voltage differences between the compensated first voltage and the compensated second voltage at each moment; then use the mean calculation method in statistics, add up all the voltage differences and divide by the number of voltage differences to obtain the average value, and determine this average value as the mean voltage difference of the target status data.

[0343] Variance is a statistical index for measuring the degree of data dispersion. The variance calculation method in statistics can be used to calculate the square of the difference between each voltage difference and the mean voltage difference to obtain multiple squared values, and then determine the value obtained by adding up all the squared values and dividing by the number of mean voltage differences as the variance of the voltage difference of the target status data.

[0344] S1302. Determine the charging characteristics of the target battery according to the mean voltage difference and the variance of the voltage difference.

[0345] The obtained mean voltage difference and variance of the voltage difference above can be determined as the charging characteristics of the target battery.

[0346] It should be noted that the charging characteristics of the target battery include, but are not limited to, the mean voltage difference and the variance of the voltage difference, and may also include the average charging rate, the variance of the charging rate, the average charging temperature, etc.

[0347] The target state data includes the charging rate and temperature at multiple moments. Therefore, the average value of the charging rates at multiple moments can be determined as the average charging rate of the target battery, the variance of the charging rates at multiple moments can be determined as the variance of the charging rate of the target battery, and the average value of the temperatures at multiple moments can be determined as the average charging temperature of the target battery.

[0348] S1303. Input the charging characteristics into the state classification model to obtain the power consistency monitoring result of the target battery at the current moment.

[0349] Input the charging characteristics of the target battery during the charging process into the state classification model. Through the analysis of the charging characteristics by the state classification model, the power consistency monitoring result of the target battery at the current moment output by the state classification model is obtained.

[0350] Among them, the power consistency monitoring result may include the power consistency rating of the target battery, the specific value corresponding to the power consistency, etc.

[0351] The power consistency determination method provided by the embodiments of the present application determines the mean voltage difference and the variance of the voltage difference of the target state data according to the compensated first voltage and the compensated second voltage at each moment, and determines the charging characteristics of the target battery according to the mean voltage difference and the variance of the voltage difference. Then, the charging characteristics are input into the state classification model to obtain the power consistency monitoring result of the target battery at the current moment. In this method, since the mean voltage difference and the variance of the voltage difference can reflect the health state of the battery, using the mean voltage difference and the variance of the voltage difference as the charging characteristics of the target terminal battery can achieve accurate monitoring of the power consistency of the target battery; moreover, by determining the power consistency monitoring result through the preset state classification model, directly inputting the charging characteristics into the state classification model can obtain the power consistency monitoring result of the target battery, which improves the speed of determining the power consistency monitoring result.

[0352] The state classification model includes a first-level classification model and a second-level classification model; in an exemplary embodiment, as Figure 14 shown, inputting the charging characteristics into the state classification model to obtain the power consistency monitoring result of the target battery at the current moment includes the following steps:

[0353] S1401. Input the charging characteristics into the first-level classification model to obtain a preliminary classification result.

[0354] Among them, the first-level classification model can perform a preliminary classification on the power consistency of the target battery, which can be a rough classification.

[0355] Therefore, the charging characteristics of the target battery can be input into the first-level classification model, and the first-level classification model analyzes the charging characteristics to obtain the initial classification result output by the first-level classification model.

[0356] Among them, the initial classification result can include the consistency label value of the power consistency of the target battery.

[0357] S1402. When the preliminary classification result is the first value, it is determined that the power consistency monitoring result at the current moment is good; good means that the power consistency is less than or equal to the first consistency threshold.

[0358] If the initial classification result is the first value, it means that the power consistency of the target battery is less than or equal to the first consistency threshold, that is, the power consistency of the target battery is less than or equal to the first consistency threshold, then it is determined that the power consistency monitoring result of the target battery at the current moment is good.

[0359] Among them, the first value can be 1, and the first consistency threshold can be 7% SOC.

[0360] S1403. When the preliminary classification result is the second value, the charging characteristics are input into the second-level classification model to obtain the power consistency monitoring result of the target battery at the current moment.

[0361] If the initial classification result is the second value, it means that the power consistency monitoring result of the target battery at the current moment is not good, and the power consistency is greater than the first consistency threshold; the second value can be 0. At this time, the second-level classification model can be used to further determine the degree of power consistency of the target battery at the current moment.

[0362] In an exemplary embodiment, as Figure 15 shown, inputting the charging characteristics into the second-level classification model to obtain the power consistency monitoring result of the target battery at the current moment includes the following steps:

[0363] S1501. Input the charging characteristics into the second-level classification model to obtain an advanced classification result.

[0364] Input the charging characteristics into the second-level classification model, and the second-level classification model analyzes the charging characteristics to obtain the advanced classification result output by the second-level classification model.

[0365] Among them, the advanced classification result includes the power consistency label value.

[0366] S1502. When the advanced classification result is the third value, it is determined that the power consistency monitoring result at the current moment is average, and average means that the power consistency is in the interval between the first consistency threshold and the second consistency threshold.

[0367] When the advanced classification result is the third value, it indicates that the power consistency of the target battery at the current moment is within the range of the first consistency threshold and the second consistency threshold, and it can be determined that the monitoring result of the power consistency of the target battery at the current moment is average.

[0368] Among them, the third value can be 1; the second consistency threshold is greater than the first consistency threshold, and the second consistency threshold can be 10% SOC.

[0369] S1503. When the advanced classification result is the fourth value, determine that the monitoring result of the power consistency at the current moment is poor; poor means that the power consistency is greater than or equal to the second consistency threshold.

[0370] When the advanced classification result is the fourth value, it indicates that the power consistency of the target battery at the current moment is greater than or equal to the second consistency threshold, and it can be determined that the monitoring result of the power consistency of the target battery at the current moment is poor.

[0371] In the embodiments of the present application, the charging characteristics are input into the secondary classification model to obtain the advanced classification result, and when the advanced classification result is the third value, it is determined that the monitoring result of the power consistency at the current moment is average, where average means that the power consistency is within the range of the first consistency threshold and the second consistency threshold; when the advanced classification result is the fourth value, it is determined that the monitoring result of the power consistency at the current moment is poor; poor means that the power consistency is greater than or equal to the second consistency threshold. In this method, when the power consistency of the target battery is not good, the secondary classification model is used to perform a fine classification of the average level and poor situation of the power consistency, improving the accuracy of the monitoring result of the power consistency of the target battery.

[0372] The power consistency determination method provided by the embodiments of the present application inputs the charging characteristics into the primary classification model to obtain the preliminary classification result; when the preliminary classification result is the first value, it is determined that the monitoring result of the power consistency at the current moment is good; good means that the power consistency is less than or equal to the first consistency threshold; when the preliminary classification result is the second value, the charging characteristics are input into the secondary classification model to obtain the monitoring result of the power consistency of the target battery at the current moment. In this method, by inputting the charging characteristics into the primary classification model, the preliminary classification result can be quickly obtained. When the initial classification result is the first value, the monitoring result of the power consistency of the target battery can be directly determined, reducing the workload of calculating the power consistency; when the initial classification result is the second value, the secondary classification model is continued to perform a detailed classification of the power consistency, refining the monitoring result of the power consistency and improving the accuracy of the monitoring result of the power consistency.

[0373] The above embodiments illustrate the determination of the power consistency monitoring result of the target battery at the current moment for the state classification model. Next, an embodiment is used to illustrate the acquisition process of the state classification model. In an exemplary embodiment, as Figure 16 shown, the acquisition process of the state classification model includes the following steps:

[0374] S1601, obtain the historical charging characteristics of the historical battery during multiple historical charging processes, and obtain the power consistency result of the historical battery at a preset power.

[0375] In an exemplary embodiment, as Figure 17 shown, obtaining the historical charging characteristics of the historical battery during multiple historical charging processes includes the following steps:

[0376] S1701, obtain the historical state data of the historical battery during at least one historical charging process.

[0377] The historical state data of the historical battery during at least one historical charging process can be obtained from the battery management system or the cloud of the historical battery.

[0378] Among them, the historical state data may include data such as the first voltage, second voltage, current, temperature, charging rate, and cumulative charged power at each moment during the historical charging process.

[0379] S1702, respectively perform working condition compensation on the historical state data of each historical charging process through the working condition compensation model to obtain compensated historical state data.

[0380] The historical state data includes the historical first voltage and historical second voltage at each moment. The historical first voltage at each moment represents the maximum voltage at each moment, and the historical second voltage at each moment represents the minimum voltage at each moment.

[0381] Therefore, the historical first voltage of each historical charging process can be compensated for the working condition through the first voltage compensation model in the working condition compensation model to obtain the historical compensated first voltage; the historical second voltage of each historical charging process can be compensated for the working condition through the second voltage compensation model in the working condition compensation model to obtain the historical compensated second voltage; the historical compensated first voltage, historical compensated second voltage, and other state data in the historical state data are determined as the compensated historical state data.

