CN117054893B - Training method of battery capacity prediction model, battery capacity prediction method and device - Google Patents
Training method of battery capacity prediction model, battery capacity prediction method and device Download PDFInfo
- Publication number
- CN117054893B CN117054893B CN202311318555.6A CN202311318555A CN117054893B CN 117054893 B CN117054893 B CN 117054893B CN 202311318555 A CN202311318555 A CN 202311318555A CN 117054893 B CN117054893 B CN 117054893B
- Authority
- CN
- China
- Prior art keywords
- data
- battery capacity
- capacity
- battery
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 71
- 238000012549 training Methods 0.000 title claims abstract description 51
- 230000008859 change Effects 0.000 claims description 27
- 230000002159 abnormal effect Effects 0.000 claims description 16
- 238000004590 computer program Methods 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 11
- 238000003860 storage Methods 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 6
- 230000001174 ascending effect Effects 0.000 claims description 3
- 230000001419 dependent effect Effects 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 239000000758 substrate Substances 0.000 claims 1
- 238000012360 testing method Methods 0.000 description 23
- 238000005562 fading Methods 0.000 description 16
- 230000008569 process Effects 0.000 description 15
- 238000010586 diagram Methods 0.000 description 12
- 238000001514 detection method Methods 0.000 description 11
- 230000005856 abnormality Effects 0.000 description 8
- 238000004364 calculation method Methods 0.000 description 7
- 230000032683 aging Effects 0.000 description 6
- 238000013500 data storage Methods 0.000 description 5
- 230000036541 health Effects 0.000 description 5
- 238000010801 machine learning Methods 0.000 description 4
- 238000009826 distribution Methods 0.000 description 3
- 238000009499 grossing Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000007704 transition Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000003491 array Methods 0.000 description 2
- 238000013475 authorization Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- HBBGRARXTFLTSG-UHFFFAOYSA-N Lithium ion Chemical compound [Li+] HBBGRARXTFLTSG-UHFFFAOYSA-N 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000007599 discharging Methods 0.000 description 1
- 229910001416 lithium ion Inorganic materials 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/367—Software therefor, e.g. for battery testing using modelling or look-up tables
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/382—Arrangements for monitoring battery or accumulator variables, e.g. SoC
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/392—Determining battery ageing or deterioration, e.g. state of health
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Charge And Discharge Circuits For Batteries Or The Like (AREA)
- Secondary Cells (AREA)
Abstract
The invention discloses a training method of a battery capacity prediction model, a battery capacity prediction method and a device. The training method of the battery capacity prediction model comprises the following steps: acquiring capacity sequence data of a battery, wherein the capacity sequence data characterizes the relationship between the capacity of the battery and the number of charging cycles; extracting at least one sub-sequence data from the capacity sequence data as sample data based on a data variation characteristic of the capacity sequence data; inputting the charge cycle times in the sample data into a battery capacity prediction model to be trained, and outputting a predicted value of the battery capacity; model parameters of a battery capacity prediction model to be trained are adjusted based on a loss value between a predicted value of the battery capacity and a sample value of the battery capacity in the sample data to obtain a trained battery capacity prediction model. The invention improves the accuracy of the battery capacity data.
Description
Technical Field
The present invention relates to the technical field of battery and machine learning, and in particular, to a training method for a battery capacity prediction model, a battery capacity prediction method and a device.
Background
The battery long-cycle test has a long test period, and can continuously occupy a test channel, and the tested battery is usually taken down and the test channel is replaced in the test process due to the consideration of test strategies, method feasibility and test cost control. The process can cause the capacity of the battery to be tested to rebound, so that huge fluctuation or bulge appears on a battery capacity attenuation curve (a fixing curve), the quality of battery capacity attenuation data is seriously reduced, the real attenuation trend of the battery capacity is covered, and the analysis based on the battery capacity attenuation data is not beneficial to evaluating the service life condition of the battery to be tested.
Disclosure of Invention
Embodiments of the present application aim to solve one of the technical problems in the related art at least to some extent. Therefore, an object of an embodiment of the present application is to provide a training method for a battery capacity prediction model, a battery capacity prediction method, a device, an electronic apparatus, a storage medium, and a program product.
The embodiment of the application provides a training method of a battery capacity prediction model, which comprises the following steps: acquiring capacity sequence data of a battery, wherein the capacity sequence data characterizes the relationship between the capacity of the battery and the number of charging cycles; extracting at least one sub-sequence data from the capacity sequence data as sample data based on a data variation characteristic of the capacity sequence data; inputting the charge cycle times in the sample data into a battery capacity prediction model to be trained, and outputting a predicted value of the battery capacity; and adjusting model parameters of the battery capacity prediction model to be trained based on a predicted value of the battery capacity and a loss value between sample values of the battery capacity in the sample data to obtain a trained battery capacity prediction model.
Another embodiment of the present application provides a battery capacity prediction method, including: acquiring the charge cycle times of a battery; and inputting the charging cycle times into a trained battery capacity prediction model, and predicting the battery capacity by the trained battery capacity prediction model, wherein the trained battery capacity prediction model is obtained by training by using the training method of the battery capacity prediction model.
Another embodiment of the present application provides a training device for a battery capacity prediction model, the device including: the device comprises an acquisition module, an extraction module, an output module and an adjustment module. The system comprises an acquisition module, a storage module and a storage module, wherein the acquisition module is used for acquiring capacity sequence data of a battery, and the capacity sequence data represents the relationship between the capacity of the battery and the number of charging cycles; an extraction module for extracting at least one sub-sequence data from the capacity sequence data as sample data based on a data variation characteristic of the capacity sequence data; the output module is used for inputting the charge cycle times in the sample data into a battery capacity prediction model to be trained and outputting a predicted value of the battery capacity; and the adjusting module is used for adjusting the model parameters of the battery capacity prediction model to be trained based on the predicted value of the battery capacity and the loss value between the sample values of the battery capacity in the sample data so as to obtain a trained battery capacity prediction model.
Another embodiment of the present application provides a battery capacity prediction apparatus, including: the system comprises an acquisition module and a prediction module. The acquisition module is used for acquiring the charging cycle times of the battery; and the prediction module is used for inputting the charging cycle times into a trained battery capacity prediction model, and predicting the battery capacity by the trained battery capacity prediction model, wherein the trained battery capacity prediction model is obtained by training by using the training device of the battery capacity prediction model.
