Prediction Method, System, Device and Medium for Power Battery Capacity Trajectory

Abstract

The present invention relates to a method, system, device and medium for predicting the capacity trajectory of a power battery. The method includes: within a preset sampling period, obtaining the driving mileage of an electrified vehicle and the corresponding battery parameters of the power battery, and dividing the sampling period into multiple operation periods according to the battery parameters; determining a target state sequence and a target driving condition sequence of the power battery according to the battery parameters and the driving mileage of each operation period; and inputting the target state sequence and the target driving condition sequence into a capacity prediction model to predict the capacity trajectory of the power battery in the next sampling period. The present invention improves the accuracy of predicting the capacity trajectory of the power battery.

Description

Technical Field

[0001] The present invention relates to the technical field of batteries, and particularly relates to a method, a system, a device and a medium for predicting the capacity trajectory of a power battery. Background Art

[0002] In order to improve the safety and economy of an energy storage system, accurately estimating the state of health (SOH) of a power battery in an electrified vehicle has become an increasingly concerned issue. Among them, electrified vehicles include, but are not limited to, electric vehicles, rail transit vehicles, electric ships, electric aircraft, electric bicycles, electric motorcycles, electric agricultural machinery or construction machinery, etc. Power batteries include, but are not limited to, lithium-ion batteries, sodium-ion batteries, quasi-solid-state batteries, all-solid-state batteries, etc.

[0003] As the power battery ages, its maximum available capacity gradually decays, and due to the large differences in the battery degradation trajectories caused by different types of power batteries and their usage conditions, how to accurately predict the degradation trajectory of the maximum available capacity of the power battery has important guiding significance for timely sending out alarm signals and optimizing user operations. Traditional methods for predicting the capacity trajectory of a power battery mainly rely on the power battery aging test data in a laboratory environment, but these methods cannot well adapt to the complex working conditions in actual applications. Therefore, there is a need to provide a method, a system, a device and a medium for predicting the capacity trajectory of a power battery. Summary of the Invention

[0004] In view of the above-mentioned disadvantages of the prior art, the purpose of the present invention is to provide a method, a system, a device and a medium for predicting the capacity trajectory of a power battery, which improves the problem of low accuracy in predicting the capacity trajectory of a power battery in the prior art.

[0005] To achieve the above purpose and other related purposes, the present invention provides a method for predicting the capacity trajectory of a power battery, which is applied to an electrified vehicle. The method includes: within a preset sampling period, obtaining the driving mileage of the electrified vehicle and the battery parameters of the corresponding power battery, and dividing the sampling period into multiple operating periods according to the battery parameters; determining a target state sequence and a target driving condition sequence of the power battery according to the battery parameters and the driving mileage of each operating period; inputting the target state sequence and the target driving condition sequence into a capacity prediction model to predict the capacity trajectory of the power battery in the next sampling period.

[0006] In an embodiment of the present invention, determining the target state sequence and the target driving condition sequence of the power battery according to the battery parameters and the driving mileage in each operation period includes: dividing all operation periods into multiple groups according to the driving mileage, and obtaining the target state parameters and the target driving condition parameters corresponding to the categories of the preset target state and the target driving condition in each group; arranging the target state parameters and the target driving condition parameters of each group in chronological order to form the target state sequence and the target driving condition sequence.

[0007] In an embodiment of the present invention, the process of determining the categories of the target state and the target driving condition is as follows: grouping the pre-acquired historical driving mileage, and obtaining the historical state sequence and the historical driving condition sequence of the power battery based on the historical battery parameters within each group; screening out the categories of the target state / categories of the target driving condition from the historical state sequence / historical driving condition sequence based on feature engineering.

[0008] In an embodiment of the present invention, screening out the categories of the target state / categories of the target driving condition from the historical state sequence / historical driving condition sequence based on feature engineering includes: respectively calculating the Pearson correlation coefficients between the parameters in the historical state sequence / historical driving condition sequence and the average maximum available capacity of the battery in the historical state sequence; statistically analyzing the Pearson correlation coefficients of each parameter, and taking the parameter categories with Pearson correlation coefficients greater than the preset correlation threshold as the corresponding categories of the target state / target driving condition.

[0009] In an embodiment of the present invention, obtaining the target state parameters and the target driving condition parameters corresponding to the categories of the preset target state and the target driving condition includes: obtaining the target state parameters of the current sampling period corresponding to the category of the target state according to the battery parameters and the corresponding driving mileage in each operation period; obtaining the target driving condition parameters of the current sampling period corresponding to the category of the target driving condition according to the battery parameters and the corresponding driving mileage in each operation period.

[0010] In an embodiment of the present invention, obtaining the target state parameters and the target driving condition parameters corresponding to the categories of the preset target state and the target driving condition includes: obtaining the target state parameters of the current sampling period corresponding to the category of the target state according to the battery parameters and the corresponding driving mileage in each operation period; monitoring the operating conditions of the electrified vehicle, predicting the operating conditions of the next sampling period based on the monitored operating conditions, and when it is predicted that the operating conditions change in the next sampling period, inputting the operating conditions of the next sampling period into a pre-trained driving condition prediction model to predict the target driving condition parameters of the next sampling period.