[0382] It should be noted that the specific implementation method for obtaining the compensated historical state data in this embodiment is the same as the specific implementation method for obtaining the compensated state data in the above embodiment, and will not be elaborated here.

[0383] S1703. Obtain the target historical state data of multiple target historical charging processes from the compensated historical state data according to a preset platform voltage range.

[0384] For any historical charging process, if the compensated voltage data in the compensated historical state data during the historical charging process exceeds the platform voltage range, then determine this historical charging process as a target historical charging process. In this way, multiple target historical charging processes can be obtained.

[0385] Then, for any target historical charging process, obtain the compensated historical state data within the platform voltage range, and determine the compensated historical state data within the platform voltage range as the target historical state data.

[0386] It should be noted that the method for judging whether the compensated voltage data in the compensated historical state data exceeds the platform voltage range in this embodiment is the same as the method for judging whether the compensated voltage in the compensated state data exceeds the platform voltage range in the above embodiment. The specific implementation method for obtaining the compensated historical state data within the platform voltage range in this embodiment is the same as the specific implementation method for obtaining the compensated state data within the platform voltage range in the above embodiment, and this application embodiment will not limit it here.

[0387] S1704. Extract features from each target historical state data to obtain the historical charging features of the historical battery during each historical charging process.

[0388] In this embodiment, the implementation method for extracting features from each target historical state data to obtain the historical charging features of the historical battery during each historical charging process is the same as the implementation method for extracting features from the target state data to obtain the charging features of the target battery in the above embodiment, and this embodiment will not elaborate here.

[0389] In the embodiment of the present application, obtain the historical state data of the historical battery during at least one historical charging process; through a working condition compensation model, perform working condition compensation on the historical state data of each historical charging process respectively to obtain compensated historical state data, and according to a preset platform voltage range, obtain the target historical state data of multiple target historical charging processes from the compensated historical state data, and then extract features from each target historical state data to obtain the historical charging features of the historical battery during each historical charging process. In this embodiment, by performing working condition compensation and screening the platform voltage range on the historical state data during multiple historical charging processes, the historical charging features that meet the conditions are obtained, improving the accuracy of the training data set of the state classification model, and thus making the obtained state classification model more accurate.

[0390] In an exemplary embodiment, as Figure 18 shown, obtaining the charge consistency result of the historical battery at a preset battery level includes the following steps:

[0391] S1801. For any preset battery level, obtain the maximum measured voltage and the minimum measured voltage of the historical battery at the preset battery level.

[0392] Among them, the preset battery level can be the battery level when the state of charge (SOC) is less than the preset SOC threshold; for example, the SOC threshold can be 30% SOC. Therefore, the preset battery level can be less than 30% SOC, and the preset battery level can correspond to the battery levels of the historical battery during multiple historical charging processes.

[0393] For any preset battery level, the maximum measured voltage and the minimum measured voltage of the historical battery at this preset battery level can be obtained.

[0394] Optionally, it can also be the case where, when the historical battery is at the preset battery level, the historical battery is charged or discharged with a small current for a preset duration, and the maximum measured voltage and the minimum measured voltage of the historical battery are obtained at the end of the duration.

[0395] Among them, the maximum measured voltage is the maximum voltage value among the voltages of all the battery cells in the historical battery, and the minimum measured voltage is the minimum voltage value among the voltages of all the battery cells in the historical battery.

[0396] S1802. Substitute the maximum measured voltage and the minimum measured voltage into the voltage - battery level relationship formula respectively to obtain the first battery level corresponding to the maximum measured voltage and the second battery level corresponding to the minimum measured voltage.

[0397] Among them, the voltage - battery level relationship formula can be obtained by mapping the SOC - OCV standard curve constructed according to the static voltage mapping method, and the voltage - battery level relationship formula can represent the relationship between voltage and SOC.

[0398] Therefore, the maximum measured voltage can be substituted into the voltage - battery level relationship formula to obtain the first battery level corresponding to the maximum measured voltage, and the minimum measured voltage can be substituted into the voltage - battery level relationship formula to obtain the second battery level corresponding to the minimum measured voltage.

[0399] S1803. Determine the battery level consistency result of the historical battery at the battery level by taking the difference between the first battery level and the second battery level.

[0400] Determine the battery level consistency result of the historical battery at this battery level by taking the difference between the first battery level and the second battery level. Based on the same method, the battery level consistency results of the historical battery at each preset battery level can be determined.

[0401] In the embodiment of the present application, for any preset power, the maximum measured voltage and the minimum measured voltage of the historical battery at the preset power are obtained, and the maximum measured voltage and the minimum measured voltage are respectively substituted into the voltage-power relationship formula to obtain the first power corresponding to the maximum measured voltage and the second power corresponding to the minimum measured voltage. Finally, the difference between the first power and the second power is determined as the power consistency result of the historical battery at the power. In this embodiment, the power consistency result at the preset power is obtained by the static voltage method, which improves the accuracy of the training data set of the state classification model, so that the obtained state classification model is more accurate.

[0402] S1602. Match each historical charging feature with each power consistency result according to time to obtain a training data set.

[0403] Match each historical charging feature with each power consistency result according to the principle of the closest time, so that the interval between the time of the charging process corresponding to the historical charging feature and the time of the voltage corresponding to the power consistency result is within a certain time range. For example, control the time corresponding to the historical charging feature and the power consistency result within two days.

[0404] For example, for any historical charging feature, obtain the power consistency result that is closest in charging time to the historical charging process corresponding to the historical charging feature and within the preset time interval, and then match the historical charging feature with the corresponding power consistency result to obtain multiple matching pairs, and determine the multiple matching pairs as the training data set.

[0405] S1603. Train the initial state classification model according to the training data set to obtain a state classification model.

[0406] Train the initial state classification model according to the training data set until the initial state classification model meets the preset convergence condition, and determine the initial state classification model that meets the preset convergence condition as the state classification model. The base model of the initial state classification model can be a decision tree.

[0407] The initial state classification model meets the preset convergence condition can be that the number of iterations of the initial state classification model reaches the preset number of iterations, or the recall rate and precision rate of the initial state classification model both reach the preset threshold.

[0408] Among them, the historical charging features in the training data set can include the mean pressure difference and the variance of the pressure difference, and a two-layer decision tree model (state classification model) is established through the mean pressure difference and the variance of the pressure difference.

[0409] The method for determining charge quantity consistency provided by the embodiments of the present application obtains the historical charging characteristics of a historical battery during multiple historical charging processes, and obtains the charge quantity consistency results of the historical battery at a preset charge quantity. Then, the historical charging characteristics and the charge quantity consistency results are matched according to time to obtain a training data set. Next, based on the training data set, an initial state classification model is trained to obtain a state classification model. In this method, the state classification model is trained with the historical charging characteristics and the charge quantity consistency results as the training data set, so that the charge quantity consistency of the battery can be predicted through the charging characteristics of the battery in the future. Moreover, when constructing the training data set, the historical charging characteristics and the charge quantity consistency results are matched according to time. In this way, the historical charging characteristics and the charge quantity consistency results are more corresponding, improving the accuracy of the training data set, and thus improving the accuracy of the state classification model.

[0410] The state classification model includes a primary classification model and a secondary classification model; in an exemplary embodiment, as Figure 19 shown, according to the training data set, training the initial state classification model to obtain the state classification model includes the following steps:

[0411] S1901, constructing a primary consistency label and a secondary consistency label according to the training data set.

[0412] According to the training data set, two sets of consistency labels can be constructed, including a primary consistency label and a secondary consistency label; each set of consistency labels includes historical charging characteristics corresponding to different charge quantity consistency threshold ranges.

[0413] Among them, the primary consistency label can include two labels. The historical charging characteristics with the charge quantity consistency result less than or equal to the third consistency threshold are used as label 1 (the first value), and the historical charging characteristics with the charge quantity consistency result greater than or equal to the first consistency threshold are used as label 0 (the second value); the primary consistency label is used to train whether the charge quantity consistency is good.

[0414] The secondary consistency label can include two labels. The historical charging characteristics with the charge quantity consistency result less than or equal to the second consistency threshold are used as label 1 (the third value), and the historical charging characteristics with the charge quantity consistency result greater than or equal to the fourth consistency threshold are used as label 0 (the fourth value); the secondary consistency label is used to train to distinguish whether the charge quantity consistency is very poor.

[0415] It should be noted that the consistency thresholds in each set of consistency labels can be adjusted according to actual business requirements. The first consistency threshold can be 7% SOC, the second consistency threshold can be 10% SOC, the third consistency threshold can be 5% SOC, and the fourth consistency threshold can be 12% SOC.

[0416] In each group of consistency tags, the difference in the power consistency threshold between tag 1 and tag 0 is mainly to control the tag error caused by the insufficient accuracy of the power consistency result calculated by the power-voltage relationship formula, so as to improve the classification accuracy.