Another embodiment of the present application provides an electronic device, including a memory storing a computer program and a processor implementing the steps of the method according to any of the above embodiments when the processor executes the computer program.
Another embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, which when executed by a processor, implements the steps of the method of any of the above embodiments.
Another embodiment of the present application provides a computer program product comprising instructions that, when executed by a processor of a computer device, enable the computer device to perform the steps of the method according to any one of the embodiments above.
In the above embodiment, the capacity sequence data of the battery is acquired; extracting at least one sub-sequence data from the capacity sequence data as sample data based on a data variation characteristic of the capacity sequence data; inputting the charge cycle times in the sample data into a battery capacity prediction model to be trained, and outputting a predicted value of the battery capacity; based on the predicted value of the battery capacity and the loss value between the sample values of the battery capacity in the sample data, the model parameters of the battery capacity prediction model to be trained are adjusted to obtain a trained battery capacity prediction model, and the accuracy of the battery capacity data is improved.
Drawings
Fig. 1A is a schematic diagram of a system architecture according to an embodiment of the present application;
fig. 1B is a flowchart of a training method of a battery capacity prediction model according to an embodiment of the present application;
FIG. 2 is a schematic diagram of acquiring sample data according to an embodiment of the present application;
FIG. 3 is a schematic diagram of a data decay curve provided by an embodiment of the present application;
fig. 4 is a flowchart of a training method of a battery capacity prediction model according to another embodiment of the present application;
fig. 5 is a schematic flow chart of a battery capacity prediction method according to an embodiment of the present application;
Fig. 6 is a schematic diagram of a training device for a battery capacity prediction model according to an embodiment of the present application;
fig. 7 is a schematic diagram of a battery capacity prediction apparatus according to an embodiment of the present application.
Detailed Description
Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar elements or elements having like or similar functions throughout. The embodiments described below by referring to the drawings are illustrative and intended to explain the present invention and should not be construed as limiting the invention.
The battery long-cycle test has a long test period, and can continuously occupy a test channel, and the tested battery is usually taken down and the test channel is replaced in the test process due to the consideration of test strategies, method feasibility and test cost control. The process can cause the capacity of the battery to be tested to rebound, so that huge fluctuation or bulge appears on a battery capacity attenuation curve (a fixing curve), the quality of battery capacity attenuation data is seriously reduced, the real attenuation trend of the battery capacity is covered, and the analysis based on the battery capacity attenuation data is not beneficial to evaluating the service life condition of the battery to be tested.
For optimization of a data curve, data smoothing can be performed on data points of the curve in some cases to obtain edge profiles of the data points, and then weighted average is performed on intersecting portions of adjacent data points in the edge profiles to obtain target values, so that deviation of the data is reduced to a certain extent. But the choice of weights will affect the adaptivity of this approach. Therefore, there is a certain disadvantage in optimizing the battery capacity fade curve in this manner.
In some cases, when the state of health of the battery is estimated, the battery aging observation equation can be constructed by performing data fitting on the voltage and the capacity of the battery to obtain the equivalent end voltage rebound and capacity mapping relationship after the charging is completed. And then extracting historical capacity data and cycle times of the battery to be estimated, and constructing a state transition equation of battery aging according to the historical capacity data and cycle times. Then, a state space model representing battery aging is obtained by fusing a state transition equation and an observation equation of battery aging, and the state space model is used for estimating the health state of the lithium ion battery and corresponding probability distribution. However, this method does not sufficiently consider the influence of the battery capacity rebound, which causes a large fluctuation or protrusion of the battery capacity fading curve, when performing data fitting, resulting in insufficient accuracy of the model. In addition, the mode needs to construct a state transition equation and an observation equation of battery aging, and the calculation process is complicated.
In view of this, the embodiments of the present application provide a training method of an optimized battery capacity prediction model and a battery capacity prediction method. The system architecture shown in fig. 1A may be used to implement a training method and a battery capacity prediction method of a battery capacity prediction model.
As shown in fig. 1A, system architecture 10 includes an execution device 110, a training device 120, a database 130, a client device 140, a data storage system 150, and a data acquisition device 160. The execution device 110 includes a calculation module 111 and an I/O interface 112, the calculation module 111 including the trained battery capacity prediction model 101.
The data acquisition device 160 is configured to acquire capacity sequence data of the battery and store the capacity sequence data in the database 130, wherein the capacity sequence data characterizes a relationship between a capacity of the battery and a number of charging cycles; the training device 120 stores therein a battery capacity prediction model to be trained, and the training device 120 performs model training based on the capacity sequence data of the battery maintained in the database 130, generating a trained battery capacity prediction model 101. Specifically, the training device 120 extracts at least one sub-sequence data from the capacity sequence data as sample data based on the data variation characteristics of the capacity sequence data; inputting the charge cycle times in the sample data into a battery capacity prediction model to be trained, and outputting a predicted value of the battery capacity; and then, based on the predicted value of the battery capacity and the loss value between the sample values of the battery capacity in the sample data, adjusting the model parameters of the battery capacity prediction model to be trained to obtain a trained battery capacity prediction model 101.
The trained battery state prediction model 101 obtained by the training device 120 may be applied in different systems or devices, for example in the execution device 110.
The execution device 110 is configured with an I/O interface 112, through which data interaction with external devices, including, for example, a client device 140, is performed via the I/O interface 112. The user may input the number of charge cycles of the battery to be predicted to the I/O interface 112 via the client device 140.
The execution device 110 may call data, code, etc. in the data storage system 150, or may store data, instructions, etc. in the data storage system 150.
The calculation module 111 stores the trained battery capacity prediction model 101, and the calculation module 111 predicts the battery capacity based on the input number of charge cycles using the trained battery capacity prediction model 101.
Next, I/O interface 112 returns the predicted results to client device 140 for presentation to the user.
In the case shown in fig. 1A, the user may manually specify the number of charge cycles of the battery of the input execution device 110, for example, to execute an input operation in an interface provided by the I/O interface 112. In another case, the client device 140 may automatically input the number of charging cycles of the battery to the I/O interface 112, for example, the number of charging cycles of the battery is automatically input after the client device 140 obtains the authorization of the user, and the user may set the corresponding authorization in the client device 140. The user may view the battery capacity output by the execution device 110 at the client device 140, and the specific presentation may be in the form of a display, sound, action, etc. The client device 140 may also be used as a data collection terminal to store the collected capacity sequence data of the battery into the database 130.