[0011] In one embodiment of the present invention, inputting the target state sequence and the target driving condition sequence into the capacity prediction model to predict the capacity trajectory of the power battery in the next sampling period includes: inputting the target state sequence and the target driving condition sequence into the encoder of the capacity prediction model, and obtaining a hidden state vector based on the time dependence between various parameters; inputting the hidden state vector into the decoder of the capacity prediction model, and gradually decoding the hidden state vector to predict the capacity trajectory of the power battery in the next sampling period.

[0012] In one embodiment of the present invention, there is also provided a prediction system for the capacity trajectory of a power battery. The system includes: a data acquisition module, configured to acquire the driving mileage of an electrified vehicle and the corresponding battery parameters of the power battery within a preset sampling period, and divide the sampling period into multiple operation periods according to the battery parameters; a grouped feature extraction module, configured to determine the target state sequence and the target driving condition sequence of the power battery according to the battery parameters and the driving mileage of each operation period; a capacity prediction module, configured to input the target state sequence and the target driving condition sequence into the capacity prediction model to predict the capacity trajectory of the power battery in the next sampling period.

[0013] In one embodiment of the present invention, there is also provided an electronic device, including: one or more processors; a storage device, configured to store one or more programs, and when the one or more programs are executed by the one or more processors, enable the electronic device to implement the prediction method for the capacity trajectory of the power battery described in any one of the above.

[0014] In one embodiment of the present invention, there is also provided a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor of a computer, enable the computer to execute the prediction method for the capacity trajectory of the power battery described in any one of the above.

[0015] As described above, the prediction method, system, device, and medium for the capacity trajectory of a power battery according to the present invention have the following beneficial effects: By acquiring and subdividing the driving mileage and battery parameters of an electrified vehicle within a sampling period, a target state sequence and a target driving condition sequence are generated. Since the data in the target driving condition sequence is closely related to the user's driving behavior, inputting these data into the capacity prediction model realizes more accurate prediction of the battery capacity trajectory considering the user's driving behavior, and improves the accuracy and reliability of the battery capacity trajectory prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic flowchart of a prediction method for the capacity trajectory of a power battery provided by an embodiment of the present invention;

[0017] Figure 2 It shows a structural block diagram of a prediction system for the power battery capacity trajectory provided by an embodiment of the present invention;

[0018] Figure 3 It shows a structural schematic diagram of an electronic device according to an embodiment of the present invention. Specific embodiments

[0019] The following uses specific specific examples to illustrate the implementation manners of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific implementation manners. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0020] It should be noted that the diagrams provided in the following embodiments only schematically illustrate the basic concept of the present invention. Therefore, only the components related to the present invention are shown in the diagrams, rather than being drawn according to the number, shape, and size of the components in actual implementation. The type, quantity, and ratio of each component in actual implementation can be arbitrarily changed, and the component layout type may also be more complex.

[0021] In the following description, a large number of details are explored to provide a more thorough explanation of the embodiments of the present invention. However, it is obvious to those skilled in the art that the embodiments of the present invention can be implemented without these specific details. In other embodiments, the structure and equipment are shown in the form of block diagrams rather than in detail to avoid making the embodiments of the present invention difficult to understand.

[0022] The inventor found that as the operating time of the electrified vehicle increases, the power battery installed therein will gradually age, and the maximum available capacity of the power battery will also decay accordingly. Accurately predicting this decay trajectory has important guiding significance for timely sending alarm signals and optimizing user operations. In addition, such electrified vehicles are usually operated by humans, and different driving behaviors will have different effects on the decay trajectory of the maximum available capacity of the power battery.

[0023] In view of the above situation, the present invention provides a method for predicting the power battery capacity trajectory. By obtaining and subdividing the driving mileage and battery parameters of the electrified vehicle within the sampling period, a target state sequence and a target driving condition sequence are generated. Since the data in the target driving condition sequence is closely related to the user's driving behavior, inputting this data into the capacity prediction model can achieve more accurate prediction of the battery capacity trajectory and improve the accuracy and reliability of the battery capacity trajectory prediction under the consideration of the user's driving behavior.

[0024] Please refer to Figure 1 , and the prediction method of the power battery capacity trajectory includes the following steps:

[0025] S1. During a preset sampling period, obtain the driving mileage of the electrified vehicle and the corresponding battery parameters of the power battery, and divide the sampling period into multiple operation periods according to the battery parameters.