[0417] S1902. Train the initial first-level state classification model according to the first-level consistency tags until the initial first-level state classification model converges to obtain the first-level classification model.

[0418] Train the initial first-level state classification model according to the first-level consistency tags until the initial first-level state classification model converges, and determine the converged initial first-level state classification model as the first-level classification model; among them, the condition for the initial first-level state classification model to converge can be that the recall rate of the initial first-level state classification model is greater than the preset comprehensiveness threshold, the precision rate is greater than the preset precision rate threshold, and the accuracy rate is greater than the preset accuracy rate threshold.

[0419] Among them, the initial first-level state classification model can be a decision tree, and the historical charging characteristics can include the mean pressure difference and the variance of the pressure difference. A two-layer decision tree model is established through the mean pressure difference and the variance of the pressure difference.

[0420] Using the first-level classification model obtained by training the initial first-level state classification model with 3190 data volumes for tag 1 and 450 data volumes for tag 0 in the first-level consistency tags, evaluate the first-level classification model through evaluation indicators using the training dataset and the test set, and the evaluation results are shown in Table 1.

[0421] Table 1

[0422]

[0423] S1903. Train the initial second-level state classification model according to the second-level consistency tags until the initial second-level state classification model converges to obtain the second-level classification model.

[0424] Train the initial second-level state classification model according to the second-level consistency tags until the initial second-level state classification model converges, and determine the converged initial second-level state classification model as the second-level classification model; among them, the condition for the initial second-level state classification model to converge can be that the recall rate of the initial second-level state classification model is greater than the preset comprehensiveness threshold and the precision rate is greater than the preset accuracy rate threshold.

[0425] Among them, the initial second-level state classification model can be a decision tree, and the historical charging characteristics can include the mean pressure difference and the variance of the pressure difference. A two-layer decision tree model is established through the mean pressure difference and the variance of the pressure difference.

[0426] The method for determining the power consistency provided by the embodiments of the present application constructs a first-level consistency label and a second-level consistency label based on a training data set, and trains an initial first-level state classification model according to the first-level consistency label until the initial first-level state classification model converges to obtain a first-level classification model, and trains an initial second-level state classification model according to the second-level consistency label until the initial second-level state classification model converges to obtain a second-level classification model. In this method, a first-level classification model and a second-level classification model are trained through two-level consistency labels. The first-level classification model is used to distinguish whether the power consistency is good, and the second-level classification model is used to distinguish whether the power consistency is very poor when the power consistency is not good. In this way, the fine classification of the power consistency of the target battery can be accurately achieved through the two-level classification models.

[0427] Taking the first consistency threshold as 7% SOC, the second consistency threshold as 10% SOC, the third consistency threshold as 5% SOC, and the fourth consistency threshold as 12% SOC as an example, the state of the power consistency of the target battery is divided by combining the first-level state classification model and the second-level state classification model; specifically, the charging characteristics of the target battery are input into the first-level state classification model. If the classification result output by the first-level state classification model is 1 (label 1), it means that the power consistency of the target battery is good and the power consistency <= 7% SOC; if the classification result output by the first-level state classification model is 0, it enters the second-level state classification model. At this time, if the classification result output by the second-level state classification model is 1, the power consistency of the target battery is at a general level and the power consistency is in the range of 7% to 10% SOC. If the classification result output by the second-level state classification model is 0, it means that the power consistency of the target battery is poor and the power consistency >= 10% SOC.

[0428] Based on the analysis of the correlation of charging characteristics, it is determined that under specific working conditions (the charging characteristics meet the preset conditions), the correlation between the power consistency and some charging characteristics exceeds 0.9. Therefore, when the target battery meets the specific working conditions, the monitoring result of the power consistency of the target battery at the current moment can be determined. The following will be described in detail through an embodiment. In an exemplary embodiment, as Figure 20 shown, to determine the monitoring result of the power consistency of the target battery at the current moment according to the target state data, the following steps are included:

[0429] S2001, extract features from the target state data to obtain the charging characteristics of the target battery; the charging characteristics include the bulging signal state and the average charging rate.

[0430] The bulging signal can be used to describe the phenomenon of sudden enhancement or weakening in the signal.

[0431] The target state data of the target battery can be input into the feature extraction model, and the feature extraction model can output the bulging signal state corresponding to the target battery. Among them, the bulging signal state can include that there is a bulging signal in the target state data, or there is no bulging signal in the target state data.

[0432] It should be noted that in addition to the bulging signal state and the average charging rate, the charging characteristics of the target battery can also include the current mean value, the charging rate variance, the average charging temperature, the pressure difference mean value, and the pressure difference variance, etc.

[0433] Among them, the acquisition methods of the current mean value, the average charging rate, the charging rate variance, the average charging temperature, the pressure difference mean value, the pressure difference variance and other characteristics of the target battery are the same as those in the above embodiments, and will not be elaborated herein.

[0434] S2002. When the bulging signal state is that there is a bulging signal and the average charging rate is less than the preset charging threshold, input the charging characteristics into the regression quantization model to obtain the predicted result of the power consistency of the target battery.

[0435] Among them, the predicted result of the power consistency can be the specific value of the power consistency.

[0436] When there is a bulging signal in the target state data and the average charging rate is less than the preset charging threshold, input the charging characteristics into the regression quantization model, and analyze the charging characteristics through the regression quantization model to obtain the predicted result of the power consistency of the target battery output by the regression quantization model; the charging characteristics can include the current mean value, the average charging temperature, the pressure difference mean value, and the pressure difference variance; that is, input the current mean value, the average charging temperature, the pressure difference mean value, and the pressure difference variance into the regression quantization model to obtain the predicted result of the power consistency of the target battery.

[0437] S2003. Determine the power consistency monitoring result of the target battery at the current moment according to the predicted result of the power consistency.

[0438] The predicted result of the power consistency can be determined as the power consistency monitoring result of the target battery at the current moment.

[0439] In an exemplary embodiment, as Figure 21 shown, determining the power consistency monitoring result of the target battery at the current moment according to the predicted result of the power consistency includes:

[0440] S2101. Obtain the prediction error range of the regression quantization model under the preset consistency confidence level.

[0441] Among them, the consistency confidence level can be the credibility of the prediction result of the regression quantization model. For example, the consistency confidence level can be 95%. The prediction error range of the regression quantization model under the preset consistency confidence level can represent the prediction error range with 95% credibility of the regression quantization model.

[0442] It should be noted that the prediction error range of the regression quantization model under the preset consistency confidence level can be pre-determined and stored in the database when constructing the regression quantization model. Therefore, the computer device can directly obtain the prediction error range of the regression quantization model under the preset consistency confidence level from the database.

[0443] S2102. Determine the confidence interval of the current power consistency of the target battery according to the power consistency prediction result and the prediction error range; different confidence intervals represent different degrees of power consistency.

[0444] The prediction error range includes an upper error limit and a lower error limit. Therefore, the sum of the upper error limit and the lower error limit of the prediction error range and the power consistency prediction result can be obtained respectively. The sum of the upper error limit of the prediction error range and the power consistency prediction result is determined as the upper confidence limit of the power consistency, and the sum of the lower error limit of the prediction error range and the power consistency prediction result is determined as the lower confidence limit of the power consistency; then the lower confidence limit of the power consistency is used as the minimum value of the confidence interval, and the upper confidence limit of the power consistency is used as the maximum value of the confidence interval. The confidence interval of the current power consistency of the target battery is determined according to the minimum value and the maximum value of the confidence interval.

[0445] For example, if the prediction error range of the regression quantization model under the preset consistency confidence level is [Q0, Q1], and the power consistency prediction result is Y, then the confidence interval of the current power consistency of the target battery is [Y + Q0, Y + Q1].

[0446] S2103. Determine the confidence interval of the current power consistency as the power consistency monitoring result of the target battery at the current moment.

[0447] Determine the confidence interval of the current power consistency as the power consistency monitoring result of the target battery at the current moment; that is, if the confidence interval of the current power consistency is [Y + Q0, Y + Q1], then the power consistency monitoring result of the target battery at the current moment is [Y + Q0, Y + Q1].

[0448] In the embodiment of the present application, the prediction error range of the regression quantization model at a preset consistency confidence level is obtained, and the confidence interval of the current charge consistency of the target battery is determined according to the charge consistency prediction result and the prediction error range; wherein, different confidence intervals represent different degrees of charge consistency; then the confidence interval of the current charge consistency is determined as the charge consistency monitoring result of the target battery at the current moment. In this embodiment, the confidence interval where the charge consistency of the target battery is located is determined through the prediction error range of the regression quantization model at the consistency confidence level, so as to determine the distribution interval of the charge consistency of the target battery at the preset consistency confidence level, effectively measuring the accuracy of the charge consistency monitoring result of the target battery at the current moment.