It should be noted that fig. 1A is only a schematic diagram of a system architecture provided in the embodiments of the present application, and the positional relationship among devices, apparatuses, modules, etc. shown in the drawings is not limited in any way, for example, in fig. 1A, the data storage system 150 is an external memory with respect to the execution device 110, and in other cases, the data storage system 150 may also be disposed in the execution device 110.
Fig. 1B is a flowchart of a training method of a battery capacity prediction model according to an embodiment of the present application.
As shown in fig. 1B, the training method 100 of the first battery capacity prediction model provided in the embodiment of the present application includes, for example, steps S110 to S140.
Step S110, acquiring capacity sequence data of the battery.
Illustratively, the capacity sequence data characterizes a relationship between battery capacity and a number of charge cycles, the capacity sequence data including capacity fade sequence data. The number of battery charging cycles means that the battery completes a complete charging cycle, and a complete charging cycle includes a complete charging and discharging process. When a complete charge cycle is completed, the number of battery charge cycles is increased by 1. The number of battery charging cycles may also be referred to as the number of cycles.
The battery can comprise a battery core, when the capacity of the battery is tested, the capacity of the battery under the number of the charging cycles is recorded every time a complete charging cycle is completed, and the capacity attenuation sequence data can be obtained by performing multiple tests.
In one example, the capacity fade sequence data includes a capacity fade curve representing a relationship between battery capacity and a number of charge cycles. Generally, the capacity fade curve should theoretically be a decreasing curve because the battery capacity decreases with increasing charge cycles due to aging of the battery. However, the capacity of the battery to be tested is rebounded due to the influence of other factors such as removing the battery to be tested and replacing the test channel in the test process, so that huge fluctuation or bulge is shown on the capacity attenuation curve. Therefore, it is difficult to accurately know the true relationship between the battery capacity and the number of charge cycles directly based on the capacity fade curve obtained by the test, resulting in difficulty in accurately predicting the state of health of the battery.
Step S120 extracts at least one sub-sequence data from the capacity sequence data as sample data based on the data variation characteristics of the capacity sequence data.
Illustratively, the data change characteristic includes a characteristic in which the battery capacity changes with the change in the number of charging cycles, and based on the data change characteristic of the capacity fade sequence data, it can be known which part of the data in the capacity fade sequence data has a large fluctuation or protrusion. At least one sub-sequence data extracted from the capacity fade sequence data is, for example, data in which the degree of fluctuation or the degree of protrusion is small in the capacity fade sequence data, and it can be seen that the extracted sub-sequence data is an effective portion in which no capacity rebound exists in the capacity fade sequence data. In an example, sub-sequence data with a data change amplitude value smaller than a preset amplitude threshold value can be determined as sample data from the content decay sequence data based on the data change characteristics, the preset amplitude threshold value can be specifically set according to practical situations, and the data change amplitude value greater than or equal to the preset amplitude threshold value indicates that the data change is severe and the probability is abnormal, so that the part of data needs to be removed, and the situation that the model training effect is poor or the model is over-fitted due to the fact that the data change is severe as the sample data is avoided. The subsequence data with the data change amplitude value smaller than the preset amplitude threshold value is smooth-change data, the smooth-change subsequence data can reflect the real relation between the battery capacity and the charging cycle number to a large extent, and the smooth-change subsequence data is used as a training sample for model training, so that the accuracy of a model can be improved. The sample data comprises a plurality of data pairs, and each data pair comprises a charging cycle number and a corresponding battery capacity, and represents the battery capacity of the battery under the charging cycle number.
Step S130, the charging cycle times in the sample data are input into a battery capacity prediction model to be trained, and a predicted value of the battery capacity is output.
Step S140, adjusting model parameters of the battery capacity prediction model to be trained based on the predicted value of the battery capacity and the loss value between the sample values of the battery capacity in the sample data, so as to obtain a trained battery capacity prediction model.
Illustratively, the number of charge cycles in the sample data is taken as the input of the model, and the sample value of the battery capacity in the sample data is taken as the label. When the battery capacity prediction model to be trained is trained by using the sample data, the charging cycle times are input into the battery capacity prediction model to be trained for prediction, and the prediction result output by the model is the prediction value of the battery capacity. Comparing the output battery capacity predicted value with a sample value (label) of the battery capacity in sample data to obtain a loss value of the battery capacity predicted value and the sample value, reversely adjusting model parameters of a battery capacity predicted model to be trained based on the loss value, performing iterative calculation for a plurality of times to enable the deviation between the output of the model and the label to be smaller and smaller until the deviation is smaller than a preset deviation value or the iteration number reaches a preset number, and stopping training to obtain the trained battery capacity predicted model. The trained battery capacity prediction model can more accurately represent the relationship between battery capacity and the number of charging cycles.
It can be appreciated that, according to the embodiment of the application, the sub-sequence data is extracted from the capacity fading sequence data based on the data change characteristics as the sample data of model training, and the extracted sub-sequence data can accurately reflect the relationship between the battery capacity and the charging cycle number to a large extent. Model training is performed by using the extracted subsequence data, so that a trained battery capacity prediction model can accurately reflect the relationship between battery capacity and charging cycle times. In addition, the extracted sub-sequence data is a data segment, and the data segment contains a part of the battery capacity corresponding to the number of charging cycles, so that the sub-sequence data reflects the battery capacity under the limited (part of) charging cycles, that is, the extracted sub-sequence data has data missing. The trained battery capacity prediction model can complement missing data, namely the trained battery capacity prediction model can reflect the battery capacity under any charging cycle times, so that the trained battery capacity prediction model can completely reflect the relationship between the battery capacity and the charging cycle times, the prediction requirement of the battery capacity is met, and the prediction accuracy of the battery health state is improved.
In an example of the present application, the battery capacity prediction Model to be trained may select a Model suitable for processing the sequence data, and the battery capacity prediction Model to be trained is, for example, a machine learning Model, and may specifically include a Regression Model (Regression Model). Regression models include, for example, polynomial regression (Polynomial Regression) models, gaussian process regression (Gaussian Process Regression, GPR) models, and the like. Embodiments of the present application are described with respect to a gaussian process regression model.
The capacity fade sequence data may be preprocessed before extracting at least one sub-sequence data from the capacity fade sequence data based on a data variation characteristic of the capacity fade sequence data. Preprocessing includes data cleansing by which outliers in the capacity fade sequence data can be removed to reduce the impact of outliers on model training.