[0026] When predicting the maximum available capacity trajectory of the power battery in the electrified vehicle, first, through the cloud server, during the preset sampling period, obtain the driving mileage of the electrified vehicle at each sampling point and various battery parameters of the power battery to be measured in the electrified vehicle. Among them, the driving mileage refers to the cumulative driving distance of the electrified vehicle, and the battery parameters are used to reflect the operating state of the power battery under different working conditions. The battery parameters include, but are not limited to, the state of charge, voltage, current, etc. According to the change trend of the state of charge or current in the battery parameters, divide the sampling period into multiple operation periods, where the operation periods include a charging period and a discharging period. Exemplarily, according to the change of the state of charge in the battery parameters, divide the sampling period into multiple charging periods and multiple discharging periods. Among them, the charging period is a period during which the power battery to be measured is receiving energy. During this period, the power battery to be measured is charged by an external power source, and the state of charge of the power battery to be measured gradually increases. The discharging period is a period during which the power battery to be measured releases energy for external use (such as driving the electrified vehicle or power supply equipment). During this period, the state of charge of the power battery to be measured gradually decreases.

[0027] It should be noted that in the present invention, data processing can be performed on data with an operation period of the charging period, or data processing can also be performed on data with a discharging period. Therefore, the subsequent "operation period" refers to the charging period or the discharging period within the sampling period.

[0028] It can be understood that considering that during the process of data interaction between the cloud server and the electrified vehicle, due to reasons such as communication errors, abnormal data points may occur in the received battery parameters or driving mileage and other data. To ensure the accuracy of subsequent data processing, in the present invention, data cleaning is also performed on each received battery parameter and driving mileage to remove abnormal values in the data. Among them, the data cleaning methods include, but are not limited to, interpolation or deletion of abnormal values, as long as the data quality can be improved, and no limitation is made here. Further, it should be noted that to ensure the time consistency of the battery parameters and driving mileage, in the present invention, the obtained battery parameters and driving mileage are also time-synchronized according to their respective timestamps to ensure that at each sampling point, the relevant data of the driving mileage and the battery parameters can be correctly corresponded.

[0029] S2. Determine the target state sequence and the target driving condition sequence of the power battery according to the battery parameters and the driving mileage in each operation period.

[0030] According to the battery parameters (such as voltage, current, state of charge, etc.) in different operation periods and the driving mileage corresponding to these periods, several different categories of state parameters are obtained, and these state parameters are used to reflect the operation state of the power battery. In addition, several different categories of driving condition parameters are obtained, and these driving condition parameters are used to characterize the operation conditions related to the battery pack or battery cell in the power battery. Among them, the state parameters include but are not limited to the average maximum available capacity of the battery, the cumulative driving mileage, the average driving mileage in the operation period, etc., and the driving condition parameters include but are not limited to the average battery pack voltage, the average battery pack current, the average ambient temperature, etc. Arrange these state parameters in chronological order to obtain the target state sequence. Similarly, arrange these driving condition parameters in chronological order to obtain the target driving condition sequence.

[0031] Specifically, in an embodiment of the present invention, the determining the target state sequence and the target driving condition sequence of the power battery according to the battery parameters and the driving mileage in each operation period includes:

[0032] Divide all operation periods into multiple groups according to the driving mileage, and obtain the target state parameters and the target driving condition parameters corresponding to the categories of the preset target states and the categories of the target driving conditions in each group;

[0033] Arrange the target state parameters and the target driving condition parameters of each group in chronological order to form the target state sequence and the target driving condition sequence.

[0034] Divide all operation periods into multiple groups according to the preset driving mileage interval (such as 500 kilometers), and each group corresponds to a time period with a fixed driving mileage interval. According to the categories of the preset target states and the categories of the target driving conditions, calculate the corresponding target state parameters and the target driving condition parameters according to the driving mileage and the battery parameters in the corresponding period in each group. Arrange the target state parameters and the target driving condition parameters of all groups in chronological order to form the target state sequence and the target driving condition sequence. It should be noted that since the operation states of the electrified vehicle are different in the charging period and the discharging period, the categories of the target states may also be different in these two periods.

[0035] Specifically, in an embodiment of the present invention, the process of determining the categories of the target state and the target driving condition is:

[0036] Group the historical driving mileage obtained in advance, and obtain the historical state sequence and historical driving condition sequence of the power battery based on the historical battery parameters within each group;

[0037] Based on feature engineering, screen out the categories of target states / categories of target driving conditions from the historical state sequence / historical driving condition sequence.

[0038] The categories of target states and target driving conditions are determined in advance according to the operating characteristics of the power battery and the electric vehicle and historical data analysis. When determining the categories of target states and target driving conditions, it is first necessary to collect the historical driving mileage and corresponding historical battery parameters of the electric vehicle within a historical sampling period (such as 1 month), and divide the sampling period into multiple operating periods according to the historical battery parameters. Group the historical operating periods according to the aforementioned driving mileage interval (such as 500 kilometers), and each group corresponds to a fixed driving mileage interval. For each group: extract the historical driving mileage and battery parameters of the corresponding period of each operating period within the group, and calculate the historical state parameters and historical driving condition parameters of the group according to the extracted data and the category of the operating period. Arrange the historical state parameters and historical driving condition parameters of each group in chronological order to form a historical state sequence and a historical driving condition sequence. And through feature engineering, screen out the categories of target states and target driving conditions from them.