[0449] The charge consistency determination method provided by the embodiment of the present application extracts features from the target state data to obtain the charging features of the target battery; the charging features include the bulging signal state and the average charging rate; in the case where the bulging signal state is that there is a bulging signal and the average charging rate is less than the preset charging threshold, the charging features are input into the regression quantization model to obtain the charge consistency prediction result of the target battery; and the charge consistency monitoring result of the target battery at the current moment is determined according to the charge consistency prediction result. In this method, under specific working conditions of the battery, the charge consistency of the battery has a strong correlation with the charging features of the battery. Therefore, in the case where the bulging signal state is that there is a bulging signal and the average charging rate is less than the preset charging threshold, the charge consistency monitoring result of the target battery can be determined through the charging features, thereby improving the accuracy of the charge consistency monitoring result; moreover, the charge consistency of the target battery is predicted through the regression quantization model, and then the true charge consistency monitoring result of the target battery is determined based on the charge consistency prediction result, improving the efficiency and accuracy of determining the charge consistency monitoring result.

[0450] Next, an embodiment is used to illustrate how to obtain the bulging signal state of the target state data. In an exemplary embodiment, as Figure 22 shown, features are extracted from the target state data to obtain the charging features of the target battery, including:

[0451] S2201, Determine the pressure difference at each moment according to the compensated first voltage and the compensated second voltage at multiple moments in the target state data.

[0452] For any moment in the target state data, the compensated first voltage and the compensated second voltage at each moment are respectively determined as the pressure difference at the corresponding moment.

[0453] S2202, Determine the segmentation threshold and the pressure difference time series curve according to the pressure differences at each moment.

[0454] Among them, the segmentation threshold can determine the state of charge of the target battery.

[0455] The segmentation threshold can be determined according to a pre-trained segmentation threshold calculation model; specifically, the pressure difference at each moment can be input into the segmentation threshold calculation model, and the pressure difference at each moment can be analyzed by the segmentation threshold calculation model to obtain the segmentation threshold output by the segmentation threshold calculation model.

[0456] In an exemplary embodiment, as Figure 23 shown, determining the segmentation threshold according to the pressure difference at each moment includes the following steps:

[0457] S2301, downsample the pressure difference at each moment to obtain sampled pressure difference data.

[0458] Downsample the pressure difference at each moment, and then determine the downsampled pressure difference data as the sampled pressure difference data.

[0459] Specifically, according to a preset step size, the pressure difference can be obtained from the pressure difference at each moment at equal intervals in ascending order of time; for example, the pressure differences at each moment sorted in chronological order include: A1, A2, A3, A4, A5, A6; the preset step size is 2, then A2, A4, A6 can be determined as the sampled pressure difference data.

[0460] S2302, perform filtering processing on the sampled pressure difference data to obtain filtered pressure difference data.

[0461] The way to perform filtering processing on the sampled pressure difference data can be to input the sampled pressure difference data into a preset filtering model, and the filtering model performs filtering processing on the sampled pressure difference data to obtain the filtered pressure difference data output by the filtering model.

[0462] Optionally, the sampled pressure difference data can also be filtered according to the moving average filtering algorithm, and the filtered sampled pressure difference data is determined as the filtered pressure difference data; specifically, the size of the moving window can be set, and according to the size of the moving window, the moving window is moved from the sampled pressure difference data to take the pressure difference mean value to obtain the filtered sampled pressure difference data.

[0463] S2303, obtain the pressure difference mean value of the filtered pressure difference data.

[0464] Among them, the filtered pressure difference data includes multiple pressure differences, and the average value of the multiple pressure differences in the filtered pressure difference data can be determined as the pressure difference mean value of the filtered pressure difference data.

[0465] S2304, determine the segmentation threshold according to the pressure difference mean value and the maximum pressure difference in the filtered pressure difference data.

[0466] The segmentation threshold can be determined by a preset threshold determination model. For example, the average pressure difference and the filtered pressure difference data are input into the threshold determination model, and the threshold determination model analyzes the average pressure difference and the filtered pressure difference data to obtain the segmentation threshold output by the threshold determination model.

[0467] Optionally, the segmentation threshold can also be calculated according to the segmentation threshold calculation formula, as shown in formula (1).

[0468] (1)

[0469] Wherein, represents the segmentation threshold, represents the filtered pressure difference data, represents the maximum pressure difference in the filtered pressure difference data, represents the average pressure difference, represents the weight parameter.

[0470] The method for determining the power consistency provided by the embodiments of the present application performs downsampling on the pressure differences at each moment to obtain sampled pressure difference data, performs filtering processing on the sampled pressure difference data to obtain filtered pressure difference data, then obtains the average pressure difference of the filtered pressure difference data, and finally determines the segmentation threshold according to the average pressure difference and the maximum pressure difference in the filtered pressure difference data. In this method, the segmentation threshold is the segmentation threshold of the pressure difference time series curve. Before calculating the segmentation threshold through the pressure differences at each moment, the pressure differences at each moment are downsampled and filtered, so that the data volume for calculating the segmentation threshold is smaller and more accurate, and the filtered pressure difference data after sampling and filtering is used to calculate the segmentation threshold, thereby improving the accuracy of the segmentation threshold.

[0471] Among them, the pressure difference time series curve is drawn by arranging the pressure differences at each moment in chronological order; since the target state data is the data of the target battery during the charging process, as time increases, the SOC of the target battery increases. Therefore, the abscissa of the pressure difference time series curve can be converted into the SOC value of the target battery. As Figure 24 shown, Figure 24 is the pressure difference time series curve, the abscissa of the pressure difference time series curve is the SOC value of the target battery, and the ordinate is the pressure difference.

[0472] S2203. When the segmentation threshold divides the pressure difference time series curve into three consecutive regions and the maximum pressure difference of the pressure difference time series curve is in the second region of the three consecutive regions, it is determined that the bulging signal state of the target state data is a bulging signal existing.

[0473] The segmentation threshold line can be drawn according to the segmentation threshold, and the differential pressure time series curve is segmented by the segmentation threshold line. The way to segment the differential pressure time series curve can be that the abscissa of the differential pressure time series curve is used as the parallel line of the segmentation threshold line, and the intersection of the segmentation threshold line and the ordinate of the differential pressure time series curve is at the size of the segmentation threshold corresponding to the ordinate. As Figure 25 shown, Figure 25 This is the segmented differential pressure time series curve.

[0474] If the segmentation threshold divides the differential pressure time series curve into three consecutive regions, and the maximum differential pressure of the differential pressure time series curve is in the second region of the three consecutive regions, then it is determined that the bulge signal state of the target state data is that there is a bulge signal; otherwise, it is determined that the bulge signal state of the target state data is that there is no bulge signal.

[0475] Please continue to refer to Figure 25 , Figure 25 This is the situation where the segmentation threshold divides the differential pressure time series curve into three consecutive regions, and the maximum differential pressure is in the second region of the three consecutive regions.

[0476] The power quantity consistency determination method provided by the embodiment of the present application determines the differential pressure at each moment according to the compensated first voltage and the compensated second voltage at multiple moments in the target state data, and determines the segmentation threshold and the differential pressure time series curve according to the differential pressure at each moment. Then, when the segmentation threshold divides the differential pressure time series curve into three consecutive regions, and the maximum differential pressure of the differential pressure time series curve is in the second region of the three consecutive regions, it is determined that the bulge signal state of the target state data is that there is a bulge signal. In this method, since the segmentation threshold is determined according to the differential pressure at each moment in the target state data, and the differential pressure time series curve is segmented by the segmentation threshold, it can accurately determine whether there is a bulge signal in the target state data, improve the accuracy of determining the bulge signal state, and thus improve the reliability of the bulge signal state of the target state data.

[0477] In an exemplary embodiment, the charging characteristic further includes the average differential pressure greater than the segmentation threshold, that is, the average of the differential pressures greater than the segmentation threshold among the differential pressures at each moment; please continue to refer to Figure 25 , the average differential pressure greater than the segmentation threshold represents the average of the differential pressures in the second region of the three consecutive regions.

[0478] In an exemplary embodiment, as Figure 26 shown, obtaining the prediction error range of the regression quantization model under a preset consistency confidence level includes the following steps:

[0479] S2601, input the preset test data set into the regression quantization model, and obtain multiple power quantity consistency prediction results output by the regression quantization model.

[0480] Among them, the test data set includes the sample charging characteristics of the sample battery during multiple historical charging processes and the corresponding sample state-of-charge consistency results.

[0481] Each sample charging characteristic in the test data set can be respectively input into the regression quantization model. Through the regression quantization model, each sample charging characteristic is analyzed to obtain the state-of-charge consistency prediction results corresponding to each sample historical characteristic output by the regression quantization model.

[0482] S2602. Obtain the difference between each state-of-charge consistency prediction result and the corresponding sample state-of-charge consistency result.

[0483] Among them, the state-of-charge consistency prediction result is the predicted value obtained by fitting the regression quantization model, and the sample state-of-charge consistency result is the true value of the state-of-charge consistency result corresponding to the sample charging characteristic.