Illustratively, the outliers in the capacity fade sequence data include outliers that differ significantly from most values. For example, the capacity fading sequence data may be detected by an anomaly detection method to identify an anomaly value therein, or may be smoothed by a data smoothing method. The abnormality detection method includes, for example, a box-diagram abnormality detection method, but may include other abnormality detection methods.
Illustratively, performing abnormality detection based on the box-diagram abnormality detection scheme includes: sequencing the values of the battery capacities in the capacity sequence data in an ascending order; determining a first value, a second value and a third value from the sequenced values; determining a first edge value based on the difference IQR between the third value and the first value, and a first preset coefficient; determining a second edge value based on the difference value IQR between the third value and the first value, the third value and a second preset coefficient; determining data of which the value of the battery capacity is smaller than a first edge value or larger than a second edge value in the capacity sequence data as abnormal data; and removing the abnormal data. It can be understood that, each time of abnormality detection, a segment of sub-sequence data of the capacity sequence data can be input into the box graph abnormality detection algorithm, and the abnormality detection for the capacity sequence data can be completed by performing multiple calculations.
Specifically, the first value is smaller than the second value, the second value is smaller than the third value, the first value is 25% of the ordered battery capacity values (the values are ordered from small to large, the first value is greater than or equal to 25%), the second value is the median of the ordered battery capacity values, the third value is 75% of the ordered battery capacity values (the values are ordered from small to large, and the third value is greater than or equal to 75%). The first preset coefficient and the second preset coefficient are both 1.5, for example, and may of course be specifically set according to practical application situations. The first edge value is a lower boundary, the first edge value is, for example, a first value minus 1.5×iqr, the second edge value is an upper boundary, and the second edge value is, for example, a third value plus 1.5×iqr. The box graph anomaly detection scheme describes the overall distribution of data based on statistics of median, 25% quantile, 75% quantile, upper boundary, lower boundary, etc. In the embodiment of the application, the capacity fading sequence data is processed by using a box graph anomaly detection mode, an outlier can be determined from the capacity fading sequence data, and the outlier is removed as an anomaly value, so that the smoothing processing of the capacity fading sequence data is realized.
After removing the abnormal value in the capacity-fading sequence data, at least one sub-sequence data needs to be extracted from the capacity-fading sequence data as sample data based on the data change characteristic of the capacity-fading sequence data, as shown in fig. 2.
Fig. 2 is a schematic diagram of acquiring sample data according to an embodiment of the present application.
As shown in fig. 2, the thin curve segments represent raw capacity fade sequence data, the abscissa represents the number of charge cycles (number of cycles), and the ordinate represents battery capacity. Each peak in the curve segment includes a rebound peak formed by the rebound of battery capacity.
When extracting valid sub-sequence data (valid data segment), at least one fluctuation sub-sequence data, which characterizes abnormal change of battery capacity with change of charge cycle number, including a rebound peak, may be determined from the capacity-fading sequence data based on data change characteristics of the capacity-fading sequence data.
After each fluctuating sub-sequence data is identified, sub-sequence data adjacent to the fluctuating sub-sequence data is extracted from the content decay sequence data to obtain at least one sub-sequence data as sample data. In one example, the subsequence data adjacent to the fluctuating subsequence data includes subsequence data prior to the occurrence of fluctuation in the data in the capacity fade sequence data. As shown in fig. 2, at least one extracted sub-sequence data is represented by a thick curve segment, and it can be seen that each extracted sub-sequence data is a segment of valid data before a certain rebound peak.
It can be understood that the data before the battery capacity rebounds is an effective data segment, and the effective data segment can accurately represent the real relationship between the battery capacity and the cycle number, so that the effective data segment is extracted from the capacity fading sequence data as sample data to perform model training, the accuracy of the model can be improved, and the trained battery capacity prediction model can accurately reflect the relationship between the battery capacity and the charging cycle number.
In another example, not every sub-sequence data before a fluctuation or a bump is a valid data segment, and thus at least one fluctuation sub-sequence data (valid data segment) needs to be determined from the capacity fade sequence data based on a preset rule. The preset rule indicates that the wave subsequence data is determined based on the preset variation trend and the wave subsequence data is determined based on the number of charging cycles at certain intervals at the same time.
First, it is necessary to determine a plurality of fluctuation subsequence data based on a preset variation trend, so that the variation trend of the battery capacity with the number of charging cycles in any one of the determined fluctuation subsequence data is a preset variation trend. For example, the fluctuation subsequence data is characterized by a function having the number of charge cycles as an independent variable and the battery capacity as a dependent variable. The change trend of the battery capacity along with the number of charging cycles is a preset change trend, and the preset change trend comprises that the partial derivative value of the function corresponding to the number of charging cycles is larger than a preset partial derivative value, for example, 0. For example, a partial derivative of the function greater than 0 indicates that the capacity fade sequence data (curve) begins to curve and a surge condition or bulge occurs. It can be seen that in order to improve accuracy, it is also necessary to determine the fluctuation subsequence data based on a preset trend of variation.
Secondly, when at least one wave subsequence data includes a plurality of wave subsequence data, the wave subsequence data is determined at intervals of a certain number of charging cycles, so that the number difference between the determined number of charging cycles respectively corresponding to two adjacent wave subsequence data is larger than a preset difference, for example, 200 times, and the preset difference can be specifically set according to practical situations. As shown in fig. 2, the difference in the number of charge cycles between the respective corresponding adjacent two peaks (3 rd and 4 th peaks from left to right) that are particularly close to each other is smaller than the preset difference, and therefore one (e.g., the latter) of the adjacent two peaks that are particularly close to each other is an abnormal peak, for example, the 4 th peak is an abnormal peak, which is not the desired fluctuation subsequence data, i.e., which is not a normal rebound peak.
In some cases, if there is only one rebound peak, the rebound peak can be determined to be fluctuating sub-sequence data.
It can be understood that the embodiments of the present application interval a certain number of charging cycles and determine the fluctuation subsequence data based on a preset variation trend, so that the determined fluctuation subsequence data has a high probability of springback of the battery capacity caused by replacing the test channel during the test, rather than abnormal springback caused by other abnormal reasons. And sub-sequence data before the fluctuation sub-sequence data is used as sample data for model training, so that the accuracy of the sample data is improved.
Fig. 3 is a schematic diagram of a data attenuation curve according to an embodiment of the present application.