[0039] Specifically, when the operating period is a charging period, group the charging periods within the historical sampling period according to the driving mileage to obtain multiple historical charging intervals. For each historical charging interval, calculate the historical state parameters of the historical charging interval according to the battery parameters such as voltage, current, temperature, charge quantity, and state of charge of each charging period. Among them, the historical state parameters of each historical charging interval include but are not limited to 7 historical state parameters: average maximum available battery capacity, cumulative driving mileage, average driving mileage per driving period, average discharge depth per driving period, average battery output energy, average rate of change of voltage with respect to state of charge, and average rate of change of voltage with respect to charge quantity. Among them, the average maximum available battery capacity is the mean value of the maximum available battery capacity of each period within the current interval, with the unit of ampere-hour. The cumulative driving mileage is the total driving mileage within the current interval, with the unit of kilometer. The average driving mileage per driving period is the average driving mileage of all periods within the current interval, with the unit of kilometer. The average discharge depth per driving period is the mean value of the discharge depth of each operating period within the current interval, with the unit of percentage (%). The average battery output energy is the mean value of the output energy of each operating period within the current interval, with the unit of kilowatt-hour. The average rate of change of voltage with respect to state of charge is the mean value of the rate of change of voltage corresponding to a unit change in state of charge within the current interval, expressed as where ΔV iΔV is the change in the battery pack voltage during the i-th operation period, obtained from the voltage difference of the battery pack at the start and end of this operation period. ΔSOC d_i ΔSOC is the change in the state of charge during the i-th operation period, obtained from the difference in the state of charge of the battery pack at the start and end of this operation period. N is the total number of operation periods within the current interval. The average rate of change of voltage with respect to charge quantity is the mean value of the rate of change of voltage corresponding to the change in unit charge quantity within the current interval, denoted as where ΔQ d_i is the change in the charge quantity during the i-th operation period. It can be understood that since these 7 historical state parameters will be obtained for each historical charging interval, arranging the historical state parameters of all historical charging intervals in sequence can form 7 historical state sequences.

[0040] Furthermore, the driving condition parameters are used to characterize the relevant parameters of the battery pack and battery cells. The historical driving condition parameters of each historical charging interval include but are not limited to the average battery pack voltage, average battery pack current, average ambient temperature, average voltage range of battery cells during the driving period, average average voltage of battery cells during the driving period, average standard deviation of battery cell voltage during the driving period, maximum and minimum values of battery cell voltage during the driving period, average range of temperature probes during the driving period, average average value of temperature probes during the driving period, and average standard deviation of temperature probes during the driving period, a total of 11 driving condition parameters. Among them, the average battery pack voltage is the mean value of the battery pack voltages during each driving period in this interval, with the unit of volt. The average battery pack current is the mean value of the battery pack currents during each driving period in this interval, with the unit of ampere. The average ambient temperature is the mean value of the ambient temperatures during each driving period in this interval, with the unit of degree Celsius. The average voltage range of battery cells during the driving period is the difference between the maximum and minimum values of the battery cell voltages in this interval, with the unit of volt. The average average voltage of battery cells during the driving period is the mean value of the battery cell voltages during all operation periods in this interval, denoted as where M is the total number of battery cells in the battery pack, N is the total number of operation periods in the interval, and V’ i,j is the voltage of the j-th battery cell during the i-th operation period in the interval. The average standard deviation of battery cell voltage during the driving period is the mean value of the standard deviations of the battery cell voltages during each operation period within the interval. The maximum / minimum value of the average battery cell voltage during the driving period is the mean value of the maximum / minimum values of the battery cell voltages during each operation period within the interval, with the unit of volt. The average range of temperature probes during the driving period is the average value of the difference between the highest and lowest temperatures measured by the temperature probes during each operation period within the interval, with the unit of degree Celsius. It can be understood that since these 11 historical driving condition parameters will be obtained for each historical charging interval, arranging the historical driving condition parameters of all historical charging intervals in sequence can form 11 historical driving condition sequences.

[0041] When the operation period is the discharge period, the vehicle may be in the running state. At this time, in addition to the above 7 parameters, the state parameters also include two state parameters: the average charge consumption per unit mileage and the average state of charge loss per unit mileage. Among them, the average charge consumption per unit mileage is the average value of the charge consumed per unit driving mileage within the interval, expressed as where, Δm i is the difference in the driving mileage during the i-th operation period, obtained from the difference in the driving mileage at the start and end times of the operation period, and ΔQ d_i is the change in the charge amount during the i-th operation period. The average state of charge loss per unit mileage is the average value of the state of charge loss per unit driving mileage within the interval, expressed as When the operation period is the discharge period, the driving condition parameters are the same as those in the charging period. Therefore, when the operation period is the discharge period, 9 historical state sequences and 11 historical driving condition sequences can be generated.

[0042] It should be noted that the categories of the above-listed historical driving condition parameters and historical state parameters are only a specific example. For different electrified vehicles, those skilled in the art can adaptively select the corresponding parameter categories based on specific needs, which are not limited herein.