[0484] Respectively subtract each state-of-charge consistency prediction result from the corresponding sample state-of-charge consistency result to obtain the difference between each state-of-charge consistency prediction result and the corresponding sample state-of-charge consistency result.

[0485] S2603. Determine the prediction error range of the regression quantization model under the consistency confidence level according to each difference.

[0486] According to each difference, determine the residual distribution. Based on the residual distribution, the residual distribution can determine the random disturbance term of the regression quantization model, and the random disturbance term conforms to the normal distribution.

[0487] As Figure 27 shown, Figure 27 is the distribution diagram of each state-of-charge consistency prediction result and the corresponding sample state-of-charge consistency result. The abscissa represents the sample points, and the ordinate is the SOC consistency value; plot the differences between each state-of-charge consistency prediction result and the corresponding sample state-of-charge consistency result. As Figure 28 shown, it can be determined that each difference approximately conforms to the normal distribution.

[0488] Therefore, according to the residual distribution, the prediction error range of the regression quantization model under the consistency confidence level can be calculated to evaluate the error distribution of the prediction results of the regression quantization model.

[0489] The way to obtain the prediction error range of the regression quantization model under the consistency confidence level can be that, given the consistency confidence level conf, the upper and lower limits of the confidence interval can be calculated to satisfy:

[0490] (2)

[0491] Among them, represents the state-of-charge consistency prediction result, Represents the true power consistency result, Represents the consistency confidence level.

[0492] Simplify formula (2) to obtain formula (3).

[0493] (3)

[0494] Take the lower confidence limit of formula (3), and the confidence interval limits can be replaced with Quantiles, as shown in formula (4).

[0495] (4)

[0496] The confidence interval at a confidence level of conf can be obtained as:

[0497] (5)

[0498] Where, Is The corresponding quantile; then the prediction error range should be .

[0499] Optionally, according to each difference, the method for determining the prediction error range of the regression quantization model at the consistency confidence level can also be to obtain the mean value and standard deviation of the differences corresponding to each difference, as well as the z value (multiple of the standard deviation) corresponding to the consistency confidence level, and then determine the prediction error range of the regression quantization model at the consistency confidence level as [mean value of differences - z * standard deviation, mean value of differences + z * standard deviation]; where, the z value corresponding to the consistency confidence level can be determined by the look-up table method. For example, when the consistency confidence level is 95%, the corresponding z value is 1.96.

[0500] Based on the above prediction error range, the corresponding confidence interval can be determined according to the corresponding predicted value, and it can be found that the true sample power consistency result is within the confidence interval, as Figure 29 Shown.

[0501] In the power consistency determination method provided by the embodiments of the present application, a preset test data set is input into the regression quantization model, multiple power consistency prediction results output by the regression quantization model are obtained, then the differences between each power consistency prediction result and the corresponding sample power consistency result are obtained, and finally, according to each difference, the prediction error range of the regression quantization model at the consistency confidence level is determined. In this method, since the power consistency prediction result is predicted by the regression quantization model, and the sample power consistency result is the true power consistency result in the test data set, therefore, according to the differences between each power consistency prediction result and the corresponding sample power consistency result, the prediction error range of the regression quantization model at the consistency confidence level can be accurately measured.

[0502] In the above embodiments, the description is about determining the quantization result of the power consistency of the target battery at the current moment through the regression quantization model. Next, an embodiment is used to illustrate the acquisition process of the regression quantization model. In an exemplary embodiment, the acquisition process of the regression quantization model includes: obtaining a training sample set; the training sample set includes multiple sample charging characteristics and corresponding sample power consistency results; the sample charging characteristics represent historical charging characteristics with a bulge signal and an average charging rate less than a preset charging threshold; according to the training sample set, training the initial regression quantization model until the initial regression quantization model converges to obtain the regression quantization model.

[0503] Obtain the training data set, and then obtain the historical charging characteristics with a bulge signal and an average charging rate less than the preset charging threshold from the training data set. Determine the data corresponding to the historical charging characteristics with a bulge signal and an average charging rate less than the preset charging threshold in the training data set as the training sample set. Among them, the training sample set includes multiple sample charging characteristics and corresponding sample power consistency results; the method of obtaining the training data set is the same as that of the training data set when constructing the state classification model in the above embodiments, and this embodiment will not be elaborated here.

[0504] The initial regression quantization model is an initial multiple regression model, and the condition for the initial regression quantization model to converge can be that the coefficient of determination R calculated by the initial regression model through the test set 2 exceeds 0.95.

[0505] Among them, under the condition that there is a bulge signal in the state data during the charging process and the average charging rate is less than the preset rate threshold, it is found that the correlation between the power consistency result and features such as the mean pressure difference, the variance of the pressure difference, and the mean pressure difference of the pressure difference greater than the segmentation threshold exceeds 0.9.

[0506] Therefore, the initial regression quantization model can be trained with the mean pressure difference, the variance of the pressure difference, the mean pressure difference of the pressure difference greater than the segmentation threshold, the average charging rate, and the average temperature in the training sample set as independent variables and the sample power consistency result in the training sample set as the dependent variable until the initial regression quantization model converges, and the initial regression quantization model obtained after the initial regression quantization model converges is determined as the regression quantization model.

[0507] The method for determining the power consistency provided by the embodiments of the present application obtains a training sample set; wherein, the training sample set includes multiple sample charging characteristics and corresponding sample power consistency results, and the sample charging characteristics represent historical charging characteristics with a bulging signal and an average charging rate less than a preset charging threshold; then, according to the training sample set, the initial regression quantization model is trained until the initial regression quantization model converges to obtain a regression quantization model. In this method, the regression quantization model is trained by the historical charging characteristics and the corresponding sample power consistency results under specific conditions, so that when the charging characteristics of the battery meet the preset conditions, the power consistency monitoring result of the battery can be accurately determined through the regression quantization model.

[0508] In an exemplary embodiment, by establishing a computer improvement mechanism, the charging segments that do not span the platform area are complemented with historical charging segments within a time window, which can better reduce the distortion of charging information. A comparison can be made between using the computer improvement mechanism and not using the computer improvement mechanism, and the comparison results are shown in Table 2.

[0509] Table 2

[0510]

[0511] In an exemplary embodiment, the embodiments of the present application also provide a method for determining power consistency, as Figure 30 shown, this embodiment includes the following steps:

[0512] S3001, obtain the state data of the target battery during the charging process.

[0513] Among them, the state data includes data such as the maximum voltage, minimum voltage, current, temperature, and cumulative charged power at multiple moments.

[0514] S3002, input the current in the state data into the working condition compensation model to obtain the maximum voltage compensation amount and the minimum voltage compensation amount.

[0515] Among them, the working condition compensation model includes a maximum voltage compensation model and a minimum voltage compensation model. Input the current in the state data into the maximum voltage compensation model to obtain the maximum voltage compensation amount; input the current in the state data into the minimum voltage compensation model to obtain the minimum voltage compensation amount.

[0516] The construction process of the operating condition compensation model includes: obtaining multiple charging segments within a historical time period, where each charging segment includes the maximum voltage, minimum voltage, current, minimum temperature, and SOC of two battery cells at multiple charging moments. For any charging segment, data features with a current change greater than a preset given threshold between adjacent moments are obtained, including: the current at the previous moment and the next moment, data such as the maximum voltage, minimum voltage, temperature, and cumulative charged amount. For all charging segments, a multiple regression model is established for the voltage change data, temperature, and cumulative charged capacity corresponding to each current change data, obtaining the maximum multiple regression model and the minimum multiple regression model. The reference temperature of 25 °C, the reference cumulative charged amount of 0, and the reference current of 0 are respectively substituted into the maximum multiple regression model and the minimum multiple regression model to obtain the maximum operating condition compensation model and the minimum operating condition compensation model.

[0517] S3003, compensate the maximum voltage and the minimum voltage according to the maximum voltage compensation amount and the minimum voltage compensation amount respectively to obtain compensated voltage data, and determine the compensated voltage data and other data in the status data as compensated status data.

[0518] Among them, the compensated voltage data includes the maximum compensated voltage and the minimum compensated voltage.

[0519] S3004, judge whether the compensated voltage data crosses the plateau region according to the platform voltage range.

[0520] Set the platform voltage range [volt_min, volt_max]. If the maximum voltage value in the compensated voltage data is greater than or equal to volt_max, and the minimum voltage value in the compensated voltage data is less than or equal to volt_min, it is determined that the compensated voltage data crosses the plateau region; otherwise, it is determined that the compensated voltage data does not cross the plateau region.

[0521] S3005, if the compensated voltage data crosses the plateau region, obtain the target status data corresponding to the compensated voltage data within the platform voltage range from the compensated status data.

[0522] For any moment in the compensated status data, the status data with the maximum voltage less than or equal to volt_max and the minimum voltage greater than or equal to volt_min at the moment is determined as the target status data.