As shown in fig. 3, the trained battery capacity prediction model trained using the sample data reflects the relationship between the number of charge cycles and the battery capacity. After the trained battery capacity prediction model is obtained, the charge cycle times (such as the value corresponding to the abscissa in fig. 3) in the capacity fading sequence data are input into the trained battery capacity prediction model, updated battery capacity is output, a data fading curve is output based on the updated battery capacity and the charge cycle times, the data fading curve represents the relationship between the updated battery capacity and the charge cycle times, and the data fading curve (fitting data) is shown as a thick curve segment in fig. 3. Therefore, the coincidence degree between the data attenuation curve and the effective data segment is higher, the accuracy of the trained battery capacity prediction model is higher, the data attenuation curve complements the rest parts except the effective data segment, and the relationship between the battery capacity and the charging cycle times can be completely reflected by the data attenuation curve.
Fig. 4 is a flowchart of a training method of a battery capacity prediction model according to another embodiment of the present application.
As shown in fig. 4, the training method 400 of the second battery capacity prediction model provided in the embodiment of the present application includes, for example, steps S410 to S450.
In step S410, original capacity fade sequence data is acquired.
Step S420, removing outliers in the original capacity fade sequence data.
For example, the original capacity fade sequence data is processed through a box plot to remove outliers in the capacity fade sequence data.
In step S430, the valid data segment is extracted as a training sample.
For example, after removing the outliers, a valid data segment before extracting the fluctuation subsequence data from the content decay sequence data is used as a training sample.
Step S440, training a battery capacity prediction model to be trained by using the training samples.
Step S450, outputting a data decay curve using the trained battery capacity prediction model.
For example, the number of charge cycles in the capacity fade sequence data is input into a trained battery capacity prediction model, an updated battery capacity is output, and a data fade curve is output based on the updated battery capacity and the number of charge cycles.
It will be appreciated that the specific implementation of the embodiments of the present application may refer to the above mentioned content, and will not be described herein.
The tested batteries have different designs and different test schemes, so that the distribution of rebound peaks of battery capacity on an attenuation curve is different when different batteries are tested. The trained battery capacity prediction model obtained through the method has strong self-adaption force, the model learning and using cost is low, and the battery capacity prediction cost is reduced.
According to the embodiment of the application, the effective data segment is extracted by combining a data cleaning (abnormal value is removed from the box-shaped graph) mode and a mode of identifying the rebound peak based on a certain cycle number interval and curve derivative change condition, so that the data quality of the effective data segment is improved. And curve fitting is performed by using a Gaussian process regression model, so that curve data parts corresponding to rebound peaks are supplemented, and the fitting effect is improved. Therefore, a life attenuation curve corresponding to a smooth battery long-cycle test can be effectively obtained by fitting by using a trained battery capacity prediction model, and the fitted curve can reflect the life attenuation trend of the tested battery core, so that the battery capacity prediction model has good self-adaption and generalization performances.
The data processing process of the present application can be implemented based on the Python language, which is a computer language that provides an efficient, high-level data structure that can also be easily and efficiently programmed towards objects. The effective data segment in the tested battery long-cycle test data is identified, the Gaussian process regression model is trained by using the effective data segment, and finally the trained Gaussian process regression model is utilized to output a final attenuation smooth curve, so that the accuracy of the attenuation smooth curve is improved.
In constructing the battery capacity prediction model, the model may be constructed based on an open source machine learning framework such as Sklearn, which is a general learning library for machine learning. And training the constructed battery capacity prediction model by inputting the extracted effective data segment, inputting the actual cycle number of the battery to be tested into the trained battery capacity prediction model, reasoning the complete attenuation curve of the battery by the trained battery capacity prediction model, and outputting a final result.
Fig. 5 is a flowchart of a battery capacity prediction method according to an embodiment of the present application.
As shown in fig. 5, the battery capacity prediction method 500 provided in the embodiment of the present application includes steps S510 to S520, for example.
Step S510, the number of charging cycles of the battery is obtained.
Step S520, the number of charge cycles is input into the trained battery capacity prediction model, and the battery capacity is predicted by the trained battery capacity prediction model.
For example, a battery capacity may be predicted using a trained battery capacity prediction model to enable prediction of battery capacity. For example, if the current charge cycle number of the battery is known in the process of actually using the battery, the charge cycle number can be input into a trained battery capacity prediction model, and the battery capacity under the current charge cycle number is predicted by the battery capacity prediction model, so that whether the battery is aged or not and other health states can be known in time. The trained battery capacity prediction model may be trained using the model training method mentioned above.
Fig. 6 is a schematic diagram of a training device for a battery capacity prediction model according to an embodiment of the present application.
Referring to fig. 6, a training apparatus 600 for a battery capacity prediction model is provided, and the training apparatus 600 for a battery capacity prediction model includes: a first acquisition module 610, an extraction module 620, an output module 630, and an adjustment module 640.
Illustratively, the first acquisition module 610 is configured to acquire capacity sequence data of the battery, wherein the capacity sequence data characterizes a relationship between a capacity of the battery and a number of charging cycles.
Illustratively, the extraction module 620 is configured to extract at least one sub-sequence data from the capacity sequence data as sample data based on a data variation characteristic of the capacity sequence data.
The output module 630 is used for inputting the number of charging cycles in the sample data into a battery capacity prediction model to be trained, and outputting a predicted value of the battery capacity.
Illustratively, the adjustment module 640 is configured to adjust model parameters of the battery capacity prediction model to be trained based on a predicted value of the battery capacity and a loss value between sample values of the battery capacity in the sample data to obtain a trained battery capacity prediction model.
It will be appreciated that for a specific description of the training apparatus 600 of the battery capacity prediction model, reference may be made to the description of the method of training the battery capacity prediction model hereinabove.
Illustratively, the extraction module is specifically configured to: determining at least one fluctuation subsequence data from the capacity sequence data based on a data change characteristic of the capacity sequence data, wherein the fluctuation subsequence data characterizes an abnormal change in battery capacity with a change in charge cycle number; based on each fluctuating sub-sequence data, sub-sequence data adjacent to the fluctuating sub-sequence data is extracted from the container sequence data to obtain at least one sub-sequence data as sample data.
The at least one wave subsequence data includes a plurality of wave subsequences, a variation trend of a battery capacity in any wave subsequence data along with a charging cycle number is a preset variation trend, and a number difference between charging cycle numbers respectively corresponding to two adjacent wave subsequence data is larger than a preset difference.