[0043] In an embodiment of the present invention, the method for screening out the category of the target state / the category of the target driving condition from the historical state sequence / the historical driving condition sequence based on feature engineering includes:

[0044] Calculating the Pearson correlation coefficient between each parameter in the historical state sequence / the historical driving condition sequence and the average maximum available capacity of the battery in the historical state sequence respectively;

[0045] Statistical analysis of the Pearson correlation coefficients of each parameter, and taking the parameter categories with Pearson correlation coefficients greater than the preset correlation threshold as the corresponding categories of the target state / the target driving condition.

[0046] Although the above-extracted state parameters and driving condition parameters can characterize the relevant performance of the power battery, the amount of data is large. To reduce the amount of computation, the present invention also screens out multiple parameters strongly correlated with the power battery capacity through feature engineering. Specifically, the state parameters and driving condition parameters of the same category extracted within each group are arranged in chronological order to correspondingly form multiple historical state sequences and historical driving condition sequences. The Pearson correlation coefficients of these parameter sequences and the average maximum available capacity sequence of the battery are calculated according to formula (1):

[0047]

[0048] where, For the i-th parameter sequence x i The Pearson correlation coefficient between and the average maximum available battery capacity, x ij Is the j-th parameter value in the i-th parameter sequence, Is the mean of the i-th parameter sequence, z j Is the j-th parameter value in the average maximum available battery capacity sequence, Is the mean of all data in the average maximum available battery capacity sequence, and n is the total number of data points in the sequence. Among them, the above parameter sequence refers to the historical state parameter sequence or the historical driving condition parameter sequence.

[0049] Through the above process, the parameters with Pearson correlation coefficients greater than the preset correlation threshold (such as 0.6) are selected as the target state parameters and the target driving condition parameters. It can be understood that for different models of electrified vehicles, due to the different categories of historical state parameters and historical driving condition parameters selected, the corresponding categories of target states and target driving conditions are different.

[0050] Exemplarily, for a certain model of electrified vehicle, when the operation period is the discharge period, the corresponding categories of target states are the average maximum available battery capacity, the cumulative driving mileage, the average charge consumption per unit mileage, the average discharge depth during the driving period, the average charge state loss per unit mileage, and the average change rate of voltage with respect to the charge state, and the corresponding categories of target driving conditions are the minimum value of the average battery cell voltage during the driving period and the average value of the temperature probe during the driving period. For another model of electrified vehicle, when the operation period is the discharge period, the corresponding categories of target states are the average maximum available battery capacity and the cumulative driving mileage, and the corresponding categories of target driving conditions are the range of the average battery cell voltage during the driving period, the average value of the average battery cell voltage during the driving period, the maximum and minimum values of the average battery cell voltage during the driving period, the range of the temperature probe during the driving period, the average value of the temperature probe during the driving period, and the standard deviation of the temperature probe during the driving period.

[0051] In an embodiment of the present invention, the obtaining of the target state parameters and the target driving condition parameters corresponding to the preset categories of target states and target driving conditions includes:

[0052] According to the battery parameters and the corresponding driving mileage of each operation period, obtain the target state parameters of the current sampling period corresponding to the category of the target state;

[0053] According to the battery parameters and the corresponding driving mileage of each operation period, obtain the target driving condition parameters of the current sampling period corresponding to the category of the target driving condition.

[0054] Based on the battery parameters and corresponding driving mileage in each current operation period, according to the category of the target state and the category of the target driving condition obtained as described above, calculate the target state parameters corresponding to these categories and the target driving condition parameters for the current sampling period.

[0055] When the operating condition of the vehicle is in a stable state in the short term, the target state parameters and target driving condition parameters for the current sampling period can be obtained based on the battery parameters and driving mileage in the current sampling period according to the categories of the target state and the target driving condition. Specifically, after grouping all the operation periods according to the driving mileage, for each group, based on the obtained battery parameters and driving mileage, calculate the target state parameters corresponding to the category of the target state. After all the groups are processed, arrange the target state parameters of the same category in sequence to obtain the corresponding target state sequence. Exemplarily, if the category of the target state is "average maximum available battery capacity", then by calculating the average value of the maximum available battery capacity in each operation period within the group, the average maximum available battery capacity of the group can be obtained. Arrange the average maximum available battery capacities of all the groups in the order of the time stamps to obtain the target state sequence "average maximum available battery capacity sequence". Similarly, for the same group, the target driving condition parameters corresponding to the category of the target driving condition can also be obtained based on the battery parameters and driving mileage in all the operation periods within the group. After all the groups are processed, arrange the target driving condition parameters of the same category in sequence to obtain the corresponding target state sequence. Exemplarily, if the category of the target driving condition is "average ambient temperature", then by calculating the average value of the ambient temperature in each operation period within the group, the average ambient temperature of the group can be obtained. Arrange the average ambient temperatures of all the groups in the order of the time stamps to obtain the target driving condition sequence "average ambient temperature sequence".

[0056] In another embodiment of the present invention, the obtaining of the target state parameters and target driving condition parameters corresponding to the categories of the preset target state and the target driving condition includes:

[0057] Based on the battery parameters and corresponding driving mileage in each operation period, obtain the target state parameters for the current sampling period corresponding to the category of the target state;

[0058] Monitor the operating condition of the electrified vehicle, and predict the operating condition in the next sampling period based on the monitored operating condition. When it is predicted that the operating condition changes in the next sampling period, input the operating condition in the next sampling period into a pre-trained driving condition prediction model to predict the target driving condition parameters in the next sampling period.