[0523] S3006, if the compensated voltage data does not cross the plateau region, obtain historical status data according to the time window, and integrate the historical status data and the target status data to obtain integrated status data, and obtain the target status data from the integrated status data.

[0524] Among them, the compensation voltage data and the historical state data are used as the evaluation priority index based on the amount of electricity charged within the platform voltage range. The more electricity is charged within the platform voltage range, the higher the priority. Then, the compensation state data and the historical state data are merged to obtain the merged state data, and the duplicate state data in the compensation state data and the historical state data is obtained. The duplicate state data with a higher priority is determined as the unique state data, and the duplicate state data in the merged state data is screened out according to the unique state data to obtain the integrated state data.

[0525] S3007, obtain the charging characteristics of the target state data.

[0526] Among them, the charging characteristics include the mean pressure difference, the variance of the pressure difference, the average temperature, the average charging rate, the mean pressure difference of the pressure difference greater than the segmentation threshold, and the bulging signal state.

[0527] S3008, classify the charging characteristics of the target state data through the state classification model to determine the SOC consistency monitoring result of the target battery at the current moment.

[0528] The state classification model includes a primary classification model and a secondary classification model. The mean pressure difference and the variance of the pressure difference in the target state data are input into the primary classification model. If the output result of the primary classification model is label 1, it is determined that the SOC consistency of the target battery is good at the current moment, and the SOC consistency threshold is less than or equal to 7% SOC; if the output result of the primary classification model is label 0, the mean pressure difference and the variance of the pressure difference in the target state data are input into the secondary classification model. If the output result of the secondary classification model is label 1, it is determined that the SOC consistency of the target battery is average at the current moment, and the SOC consistency threshold is greater than 7% SOC and less than 10% SOC; if the output result of the secondary classification model is label 0, it is determined that the SOC consistency of the target battery is poor at the current moment, and the SOC consistency threshold is greater than or equal to 10% SOC.

[0529] Among them, obtain the historical charging characteristics and the corresponding SOC consistency results in multiple historical charging processes; and match the historical charging characteristics and the SOC consistency results according to the principle of the most recent time, and use them as the training data set; among them, according to the static lookup table method, obtain the SOC corresponding to the maximum voltage and the minimum voltage of the battery at low SOC; calculate the SOC consistency result according to the SOC of the maximum voltage and the minimum voltage.

[0530] The training data set can be divided into two groups of labels. The first group of labels includes: data with SOC consistency <= 5% SOC is used as label 1, and data with SOC consistency >= 7% SOC is used as label 0, which is used to train a classification model to distinguish whether the consistency is good; the second group of labels includes: SOC consistency >= 12% SOC is used as label 0, and consistency <= 10% SOC is used as label 1, which is used to train a classification model to distinguish whether the consistency is very bad. A first-level classification model and a second-level classification model are constructed according to the first group of labels and the second group of labels respectively.

[0531] S3009, in the case that the target state data has a bulging signal and the average charging rate is less than the preset rate threshold, analyze the charging characteristics of the target state data through a quantized regression model to determine the confidence interval of the current power consistency of the target battery.

[0532] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same moment, but can be executed at different moments. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least a part of other steps or steps or stages in other steps.

[0533] Based on the same inventive concept, the embodiments of the present application also provide a power consistency determination device for implementing the power consistency determination method involved above. The implementation solutions provided by this device to solve problems are similar to the implementation solutions described in the above method. Therefore, the specific limitations in one or more embodiments of the power consistency determination device provided below can refer to the limitations on the power consistency determination method in the above text, and will not be repeated here.

[0534] In an exemplary embodiment, as Figure 31 shown, a power consistency determination device 3100 is provided, including: a first acquisition module 3101, a compensation module 3102, and a first determination module 3103, where:

[0535] The first acquisition module 3101 is configured to acquire the state data of the target battery during charging;

[0536] The compensation module 3102 is configured to perform charging condition compensation on the state data through a preset working condition compensation model to obtain compensated state data;

[0537] The first determination module 3103 is configured to determine the power consistency monitoring result of the target battery at the current moment according to the compensation status data.

[0538] In one embodiment, as Figure 32 shown, the compensation module 3102 includes:

[0539] The first acquisition unit 3201 is configured to acquire the data to be compensated during the charging process of the target battery from the status data;

[0540] The first compensation unit 3202 is configured to perform charging condition compensation on the data to be compensated through the working condition compensation model to obtain the compensated data;

[0541] The first determination unit 3203 is configured to determine the compensated data and other status data in the status data as the compensation status data.

[0542] In one embodiment, as Figure 33 shown, the data to be compensated includes the first voltage and the second voltage of the target battery at each moment during the charging process; the first compensation unit 3202 includes:

[0543] The first acquisition subunit 3301 is configured to, for any moment, obtain the first voltage compensation amount and the second voltage compensation amount at the moment through the working condition compensation model;

[0544] The superposition subunit 3302 is configured to superimpose the first voltage compensation amount and the first voltage to obtain the compensated first voltage; and superimpose the second voltage compensation amount and the second voltage to obtain the compensated second voltage;

[0545] The first determination subunit 3303 is configured to determine the compensated first voltage and the compensated second voltage as the compensated voltage data at the moment.

[0546] In one embodiment, the target battery includes at least two battery cells, and the first voltage and the second voltage at each moment are determined from the voltages of at least two battery cells at each moment.

[0547] In one embodiment, as Figure 34 shown, the working condition compensation model includes a first voltage compensation model and a second voltage compensation model; the status data includes the current of the target battery at each moment during the charging process; the first acquisition subunit 3301 includes:

[0548] The first obtaining subunit 3401 is configured to input the current at the moment into the first voltage compensation model to obtain the first voltage compensation amount;

[0549] The second obtaining subunit 3402 is configured to input the current at the moment into the second voltage compensation model to obtain the second voltage compensation amount.

[0550] In one embodiment, as Figure 35 shown, the device 3100 further includes:

[0551] A second acquisition module 3501, configured to acquire historical state data of a historical battery during multiple historical charging processes;

[0552] A third acquisition module 3502, configured to, for any one of the historical charging processes, acquire absolute values of current differences at adjacent moments during the historical charging process;

[0553] A second determination module 3503, configured to determine historical state data corresponding to adjacent moments with absolute values of current differences greater than a preset threshold as reference state data at adjacent moments during each historical charging process;

[0554] A construction module 3504, configured to construct a working condition compensation model according to the reference state data at adjacent moments.

[0555] In one embodiment, as Figure 36 shown, the construction module 3504 includes:

[0556] An extraction unit 3601, configured to perform feature extraction on the reference state data at adjacent moments to determine historical charging features at adjacent moments;

[0557] A construction unit 3602, configured to construct a working condition compensation model according to each historical charging feature.

[0558] In one embodiment, as Figure 37 shown, the historical charging features include a first voltage difference, a second voltage difference, a current difference, a temperature, and a cumulative charged amount; the construction unit 3602 includes:

[0559] A filtering subunit 3701, configured to perform outlier filtering on the historical charging features according to the corresponding relationships between each first voltage difference and each current difference, and between each second voltage difference and each current difference, to obtain filtered charging features;

[0560] A first construction subunit 3702, configured to construct a multiple regression model according to the filtered charging features;

[0561] A third obtaining subunit 3703, configured to substitute a preset reference temperature, a preset reference cumulative charged amount, and a preset reference current into the multiple regression model to obtain a working condition compensation model.

[0562] In one embodiment, as Figure 38 shown, the multiple regression model includes a first multiple regression model and a second multiple regression model; the first construction subunit 3702 includes:

[0563] The second construction subunit 3801 is configured to construct a first multiple regression model according to the first voltage difference, current difference, temperature, and cumulative charged power in the filtered charging characteristics;

[0564] The third construction subunit 3802 is configured to construct a second multiple regression model according to the second voltage difference, current difference, temperature, and cumulative charged power in the filtered charging characteristics.

[0565] In one embodiment, as Figure 39 shown, the first determination module 3103 includes:

[0566] The second acquisition unit 3901 is configured to, when the compensation voltage data in the compensation state data exceeds the platform voltage range, determine the compensation state data within the platform voltage range as the target state data; when the compensation voltage data in the compensation state data does not exceed the platform voltage range, determine the target state data according to the compensation state data and the historical state data; the historical state data includes the state data corresponding to the historical compensation voltage data not exceeding the platform voltage range within the historical time period;

[0567] The second determination unit 3902 is configured to determine the power consistency monitoring result of the target battery at the current moment according to the target state data.

[0568] In one embodiment, as Figure 40 shown, the second acquisition unit 3901 includes:

[0569] The merging subunit 4001 is configured to merge the compensation state data and the historical state data to obtain merged state data;

[0570] The second determination subunit 4002 is configured to determine the highest-priority repeated state data as the unique state data corresponding to each power according to the repeated state data corresponding to the compensation state data and the historical state data at each power;

[0571] The fourth obtaining subunit 4003 is configured to screen the merged state data according to each unique state data to obtain integrated state data;

[0572] The second acquisition subunit 4004 is configured to obtain the target state data from the integrated state data according to the platform voltage range.