Illustratively, the fluctuation subsequence data is characterized by a function having a number of charge cycles as an argument and a battery capacity as a dependent variable; the changing trend of the battery capacity along with the charging cycle times is a preset changing trend comprising the following steps: the partial derivative value of the function corresponding to the charging cycle number is larger than the preset partial derivative value.
Illustratively, the subsequence data adjacent to the fluctuating subsequence data includes: sub-sequence data before fluctuation of the data in the capacity sequence data.
Illustratively, the training apparatus 600 of the battery capacity prediction model may further include: and a removal module for removing outliers in the capacity sequence data before extracting at least one sub-sequence data from the capacity sequence data as sample data based on the data variation characteristics of the capacity sequence data.
Illustratively, the removal module is specifically configured to: sequencing the values of the battery capacities in the capacity sequence data in an ascending order; determining a first value, a second value and a third value from the sequenced values, wherein the first value is smaller than the second value, the second value is smaller than the third value, and the second value is the median of the values of the battery capacity; determining a first edge value based on a difference between the third value and the first value, and a first preset coefficient; determining a second edge value based on the difference between the third value and the first value, the third value, and a second preset coefficient; determining data of which the value of the battery capacity is smaller than a first edge value or larger than a second edge value in the capacity sequence data as abnormal data; and removing the abnormal data.
Illustratively, the training apparatus 600 of the battery capacity prediction model may further include: a first additional output module and a second additional output module. The first additional output module is used for inputting the charge cycle times in the capacity sequence data into the trained battery capacity prediction model and outputting updated battery capacity; and a second additional output module for outputting a data curve based on the updated battery capacity and the number of charge cycles, wherein the data curve characterizes a relationship between the updated battery capacity and the number of charge cycles.
Illustratively, the battery capacity prediction model to be trained includes a regression model.
Fig. 7 is a schematic diagram of a battery capacity prediction apparatus according to an embodiment of the present application.
Referring to fig. 7, a battery capacity prediction apparatus 700 according to an embodiment of the present application includes: a second acquisition module 710 and a prediction module 720.
Illustratively, the second acquisition module 710 is configured to acquire a number of charge cycles of the battery.
Illustratively, the prediction module 720 is configured to input the number of charge cycles into a trained battery capacity prediction model from which the battery capacity is predicted.
It is understood that, for the specific description of the battery capacity prediction apparatus 700, reference may be made to the above description of the battery capacity prediction method, and the detailed description thereof will not be repeated.
An embodiment of the present application provides an electronic device, including a memory storing a computer program and a processor, where the processor executes the computer program to implement the steps of the method in any of the above embodiments.
Embodiments of the present application provide a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method of any of the above embodiments.
An embodiment of the present application provides a computer program product comprising instructions which, when executed by a processor of a computer device, enable the computer device to perform the steps of the method of any one of the embodiments described above.
It should be noted that the logic and/or steps represented in the flowcharts or otherwise described herein, for example, may be considered as a ordered listing of executable instructions for implementing logical functions, and may be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium may even be paper or other suitable medium upon which the program is printed, as the program may be electronically captured, via, for instance, optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner, if necessary, and then stored in a computer memory.
It is to be understood that portions of the present application may be implemented in hardware, software, firmware, or a combination thereof. In the above-described embodiments, the various steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, may be implemented using any one or combination of the following techniques, as is well known in the art: discrete logic circuits having logic gates for implementing logic functions on data signals, application specific integrated circuits having suitable combinational logic gates, programmable Gate Arrays (PGAs), field Programmable Gate Arrays (FPGAs), and the like.
In the description of the present application, a description referring to terms "one embodiment," "some embodiments," "examples," "specific examples," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present application. In this application, schematic representations of the above terms do not necessarily refer to the same embodiments or examples. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
In the description of the present application, it should be understood that the terms "center," "longitudinal," "transverse," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," "circumferential," etc. indicate orientations or positional relationships based on the orientations or positional relationships illustrated in the drawings, are merely for convenience in describing the present application and simplifying the description, and do not indicate or imply that the device or element being referred to must have a particular orientation, be configured and operated in a particular orientation, and therefore should not be construed as limiting the present application.
Furthermore, the terms "first," "second," and the like, as used in embodiments of the present application, are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated in the present embodiment. Thus, a feature of an embodiment described herein that is termed a "first," "second," etc., may explicitly or implicitly indicate that at least one such feature is included in the embodiment. In the description of the present application, the word "plurality" means at least two or more, for example, two, three, four, etc., unless explicitly defined otherwise in the embodiments.
In this application, unless explicitly stated or limited otherwise in the examples, the terms "mounted," "connected," and "fixed" as used in the examples should be interpreted broadly, e.g., the connection may be a fixed connection, may be a removable connection, or may be integral, and it may be understood that the connection may also be a mechanical connection, an electrical connection, etc.; of course, it may be directly connected, or indirectly connected through an intermediate medium, or may be in communication with each other, or in interaction with each other. The specific meaning of the terms in this application will be understood by those of ordinary skill in the art depending on the specific implementation.
In this application, unless expressly stated or limited otherwise, a first feature "up" or "down" a second feature may be the first and second features in direct contact, or the first and second features in indirect contact via an intervening medium. Moreover, a first feature being "above," "over" and "on" a second feature may be a first feature being directly above or obliquely above the second feature, or simply indicating that the first feature is level higher than the second feature. The first feature being "under", "below" and "beneath" the second feature may be the first feature being directly under or obliquely below the second feature, or simply indicating that the first feature is less level than the second feature.
Although embodiments of the present application have been shown and described above, it will be understood that the above embodiments are illustrative and not to be construed as limiting the application, and that variations, modifications, alternatives, and variations may be made to the above embodiments by one of ordinary skill in the art within the scope of the application.
Claims (12)
1. A method of training a battery capacity prediction model, the method comprising:
acquiring capacity sequence data of a battery, wherein the capacity sequence data characterizes the relationship between the capacity of the battery and the number of charging cycles;
extracting at least one sub-sequence data from the capacity sequence data as sample data based on a data variation characteristic of the capacity sequence data, comprising: determining a plurality of fluctuation subsequence data from the capacity sequence data, wherein the change trend of the battery capacity in any fluctuation subsequence data along with the charging cycle times is a preset change trend, and the frequency difference value between the charging cycle times respectively corresponding to two adjacent fluctuation subsequence data is larger than the preset difference value; extracting, as sample data, the at least one subsequence data before fluctuation of data occurs from the capacity sequence data based on the plurality of wave subsequences data;
Inputting the charge cycle times in the sample data into a battery capacity prediction model to be trained, and outputting a predicted value of the battery capacity; and
and adjusting model parameters of the battery capacity prediction model to be trained based on a predicted value of the battery capacity and a loss value between sample values of the battery capacity in the sample data to obtain a trained battery capacity prediction model.