[0059] Predict the operating condition of the next sampling period based on the operating condition of the vehicle detected in the current sampling period, where the operating condition includes, but is not limited to, the driving speed, load condition, external environmental conditions, etc. of the electrified vehicle. When it is predicted that the operating condition of the electrified vehicle will change in the next sampling period, for example, the environmental temperature changes in the next sampling period, or the driving speed slows down, etc., these situations may cause the target driving condition parameters in the current sampling period to not accurately represent the future operating state. To achieve an accurate prediction of the battery capacity trajectory, the target driving condition parameters for the next sampling period can be predicted based on the predicted operating condition. Specifically, the predicted operating condition of the next sampling period is input into the trained driving condition prediction model to obtain the target driving condition parameters for the next sampling period. Among them, the driving condition prediction model can be any machine learning or deep learning model, which is not limited here. Preferably, the driving condition prediction model has the same model architecture as the following capacity prediction model. It can be understood that when the operating condition of the vehicle remains unchanged in the next sampling period, at this time, the target driving condition parameters of the current sampling period can be obtained according to the battery parameters and driving mileage of the current sampling period as described above. It is also possible to predict the target driving condition parameters of the next sampling period through the above method.

[0060] S3. Input the target state sequence and the target driving condition sequence into the capacity prediction model to predict the capacity trajectory of the power battery in the next sampling period.

[0061] Specifically, in an embodiment of the present invention, the step of inputting the target state sequence and the target driving condition sequence into the capacity prediction model to predict the capacity trajectory of the power battery in the next sampling period includes:

[0062] Input the target state sequence and the target driving condition sequence into the encoder of the capacity prediction model, and obtain a hidden state vector based on the time dependence between the parameters.

[0063] Input the hidden state vector into the decoder of the capacity prediction model, and gradually decode the hidden state vector to predict the capacity trajectory of the power battery in the next sampling period.

[0064] The capacity prediction model can be any deep learning model in the Sequence-to-Sequence (Seq2Seq) framework. The Seq2Seq framework includes two parts: an encoder and a decoder, both of which can be composed of neural networks. Among them, the neural networks that make up the encoder and decoder include, but are not limited to, the Bi-directional Gate Recurrent Unit (Bi-GRU), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), Bi-directional LSTM (Bi-LSTM), etc. Exemplarily, when the capacity prediction model is a Bi-directional LSTM model, the target state sequence and the target driving condition sequence are input into the encoder of the capacity prediction model to extract the time-dependent information therein and generate multiple hidden state vectors. Each hidden state vector is input into the decoder, and according to the current hidden state vector and the output of the previous time step, each time step is processed recursively to gradually generate the capacity trajectory of the next sampling period. It can be understood that since there are two types of target driving condition sequences, in the present invention, the prediction of the capacity trajectory of the next sampling period can be achieved through two different data. One is to input the target state sequence of the current sampling period and the target driving condition sequence of the current sampling period into the capacity prediction model to obtain the capacity trajectory of the next sampling period. The other is to input the target state sequence of the current sampling period and the target driving condition sequence of the next sampling period into the capacity prediction model to obtain the capacity trajectory of the next sampling period.

[0065] It should be noted that the capacity prediction model is obtained through training. The dataset for training the capacity prediction model can be obtained either through open-source channels or by self-collection. Among them, the open-source channels can be through the National Big Data Alliance of New Energy Vehicles (NDANEV), from which the operation data periods of 200 electric buses from January 2019 to December 2019 are obtained. Each period represents a charging or discharging process. The 200 vehicles include 2 different vehicle models, with 100 vehicles in each model. During the training phase, considering that some vehicle data is less, the present invention selects 172 vehicles (86 vehicles each) with a cumulative mileage exceeding 30,000 kilometers from the two models, each with 100 vehicles, for constructing the training set, validation set, and test set. For each vehicle model, 5 sets of training sets and validation sets are constructed using the cross-validation method (the division ratio of each set of training sets and validation sets is the same, the training set contains 64 vehicles, the validation set contains 16 vehicles, and the 16 vehicles in each validation set are not repeated), and the remaining 6 vehicles are used as the test set. Five Seq2Seq models are trained through these 5 sets of training sets and validation sets, and the model with the smallest validation error is selected as the final capacity trajectory prediction model. The trained capacity trajectory prediction model performs excellently on the test set, with a maximum mean absolute error of 0.78% and a maximum root mean square error of 0.94%, showing the high precision and reliability of the model.

[0066] Please refer to Figure 2 , the prediction system 100 for the power battery capacity trajectory includes: a data acquisition module 110, a grouped feature extraction module 120, and a capacity prediction module 130. Among them, the data acquisition module 110 is used to obtain the driving mileage of the electrified vehicle and the corresponding battery parameters of the power battery within a preset sampling period, and divide the sampling period into multiple operation periods according to the battery parameters. The grouped feature extraction module 120 is used to determine the target state sequence and the target driving condition sequence of the power battery according to the battery parameters and the driving mileage of each operation period. The capacity prediction module 130 is used to input the target state sequence and the target driving condition sequence into the capacity prediction model to predict the capacity trajectory of the power battery in the next sampling period.