[0573] In one embodiment, as Figure 41 shown, the apparatus 3100 further includes:

[0574] The fourth acquisition module 4101 is configured to acquire a first power charged by the compensation state data within the platform voltage range and a second power charged by the historical state data within the platform voltage range;

[0575] A third determination module 4102, configured to determine the priority of each piece of repeated status data according to the first power and the second power.

[0576] In one embodiment, the target status data includes a compensated first voltage and a compensated second voltage at multiple moments; as Figure 42 shown, the second determination unit 3902 includes:

[0577] A fifth obtaining subunit 4201, configured to determine the mean pressure difference and the variance of the pressure difference of the target status data according to the compensated first voltage and the compensated second voltage at each moment, and determine the charging characteristic of the target battery according to the mean pressure difference and the variance of the pressure difference;

[0578] A sixth obtaining subunit 4202, configured to input the charging characteristic into a status classification model to obtain a power consistency monitoring result of the target battery at the current moment.

[0579] In one embodiment, as Figure 43 shown, the status classification model includes a first-level classification model and a second-level classification model; the sixth obtaining subunit 4202 includes:

[0580] A seventh obtaining subunit 4301, configured to input the charging characteristic into the first-level classification model to obtain a preliminary classification result;

[0581] A third determination subunit 4302, configured to determine that the power consistency monitoring result at the current moment is good when the preliminary classification result is a first value; good means that the power consistency is less than or equal to a first consistency threshold;

[0582] A third obtaining subunit 4303, configured to input the charging characteristic into the second-level classification model to obtain the power consistency monitoring result of the target battery at the current moment when the preliminary classification result is a second value.

[0583] In one embodiment, as Figure 44 shown, the third obtaining subunit 4303 includes:

[0584] An eighth obtaining subunit 4401, configured to input the charging characteristic into the second-level classification model to obtain an advanced classification result;

[0585] A first judgment subunit 4402, configured to determine that the power consistency monitoring result at the current moment is average when the advanced classification result is a third value; average means that the power consistency is in the interval between the first consistency threshold and the second consistency threshold;

[0586] A second judgment subunit 4403, configured to determine that the power consistency monitoring result at the current moment is poor when the advanced classification result is a fourth value; poor means that the power consistency is greater than or equal to the second consistency threshold.

[0587] In one embodiment, as Figure 45 shown, the device 3100 further includes:

[0588] A fifth acquisition module 4501, configured to acquire historical charging characteristics of a historical battery during multiple historical charging processes, and acquire a power consistency result of the historical battery at a preset power;

[0589] A matching module 4502, configured to perform corresponding matching on each historical charging characteristic and each power consistency result according to time to obtain a training data set;

[0590] A training module 4503, configured to train an initial state classification model according to the training data set to obtain a state classification model.

[0591] In one embodiment, as Figure 46 shown, the fifth acquisition module 4501 includes:

[0592] A third acquisition unit 4601, configured to acquire a maximum measured voltage and a minimum measured voltage of a historical battery at a preset power for any preset power;

[0593] An obtaining unit 4602, configured to respectively substitute the maximum measured voltage and the minimum measured voltage into a voltage-power relational expression to obtain a first power corresponding to the maximum measured voltage and a second power corresponding to the minimum measured voltage;

[0594] A third determination unit 4603, configured to determine a difference between the first power and the second power as a power consistency result of the historical battery at the power.

[0595] In one embodiment, as Figure 47 shown, the second determination unit 3902 includes:

[0596] A ninth obtaining subunit 4701, configured to perform feature extraction on target state data to obtain a charging feature of a target battery; the charging feature includes a bulging signal state and an average charging rate;

[0597] A tenth obtaining subunit 4702, configured to input the charging feature into a regression quantization model when the bulging signal state is a bulging signal and the average charging rate is less than a preset charging threshold to obtain a power consistency prediction result of the target battery;

[0598] A fourth determination subunit 4703, configured to determine a power consistency monitoring result of the target battery at the current moment according to the power consistency prediction result.

[0599] In one embodiment, as Figure 48 shown, the charging feature includes a bulging signal state; the ninth obtaining subunit 4701 includes:

[0600] A fifth determination subunit 4801, configured to determine a pressure difference at each moment according to a compensated first voltage and a compensated second voltage at multiple moments in the target state data;

[0601] A sixth determination subunit 4802, configured to determine a segmentation threshold and a pressure difference time sequence curve according to the pressure difference at each moment;

[0602] A seventh determination subunit 4803, configured to determine that the bulge signal state of the target state data is a presence of a bulge signal when the segmentation threshold divides the pressure difference time sequence curve into three consecutive regions and the maximum pressure difference of the pressure difference time sequence curve is in the second region among the three consecutive regions.

[0603] In one embodiment, as Figure 49 shown, the sixth determination subunit 4802 includes:

[0604] A sampling subunit 4901, configured to perform downsampling on the pressure difference at each moment to obtain sampled pressure difference data;

[0605] A filtering subunit 4902, configured to perform filtering processing on the sampled pressure difference data to obtain filtered pressure difference data;

[0606] A fourth obtaining subunit 4903, configured to obtain an average value of the pressure difference of the filtered pressure difference data;

[0607] An eighth determination subunit 4904, configured to determine a segmentation threshold according to the average value of the pressure difference and the maximum pressure difference in the filtered pressure difference data.

[0608] In one embodiment, as Figure 50 shown, the fourth determination subunit 4703 includes:

[0609] A fifth obtaining subunit 5002, configured to obtain a prediction error range of the regression quantization model under a preset consistency confidence level;

[0610] A ninth determination subunit 5002, configured to determine a confidence interval of the current charge consistency of the target battery according to the charge consistency prediction result and the prediction error range; different confidence intervals represent different degrees of charge consistency;

[0611] A tenth determination subunit 5003, configured to determine the confidence interval of the current charge consistency as the charge consistency monitoring result of the target battery at the current moment.

[0612] In one embodiment, as Figure 51 shown, the fifth obtaining subunit 5002 includes:

[0613] The sixth acquisition subunit 5101 is configured to input a preset test data set into the regression quantization model, and acquire a plurality of power consistency prediction results output by the regression quantization model;

[0614] The seventh acquisition subunit 5102 is configured to acquire the difference between each power consistency prediction result and the corresponding sample power consistency result;

[0615] The eleventh determination subunit 5103 is configured to determine the prediction error range of the regression quantization model under the consistency confidence level according to each difference.

[0616] Each module in the above power consistency determination device can be implemented in whole or in part by software, hardware, and their combination. Each of the above modules can be embedded in the processor of the computer device in hardware form or independent of the processor, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each of the above modules.

[0617] In one embodiment, a computer device is further provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps in the above method embodiments are implemented.

[0618] The implementation principles and technical effects of the steps implemented by the processor in the embodiments of the present application are similar to the principles of the above power consistency determination method, and will not be elaborated here.

[0619] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0620] The implementation principles and technical effects of the steps implemented when the computer program in the embodiments of the present application is executed by a processor are similar to the principles of the above power consistency determination method, and will not be elaborated here.

[0621] In one embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0622] The implementation principles and technical effects of the steps implemented when the computer program in the embodiments of the present application is executed by a processor are similar to the principles of the above power consistency determination method, and will not be elaborated here.

[0623] It should be noted that the data involved in the present application (including but not limited to the data for analysis, the stored data, the displayed data, etc.) are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of the relevant data need to comply with the relevant regulations.

[0624] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0625] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0626] The above-described embodiments merely represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method for determining the consistency of electric quantity, characterized in that, The method includes: Obtaining state data of a target battery during the charging process; Performing charging condition compensation on the state data through a preset working condition compensation model to obtain compensated state data; Determining a monitoring result of the charge consistency of the target battery at the current moment according to the compensated state data.

2. The method according to claim 1, characterized in that, The performing charging condition compensation on the state data through a preset working condition compensation model to obtain compensated state data includes: Obtaining data to be compensated of the target battery during the charging process from the state data; Performing charging condition compensation on the data to be compensated through the working condition compensation model to obtain compensated data; Determining the compensated data and other state data in the state data as the compensated state data.

3. The method according to claim 2, characterized in that, The data to be compensated includes a first voltage and a second voltage of the target battery at each moment during the charging process; the performing charging condition compensation on the data to be compensated through the working condition compensation model to obtain compensated data includes: For any moment, obtaining a first voltage compensation amount and a second voltage compensation amount at the moment through the working condition compensation model; Superimposing the first voltage compensation amount and the first voltage to obtain a compensated first voltage; and superimposing the second voltage compensation amount and the second voltage to obtain a compensated second voltage; Determining the compensated first voltage and the compensated second voltage as the compensated voltage data at the moment.

4. The method according to claim 3, characterized in that, The working condition compensation model includes a first voltage compensation model and a second voltage compensation model; the state data includes the current of the target battery at each moment during the charging process; The obtaining a first voltage compensation amount and a second voltage compensation amount at the moment through the working condition compensation model includes: Inputting the current at the moment into the first voltage compensation model to obtain the first voltage compensation amount; Inputting the current at the moment into the second voltage compensation model to obtain the second voltage compensation amount.

5. The method according to any one of claims 1-4, characterized in that, The obtaining process of the working condition compensation model includes: Obtaining historical state data of a historical battery during multiple historical charging processes; For any one of the historical charging processes, obtaining the absolute value of the current difference between adjacent moments during the historical charging process; Determining the historical state data corresponding to adjacent moments with the absolute value of the current difference greater than a preset threshold as the reference state data for each adjacent moment during the historical charging process; Constructing the working condition compensation model according to the reference state data for each adjacent moment.

6. The method according to claim 5, characterized in that, The constructing the working condition compensation model according to the reference state data for each adjacent moment includes: Performing feature extraction on the reference state data for each adjacent moment to determine the historical charging features for each adjacent moment; Constructing the working condition compensation model according to each of the historical charging features.

7. The method according to claim 6, characterized in that, The historical charging features include a first voltage difference, a second voltage difference, a current difference, a temperature, and an accumulated charged amount; the constructing the working condition compensation model according to each of the historical charging features includes: Filter outliers from the historical charging characteristics according to the corresponding relationships between the first voltage differences and the current differences, and between the second voltage differences and the current differences, to obtain filtered charging characteristics; Construct a multiple regression model according to the filtered charging characteristics; Substitute a preset reference temperature, a preset reference cumulative charge capacity, and a preset reference current into the multiple regression model to obtain the operating condition compensation model.

8. The method according to claim 7, characterized in that, The multiple regression model includes a first multiple regression model and a second multiple regression model; the constructing a multiple regression model according to the filtered charging characteristics includes: Construct the first multiple regression model according to the first voltage difference, current difference, temperature, and cumulative charge in the filtered charging characteristics; Construct the second multiple regression model according to the second voltage difference, current difference, temperature, and cumulative charge in the filtered charging characteristics.

9. The method according to any one of claims 1-4, characterized in that,The determining the power consistency monitoring result of the target battery at the current moment according to the compensation status data includes: When the compensation voltage data in the compensation status data exceeds a preset platform voltage range, determine the compensation status data within the platform voltage range as the target status data; When the compensation voltage data in the compensation status data does not exceed the platform voltage range, determine the target status data according to the compensation status data and the historical status data; the historical status data includes the status data corresponding to the historical compensation voltage data that did not exceed the platform voltage range within a historical time period; Determine the power consistency monitoring result of the target battery at the current moment according to the target status data.

10. The method according to claim 9, wherein, The determining the target status data according to the compensation status data and the historical status data includes: Merge the compensation status data and the historical status data to obtain merged status data; According to the repeated status data corresponding to the compensation status data and the historical status data at each power level, determine the repeated status data with the highest priority as the unique status data corresponding to each power level; Filter the merged status data according to each of the unique status data to obtain integrated status data; Obtain the target status data from the integrated status data according to the platform voltage range.

11. The method according to claim 10, wherein, Before determining the repeated status data with the highest priority as the unique status data corresponding to each power level, the method further includes: Obtain a first charge amount charged by the compensation status data within the platform voltage range, and a second charge amount charged by the historical status data within the platform voltage range; Determine the priority of each of the repeated status data according to the first charge amount and the second charge amount.

12. The method according to claim 9, wherein, The target status data includes a compensated first voltage and a compensated second voltage at multiple moments; the determining the power consistency monitoring result of the target battery at the current moment according to the target status data includes: Determine the mean voltage difference and the variance of the voltage difference of the target status data according to the compensated first voltage and the compensated second voltage at each of the moments; Determine the charging characteristics of the target battery according to the mean pressure difference and the variance of the pressure difference; Input the charging characteristics into a state classification model to obtain the power consistency monitoring result of the target battery at the current moment.

13. The method according to claim 12, wherein, The state classification model includes a primary classification model and a secondary classification model; the step of inputting the charging characteristics into the state classification model to obtain the power consistency monitoring result of the target battery at the current moment includes: Input the charging characteristics into the primary classification model to obtain a preliminary classification result; When the preliminary classification result is the first value, determine that the power consistency monitoring result at the current moment is good; the good indicates that the power consistency is less than or equal to the first consistency threshold; When the preliminary classification result is the second value, input the charging characteristics into the secondary classification model to obtain the power consistency monitoring result of the target battery at the current moment.

14. The method according to claim 13, wherein, The step of inputting the charging characteristics into the secondary classification model to obtain the power consistency monitoring result of the target battery at the current moment includes: Input the charging characteristics into the secondary classification model to obtain an advanced classification result; When the advanced classification result is the third value, determine that the power consistency monitoring result at the current moment is average, and the average indicates that the power consistency is in the interval between the first consistency threshold and the second consistency threshold; When the advanced classification result is the fourth value, determine that the power consistency monitoring result at the current moment is poor; the poor indicates that the power consistency is greater than or equal to the second consistency threshold.

15. The method according to claim 12, wherein, The obtaining process of the state classification model includes: Obtain the historical charging characteristics of the historical battery during multiple historical charging processes, and obtain the power consistency results of the historical battery at a preset power; Correspondingly match each of the historical charging characteristics with each of the power consistency results according to time to obtain a training data set; Train an initial state classification model according to the training data set to obtain the state classification model.

16. The method according to claim 15, wherein, The step of obtaining the power consistency result of the historical battery at a preset power includes: For any preset power, obtain the maximum measured voltage and the minimum measured voltage of the historical battery at the preset power; Substitute the maximum measured voltage and the minimum measured voltage into the voltage-power relationship formula respectively to obtain the first power corresponding to the maximum measured voltage and the second power corresponding to the minimum measured voltage; Determine the difference between the first power and the second power as the power consistency result of the historical battery at the power.

17. The method according to claim 9, wherein, The step of determining the power consistency monitoring result of the target battery at the current moment according to the target state data includes: Extract features from the target state data to obtain the charging characteristics of the target battery; the charging characteristics include the bulging signal state and the average charging rate; When the bulging signal state is that there is a bulging signal and the average charging rate is less than the preset charging threshold, input the charging characteristics into a regression quantization model to obtain the power consistency prediction result of the target battery; Based on the predicted result of the power consistency, determine the monitoring result of the power consistency of the target battery at the current moment.

18. The method according to claim 17, wherein, The feature extraction of the target state data to obtain the charging feature of the target battery includes: Determine the pressure difference at each moment according to the compensated first voltage and the compensated second voltage at multiple moments in the target state data; Determine the segmentation threshold and the pressure difference time series curve according to the pressure differences at each moment; When the segmentation threshold divides the pressure difference time series curve into three continuous regions and the maximum pressure difference of the pressure difference time series curve is in the second region of the three continuous regions, determine that the bulging signal state of the target state data is the existence of a bulging signal.

19. The method according to claim 18, wherein, Determine the segmentation threshold according to the pressure differences at each moment, including: Perform downsampling on the pressure differences at each moment to obtain sampled pressure difference data; Perform filtering processing on the sampled pressure difference data to obtain filtered pressure difference data; Obtain the average pressure difference of the filtered pressure difference data; Determine the segmentation threshold according to the average pressure difference and the maximum pressure difference in the filtered pressure difference data.

20. The method according to claim 17, wherein, The determination of the monitoring result of the power consistency of the target battery at the current moment according to the predicted result of the power consistency includes: Obtain the prediction error range of the regression quantization model under a preset consistency confidence level; Determine the confidence interval of the current power consistency of the target battery according to the predicted result of the power consistency and the prediction error range; different confidence intervals represent different degrees of power consistency; Determine the confidence interval of the current power consistency as the monitoring result of the power consistency of the target battery at the current moment.

21. The method according to claim 20, wherein, The obtaining of the prediction error range of the regression quantization model under a preset consistency confidence level includes: Input a preset test data set into the regression quantization model to obtain multiple predicted results of the power consistency output by the regression quantization model; Obtain the difference between each predicted result of the power consistency and the corresponding sample power consistency result; Determine the prediction error range of the regression quantization model under the consistency confidence level according to each difference.

22. An apparatus for determining power consistency, wherein, The device includes: A first acquisition module for acquiring the state data of the target battery during charging; A compensation module for compensating the charging working condition of the state data through a preset working condition compensation model to obtain compensated state data; A first determination module for determining the monitoring result of the power consistency of the target battery at the current moment according to the compensated state data.

23. A computer device, comprising a memory and a processor, the memory storing a computer program, wherein, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 21.

24. A computer-readable storage medium, having a computer program stored thereon, wherein, When the computer program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 21.

25. A computer program product, comprising a computer program, wherein, When the computer program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 21.

Images