2. The method of claim 1, wherein the step of determining the position of the substrate comprises,
the determining a plurality of fluctuation subsequences from the capacity sequence data includes: determining the plurality of wave subsequences of data from the capacity sequence data based on data change characteristics of the capacity sequence data, wherein the wave subsequences of data characterize abnormal changes in battery capacity with changes in charge cycle number; and
the extracting, based on the plurality of wave subsequences, the at least one subsequence before the fluctuation of the data occurs from the capacity sequence data as sample data, includes: based on each fluctuating sub-sequence data, sub-sequence data adjacent to the fluctuating sub-sequence data is extracted from the capacity sequence data to obtain the at least one sub-sequence data as the sample data.
3. The method according to claim 2, characterized in that:
the fluctuation subsequence data is characterized by a function, wherein the function takes the number of charging cycles as an independent variable and the battery capacity as a dependent variable;
the changing trend of the battery capacity along with the charging cycle times is a preset changing trend comprising the following steps: the partial derivative value of the function corresponding to the charging cycle number is larger than a preset partial derivative value.
4. The method of claim 2, wherein the subsequence data adjacent to the fluctuating subsequence data comprises: sub-sequence data before fluctuation of the data in the capacity sequence data occurs.
5. The method according to any one of claims 1-4, further comprising, before extracting at least one sub-sequence data from the capacity sequence data as sample data based on a data change characteristic of the capacity sequence data:
sequencing the values of the battery capacities in the capacity sequence data in an ascending order;
determining a first value, a second value and a third value from the sequenced values, wherein the first value is smaller than the second value, the second value is smaller than the third value, and the second value is the median of the values of the battery capacity;
Determining a first edge value based on a difference between the third value and the first value, and a first preset coefficient;
determining a second edge value based on a difference between the third value and the first value, the third value, and a second preset coefficient;
determining data of which the value of the battery capacity is smaller than the first edge value or larger than the second edge value in the capacity sequence data as abnormal data; and
and removing the abnormal data.
6. The method according to any one of claims 1-4, further comprising:
inputting the charge cycle times in the capacity sequence data into a trained battery capacity prediction model, and outputting updated battery capacity; and
based on the updated battery capacity and the number of charge cycles, a data curve is output, wherein the data curve characterizes a relationship between the updated battery capacity and the number of charge cycles.
7. The method of any one of claims 1-4, wherein the battery capacity prediction model to be trained comprises a regression model.
8. A battery capacity prediction method, the method comprising:
Acquiring the charge cycle times of a battery; and
inputting the number of charge cycles into a trained battery capacity prediction model, predicting battery capacity from the trained battery capacity prediction model,
wherein the trained battery capacity prediction model is trained using the method according to any one of claims 1-7.
9. A training device for a battery capacity prediction model, the device comprising:
the system comprises an acquisition module, a storage module and a storage module, wherein the acquisition module is used for acquiring capacity sequence data of a battery, and the capacity sequence data represents the relationship between the capacity of the battery and the number of charging cycles;
an extraction module for extracting at least one sub-sequence data from the capacity sequence data as sample data based on a data variation characteristic of the capacity sequence data, comprising: determining a plurality of fluctuation subsequence data from the capacity sequence data, wherein the change trend of the battery capacity in any fluctuation subsequence data along with the charging cycle times is a preset change trend, and the frequency difference value between the charging cycle times respectively corresponding to two adjacent fluctuation subsequence data is larger than the preset difference value; extracting, as sample data, the at least one subsequence data before fluctuation of data occurs from the capacity sequence data based on the plurality of wave subsequences data;
The output module is used for inputting the charge cycle times in the sample data into a battery capacity prediction model to be trained and outputting a predicted value of the battery capacity; and
and the adjusting module is used for adjusting the model parameters of the battery capacity prediction model to be trained based on the predicted value of the battery capacity and the loss value between the sample values of the battery capacity in the sample data so as to obtain a trained battery capacity prediction model.
10. A battery capacity prediction apparatus, characterized in that the apparatus comprises:
the acquisition module is used for acquiring the charging cycle times of the battery; and
a prediction module for inputting the number of charge cycles into a trained battery capacity prediction model, predicting a battery capacity from the trained battery capacity prediction model,
wherein the trained battery capacity prediction model is trained using the apparatus of claim 9.