[0067] For the specific limitations of the prediction system for the power battery capacity trajectory, reference can be made to the limitations of the prediction method for the power battery capacity trajectory described above, which will not be elaborated here. Each module in the above prediction system for the power battery capacity trajectory can be implemented in whole or in part through software, hardware, and their combination. The above modules can be embedded in the processor of the computer device in a hardware format or be independent of it, or can be stored in the memory of the computer device in a software format, so that the processor can call the corresponding operations of the above modules.

[0068] It should be noted that, in order to highlight the innovative part of the present invention, modules that are not closely related to solving the technical problems proposed by the present invention are not introduced in this embodiment, but this does not mean that there are no other modules in this embodiment.

[0069] Please refer to Figure 3 , the electronic device 1 may include a memory 12, a processor 13 and a bus, and may also include a computer program stored in the memory 12 and operable on the processor 13, such as a prediction program for the power battery capacity trajectory.

[0070] Among them, the memory 12 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, mobile hard disk, multimedia card, card-type memory (such as: SD or DX memory, etc.), magnetic memory, magnetic disk, optical disc, etc. The memory 12 may be an internal storage unit of the electronic device 1 in some embodiments, such as the mobile hard disk of the electronic device 1. The memory 12 may also be an external storage device of the electronic device 1 in other embodiments, such as a plug-in mobile hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc. equipped on the electronic device 1. Further, the memory 12 may also include both the internal storage unit and the external storage device of the electronic device 1. The memory 12 can not only be used to store application software installed in the electronic device 1 and various types of data, such as the code for predicting the power battery capacity trajectory, etc., but also be used to temporarily store data that has been output or will be output.

[0071] The processor 13 may be composed of integrated circuits in some embodiments. For example, it may be composed of a single packaged integrated circuit, or may be composed of multiple integrated circuits with the same or different functions, including a combination of one or more Central Processing Units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 13 is the control core (Control Unit) of the electronic device 1, connecting various components of the entire electronic device 1 through various interfaces and lines, and by running or executing programs or modules stored in the memory 12 (such as the prediction program for the power battery capacity trajectory, etc.), and calling the data stored in the memory 12, to execute various functions of the electronic device 1 and process data.

[0072] The processor 13 executes the operating system of the electronic device 1 and various installed application programs. The processor 13 executes the application program to implement the steps in the above-mentioned method for predicting the power battery capacity trajectory.

[0073] Exemplarily, the computer program may be divided into one or more modules, which are stored in the memory 12 and executed by the processor 13 to complete the present application. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, and these instruction segments are used to describe the execution process of the computer program in the electronic device 1. For example, the computer program may be divided into a data acquisition module 110, a packet feature extraction module 120, and a capacity prediction module 130.

[0074] The integrated unit implemented in the form of the software functional module described above may be stored in a computer-readable storage medium, and the computer-readable storage medium may be non-volatile or volatile. The above software functional module is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a computer device, or a network device, etc.) or a processor to execute some functions of the method for predicting the power battery capacity trajectory according to each embodiment of the present application.

[0075] In summary, for a method, a system, a device, and a medium for predicting a power battery capacity trajectory disclosed by the present invention, by obtaining and subdividing the driving mileage and battery parameters of an electrified vehicle during a sampling period, a target state sequence and a target driving condition sequence are generated. Since the data in the target driving condition sequence is closely related to the user's driving behavior, by inputting this data into the capacity prediction model, more accurate prediction of the battery capacity trajectory is achieved while considering the user's driving behavior, and the accuracy and reliability of the prediction of the power battery capacity trajectory are improved. Therefore, the present invention effectively overcomes various disadvantages in the prior art and has high industrial utilization value.

[0076] The above embodiments are only illustrative of the principles and effects of the present invention and are not intended to limit the present invention. Any person familiar with this technology can modify or change the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or changes completed by those with ordinary knowledge in the technical field without departing from the spirit and technical idea disclosed by the present invention should still be covered by the claims of the present invention.

Claims

1. A method for predicting power battery capacity trajectory, characterized in that: Applied to an electric vehicle, the method comprises: Acquire the mileage of the electric vehicle and the battery parameters of the corresponding power battery within a preset sampling period, and divide the sampling period into a plurality of operating periods according to the battery parameters; Determine the target state sequence and target driving condition sequence of the power battery based on the battery parameters and driving mileage in each operating period; Inputting the target state sequence and the target driving condition sequence into a capacity prediction model to predict the capacity trajectory of the power battery in the next sampling period; Determining the target state sequence and target driving condition sequence of the power battery according to the battery parameters and driving mileage in each operating period includes: Dividing all the operating time periods into a plurality of groups according to the driving mileage, and obtaining a target state parameter and a target driving condition parameter corresponding to a preset target state category and a preset target driving condition category in each group; Arrange the target state parameters and target driving condition parameters of each group in chronological order to form a target state sequence and a target driving condition sequence; wherein the target state parameter is used to characterize the operating state of the power battery, and the target driving condition parameter is used to characterize the operating conditions related to the battery pack or battery cell in the power battery; The process of determining the target state and the category of the target driving condition is as follows: According to the historical driving mileage acquired in advance, the historical driving mileage is grouped, and based on the historical battery parameters in each group, a historical state sequence and a historical driving condition sequence of the power battery are acquired; Filtering out a target state category / target driving condition category from the historical state sequence / the historical driving condition sequence based on feature engineering; The step of filtering out the target state category / target driving condition category from the historical state sequence / the historical driving condition sequence based on feature engineering includes: respectively calculating the Pearson correlation coefficient between each parameter in the historical state sequence / the historical driving condition sequence and the average maximum available battery capacity of the historical state sequence; The Pearson correlation coefficient of each parameter is counted, and the parameter category with a Pearson correlation coefficient greater than a preset correlation threshold is used as the category of the corresponding target state / target driving condition.

2. The method for predicting the power battery capacity trajectory according to claim 1, characterized in that: The step of obtaining target state parameters and target driving condition parameters corresponding to preset target state categories and target driving condition categories includes: According to the battery parameters and the corresponding driving mileage of each operating period, obtaining the target state parameters of the current sampling period corresponding to the category of the target state; According to the battery parameters and the corresponding driving mileage in each operating period, the target driving condition parameters of the current sampling period corresponding to the type of the target driving condition are obtained.

3. The method for predicting the power battery capacity trajectory according to claim 1, characterized in that: The step of obtaining target state parameters and target driving condition parameters corresponding to preset target state categories and target driving condition categories includes: According to the battery parameters and the corresponding driving mileage of each operating period, obtaining the target state parameters of the current sampling period corresponding to the category of the target state; Monitor the operating conditions of the electrified vehicle and predict the operating conditions of the next sampling period based on the monitored operating conditions. When it is predicted that the operating conditions will change within the next sampling period, input the operating conditions of the next sampling period into a pre-trained driving condition prediction model to predict the target driving condition parameters of the next sampling period.

4. The method for predicting the power battery capacity trajectory according to claim 1, characterized in that: The step of inputting the target state sequence and the target driving condition sequence into a capacity prediction model to predict the capacity trajectory of the power battery in the next sampling period includes: Inputting the target state sequence and the target driving condition sequence into an encoder of the capacity prediction model, and obtaining a hidden state vector based on the time dependency between various parameters; The latent state vector is input into the decoder of the capacity prediction model, the latent state vector is decoded step by step, and the capacity trajectory of the power battery in the next sampling period is predicted.

5. A prediction system for power battery capacity trajectory, characterized in that: The system comprises: A data acquisition module, used to acquire the mileage of the electric vehicle and the battery parameters of the corresponding power battery within a preset sampling period, and divide the sampling period into a plurality of operating time periods according to the battery parameters; A grouping feature extraction module is used to determine the target state sequence and target driving condition sequence of the power battery according to the battery parameters and driving mileage in each operating period; A capacity prediction module, used for inputting the target state sequence and the target driving condition sequence into a capacity prediction model to predict the capacity trajectory of the power battery in the next sampling period; Determining the target state sequence and target driving condition sequence of the power battery according to the battery parameters and driving mileage in each operating period includes: Dividing all the operating time periods into a plurality of groups according to the driving mileage, and obtaining a target state parameter and a target driving condition parameter corresponding to a preset target state category and a preset target driving condition category in each group; Arrange the target state parameters and target driving condition parameters of each group in chronological order to form a target state sequence and a target driving condition sequence; wherein the target state parameter is used to characterize the operating state of the power battery, and the target driving condition parameter is used to characterize the operating conditions related to the battery pack or battery cell in the power battery; The process of determining the target state and the category of the target driving condition is as follows: According to the historical driving mileage acquired in advance, the historical driving mileage is grouped, and based on the historical battery parameters in each group, a historical state sequence and a historical driving condition sequence of the power battery are acquired; Filtering out a target state category / target driving condition category from the historical state sequence / the historical driving condition sequence based on feature engineering; The step of filtering out the target state category / target driving condition category from the historical state sequence / the historical driving condition sequence based on feature engineering includes: respectively calculating the Pearson correlation coefficient between each parameter in the historical state sequence / the historical driving condition sequence and the average maximum available battery capacity of the historical state sequence; The Pearson correlation coefficient of each parameter is counted, and the parameter category with a Pearson correlation coefficient greater than a preset correlation threshold is used as the category of the corresponding target state / target driving condition.

6. An electronic device, characterized in that: The electronic device comprises: one or more processors; A storage device for storing one or more programs, which, when executed by the one or more processors, enables the electronic device to implement the method for predicting the power battery capacity trajectory as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor of a computer, the computer is enabled to execute the method for predicting the power battery capacity trajectory as described in any one of claims 1 to 4.

Images