11. An electronic device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1-8 when the computer program is executed.
12. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311318555.6A CN117054893B (en) | 2023-10-12 | 2023-10-12 | Training method of battery capacity prediction model, battery capacity prediction method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311318555.6A CN117054893B (en) | 2023-10-12 | 2023-10-12 | Training method of battery capacity prediction model, battery capacity prediction method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117054893A CN117054893A (en) | 2023-11-14 |
| CN117054893B true CN117054893B (en) | 2024-03-29 |
Family
ID=88663127
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311318555.6A Active CN117054893B (en) | 2023-10-12 | 2023-10-12 | Training method of battery capacity prediction model, battery capacity prediction method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117054893B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118625154A (en) * | 2024-08-09 | 2024-09-10 | 长江三峡集团实业发展(北京)有限公司 | Battery remaining life prediction method, device, computer equipment and storage medium |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109031153A (en) * | 2018-10-16 | 2018-12-18 | 北京交通大学 | A kind of health status On-line Estimation method of lithium ion battery |
| CN111490301A (en) * | 2020-03-17 | 2020-08-04 | 湖南立方新能源科技有限责任公司 | Method and system for judging battery failure remediation time |
| CN112213643A (en) * | 2020-09-30 | 2021-01-12 | 蜂巢能源科技有限公司 | Method, system and equipment for predicting initial capacity and health state of battery |
| CN112487702A (en) * | 2020-10-26 | 2021-03-12 | 湖州师范学院 | Method for predicting remaining service life of lithium ion battery |
| CN113640690A (en) * | 2021-06-24 | 2021-11-12 | 河南科技大学 | Method for predicting residual life of power battery of electric vehicle |
| CN114282169A (en) * | 2021-10-12 | 2022-04-05 | 腾讯科技（深圳）有限公司 | Abnormal data detection method and related device |
| CN114781262A (en) * | 2022-04-21 | 2022-07-22 | 哈尔滨工业大学 | Long-life lithium ion battery early life prediction method and system based on degradation sequence time sequence correlation modeling |
| CN115221795A (en) * | 2022-08-03 | 2022-10-21 | 一汽解放汽车有限公司 | Training method, prediction method, device, equipment and medium of capacity prediction model |
| CN115621573A (en) * | 2022-10-14 | 2023-01-17 | 佛山科学技术学院 | System and method for predicting health state of lithium battery of energy storage power station |
| CN116577668A (en) * | 2023-05-31 | 2023-08-11 | 徐州徐工矿业机械有限公司 | SAE-QPSO-RVM-based mining dump truck starting lead-acid storage battery life prediction method |
| CN116626502A (en) * | 2023-05-23 | 2023-08-22 | 南斗六星系统集成有限公司 | Battery capacity prediction method, device, equipment and storage medium |
| EP4231027A1 (en) * | 2020-10-15 | 2023-08-23 | Resonac Corporation | Lithium ion battery life span prediction method, discharge capacity retention rate prediction method, life span prediction program, discharge capacity retention rate prediction program, and information processing device |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11313894B2 (en) * | 2020-06-29 | 2022-04-26 | Intelematics Australia Pty Limited | Automobile battery failure prediction method and system |
-
2023
- 2023-10-12 CN CN202311318555.6A patent/CN117054893B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109031153A (en) * | 2018-10-16 | 2018-12-18 | 北京交通大学 | A kind of health status On-line Estimation method of lithium ion battery |
| CN111490301A (en) * | 2020-03-17 | 2020-08-04 | 湖南立方新能源科技有限责任公司 | Method and system for judging battery failure remediation time |
| CN112213643A (en) * | 2020-09-30 | 2021-01-12 | 蜂巢能源科技有限公司 | Method, system and equipment for predicting initial capacity and health state of battery |
| EP4231027A1 (en) * | 2020-10-15 | 2023-08-23 | Resonac Corporation | Lithium ion battery life span prediction method, discharge capacity retention rate prediction method, life span prediction program, discharge capacity retention rate prediction program, and information processing device |
| CN112487702A (en) * | 2020-10-26 | 2021-03-12 | 湖州师范学院 | Method for predicting remaining service life of lithium ion battery |
| CN113640690A (en) * | 2021-06-24 | 2021-11-12 | 河南科技大学 | Method for predicting residual life of power battery of electric vehicle |
| CN114282169A (en) * | 2021-10-12 | 2022-04-05 | 腾讯科技（深圳）有限公司 | Abnormal data detection method and related device |
| CN114781262A (en) * | 2022-04-21 | 2022-07-22 | 哈尔滨工业大学 | Long-life lithium ion battery early life prediction method and system based on degradation sequence time sequence correlation modeling |
| CN115221795A (en) * | 2022-08-03 | 2022-10-21 | 一汽解放汽车有限公司 | Training method, prediction method, device, equipment and medium of capacity prediction model |
| CN115621573A (en) * | 2022-10-14 | 2023-01-17 | 佛山科学技术学院 | System and method for predicting health state of lithium battery of energy storage power station |
| CN116626502A (en) * | 2023-05-23 | 2023-08-22 | 南斗六星系统集成有限公司 | Battery capacity prediction method, device, equipment and storage medium |
| CN116577668A (en) * | 2023-05-31 | 2023-08-11 | 徐州徐工矿业机械有限公司 | SAE-QPSO-RVM-based mining dump truck starting lead-acid storage battery life prediction method |
Non-Patent Citations (2)
| Title |
|---|
| "A method for capacity prediction of lithium-ion batteries under small sample conditions";Zhang, Meng 等;《Energy》;第238卷;正文第1-2页 * |
| "基于双高斯模型的锂电池剩余使用寿命预测方法";李彦梅 等;《 电气工程学报》;第17卷(第4期);正文第32-40页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117054893A (en) | 2023-11-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117054893B (en) | Training method of battery capacity prediction model, battery capacity prediction method and device | |
| US20230366936A1 (en) | Lithium ion battery lifetime prediction method, discharge capacity retention rate prediction method, lifetime prediction program, discharge capacity retention rate prediction program, and information processing device | |
| CN111639798A (en) | Intelligent prediction model selection method and device | |
| CN112085111B (en) | Load identification method and device | |
| CN113125095A (en) | Universal circuit breaker contact system residual mechanical life prediction method based on deep learning | |
| CN115440292B (en) | Method, device and equipment for testing memory chip and storage medium | |
| JP2023017480A (en) | Battery model construction method and battery degradation prediction device | |
| CN115408864B (en) | Electronic transformer error state self-adaptive prediction method, system and equipment | |
| CN111881023A (en) | Software aging prediction method and device based on multi-model comparison | |
| CN116643178B (en) | SOC estimation method and related device of battery management system | |
| CN117150276B (en) | Machine learning model construction method, vehicle driving risk prediction method and device | |
| CN116739829B (en) | Big data-based power data analysis method, system and medium | |
| CN116559667A (en) | Model training method and device, battery detection method and device, equipment and medium | |
| CN117972358A (en) | AI-based power distribution cabinet state identification method, device, equipment and storage medium | |
| CN116736166A (en) | Battery cell abnormality detection method and device of battery pack and battery pack | |
| JP2003044123A (en) | Plant diagnostic device | |
| CN112664410B (en) | Big data-based modeling method for unit online monitoring system | |
| CN117825970A (en) | Battery degradation analysis method, device, equipment and storage medium | |
| CN117783872A (en) | Power battery data abnormality detection method, device, server and storage medium | |
| CN113537693A (en) | Personnel risk level obtaining method, terminal and storage device | |
| CN116413606A (en) | Power battery health state estimation method, system and medium | |
| CN115544803A (en) | Method, device, equipment and storage medium for predicting residual life of transformer | |
| CN114840983A (en) | Power storage battery SOH calculation model construction method and calculation method | |
| CN113283157A (en) | System, method, terminal and medium for predicting life cycle of intelligent stamping press part | |
| CN117829002B (en) | Aging diagnosis monitoring method and system for power cable |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |