CN116401954A - Prediction method, prediction device, equipment and medium for cycle life of lithium battery - Google Patents
Prediction method, prediction device, equipment and medium for cycle life of lithium battery Download PDFInfo
- Publication number
- CN116401954A CN116401954A CN202310439970.0A CN202310439970A CN116401954A CN 116401954 A CN116401954 A CN 116401954A CN 202310439970 A CN202310439970 A CN 202310439970A CN 116401954 A CN116401954 A CN 116401954A
- Authority
- CN
- China
- Prior art keywords
- lithium battery
- target
- curve
- data
- determining
- 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.)
- Pending
Links
- WHXSMMKQMYFTQS-UHFFFAOYSA-N Lithium Chemical compound [Li] WHXSMMKQMYFTQS-UHFFFAOYSA-N 0.000 title claims abstract description 188
- 229910052744 lithium Inorganic materials 0.000 title claims abstract description 188
- 238000000034 method Methods 0.000 title claims abstract description 62
- 230000015556 catabolic process Effects 0.000 claims abstract description 30
- 238000006731 degradation reaction Methods 0.000 claims abstract description 30
- 238000004140 cleaning Methods 0.000 claims abstract description 23
- 238000012216 screening Methods 0.000 claims description 25
- 238000012545 processing Methods 0.000 claims description 22
- 238000012549 training Methods 0.000 claims description 22
- 238000012360 testing method Methods 0.000 claims description 11
- 238000013075 data extraction Methods 0.000 claims description 9
- 238000002790 cross-validation Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 5
- 230000008569 process Effects 0.000 description 9
- 230000002159 abnormal effect Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 238000012417 linear regression Methods 0.000 description 6
- 238000011156 evaluation Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 230000032683 aging Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 238000007637 random forest analysis Methods 0.000 description 2
- 238000012827 research and development Methods 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 238000010200 validation analysis Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000002301 combined effect Effects 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000007786 learning performance Effects 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
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/04—Ageing analysis or optimisation against ageing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
- Y02E60/10—Energy storage using batteries
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Mathematical Physics (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Pure & Applied Mathematics (AREA)
- Computational Mathematics (AREA)
- Mathematical Optimization (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Mathematical Analysis (AREA)
- Artificial Intelligence (AREA)
- Computer Hardware Design (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Quality & Reliability (AREA)
- Computing Systems (AREA)
- Operations Research (AREA)
- Probability & Statistics with Applications (AREA)
- Geometry (AREA)
- Algebra (AREA)
- Secondary Cells (AREA)
Abstract
The application provides a prediction method, a prediction device, equipment and a medium for the cycle life of a lithium battery, wherein the prediction method comprises the following steps: acquiring real-time operation data of the lithium battery to be predicted in a plurality of charge-discharge cycle periods, and performing data cleaning treatment on the real-time operation data to obtain target operation data; determining a plurality of dimension curves corresponding to a plurality of charge-discharge cycle periods according to the target operation data, and extracting target characteristic data reflecting capacity degradation of the lithium battery to be predicted from the plurality of dimension curves; and inputting the target characteristic data into a pre-trained lithium battery cycle life prediction model to determine a cycle life prediction value of the lithium battery to be predicted. According to the method and the device, the limitation of a single model in the aspects of generalization performance and the like is overcome, and the cycle life of the lithium battery is accurately predicted by using early cycle data.
Description
Technical Field
The present invention relates to the field of battery life prediction, and in particular, to a method, a device, equipment and a medium for predicting a cycle life of a lithium battery.
Background
The state of health of lithium batteries is affected by various factors such as temperature, current multiplying power, cut-off voltage, etc., and aging is a long-term gradual change process. Therefore, feedback of the cycle life of a lithium battery is typically a time-lapse feedback, possibly months, years or even decades. If the early cycle data can be used for accurately predicting the service life of the lithium battery, the development period and the process optimization period of the lithium battery can be accelerated, and the service life expectation of the battery can be estimated for customers. Therefore, how to accurately predict the cycle life of a lithium battery based on less cycle data is important.
At present, the prediction of the cycle life of the lithium battery pays more attention to the residual service life, and the research for predicting the cycle life of the lithium battery by utilizing the early cycle data characteristics is less, so that the accelerated research and development of lithium battery enterprises and the early warning of the service cycle of a client are not facilitated.
Disclosure of Invention
In view of the foregoing, an object of the present application is to provide a method, apparatus, device and medium for predicting the cycle life of a lithium battery, which overcome the limitation of a single model in terms of generalization performance and the like, and accurately predict the cycle life of the lithium battery by using early cycle data.
In a first aspect, an embodiment of the present application provides a method for predicting a cycle life of a lithium battery, where the method includes:
acquiring real-time operation data of the lithium battery to be predicted in a plurality of charge-discharge cycle periods, and performing data cleaning treatment on the real-time operation data to obtain target operation data; the data cleaning processing comprises outlier screening and missing value processing;
determining a plurality of dimension curves corresponding to a plurality of charge-discharge cycle periods according to the target operation data, and extracting target characteristic data reflecting capacity degradation of the lithium battery to be predicted from the plurality of dimension curves;
inputting the target characteristic data into a pre-trained lithium battery cycle life prediction model to determine a cycle life prediction value of the lithium battery to be predicted; the lithium battery cycle life prediction model is a Stacking integrated learning model combined by a target base learner and a target element learner.
Further, the extracting target feature data for reflecting the capacity degradation of the lithium battery from the dimension curve includes:
extracting a plurality of original characteristic data reflecting the capacity degradation of the lithium battery to be predicted from a plurality of dimension curves;
Determining a correlation coefficient and an XgBoost characteristic importance index of each original characteristic data;
screening a plurality of original characteristic data based on the correlation coefficient of each original characteristic data to determine a plurality of characteristic data to be screened;
and screening the plurality of feature data to be screened based on the XgBoost feature importance index of each feature data to be screened, and determining the target feature data.
Further, the dimension curves include a discharge capacity curve, a differential capacity curve, and a temperature curve.
Further, when the dimension curve is the discharge capacity curve, the raw characteristic data is determined by:
determining a plurality of first target discharge capacity curves from the plurality of discharge capacity curves, and determining a target curve linearly fitted with the plurality of target discharge capacity curves based on the plurality of first target discharge capacity curves;
for each first target discharge capacity curve, determining the first discharge capacity of the lithium battery to be predicted in a charge-discharge cycle period corresponding to the first target discharge capacity curve based on the first target discharge capacity curve;
determining a maximum discharge capacity from a plurality of first discharge capacities;
Determining two second target discharge capacity curves from the plurality of discharge capacity curves;
for each second target discharge capacity curve, determining the second discharge capacity of the lithium battery to be predicted in the charge-discharge cycle period corresponding to the second target discharge capacity curve based on the second target discharge capacity curve;
determining the slope and intercept of the target curve, the second discharge capacity, the difference between the maximum discharge capacity and the second discharge capacity as the original characteristic data;
when the dimension curve is the differential capacity curve, the raw feature data is determined by:
determining a first differential capacity curve and a second differential capacity curve from a plurality of differential capacity curves, and performing difference on the first differential capacity curve and the second differential capacity curve to obtain a difference value curve;
determining the maximum value, the minimum value, the variance, the bias state and the kurtosis in the difference curve as the original characteristic data;
when the dimension curve is the temperature curve, the raw characteristic data is determined by:
determining a plurality of target temperature curves from a plurality of temperature curves;
Determining the highest temperature value, the lowest temperature value and the average temperature value of the lithium battery to be predicted in the charge-discharge cycle period corresponding to the target temperature curves from the target temperature curves;
and determining the difference between the highest temperature value and the lowest temperature value, the highest temperature value, the lowest temperature value and the average temperature value as the original characteristic data.
Further, the lithium battery cycle life prediction model is trained by:
acquiring historical operation data of a sample lithium battery in a plurality of sample periods and actual values of cycle life in the plurality of sample periods, and performing data cleaning treatment on the historical operation data to obtain sample operation data;
determining sample dimension curves corresponding to the sample periods according to the sample operation data, and extracting sample characteristic data reflecting capacity degradation of the sample lithium battery from the sample dimension curves;
determining a training sample set and a testing sample set according to the sample characteristic data and the actual value of the cycle life;
combining different base learners and meta learners on the plurality of base models to obtain a plurality of initial Stacking integrated learning models; the Stacking integrated learning model consists of the target base learner and an initial element learner;
For each initial Stacking integrated learning model, performing ten-fold cross validation on the initial Stacking integrated learning model by using the training sample set, and determining a decision coefficient of the initial Stacking integrated learning model;
determining a Stacking integrated learning target model with the largest decision coefficient from a plurality of initial Stacking integrated learning models, and determining the Stacking integrated learning target model as the lithium battery cycle life prediction model;
and performing super-parameter learning on the initial element learner in the lithium battery cycle life prediction model according to the test sample set until the score of the initial element learner in the lithium battery cycle life prediction model reaches a score threshold value, and determining the initial element learner when the score reaches the score threshold value as a target element learner in the lithium battery cycle life prediction model.
In a second aspect, embodiments of the present application further provide a prediction apparatus for cycle life of a lithium battery, where the prediction apparatus includes:
the data acquisition module is used for acquiring real-time operation data of the lithium battery to be predicted in a plurality of charge-discharge cycle periods, and performing data cleaning treatment on the real-time operation data to obtain target operation data; the data cleaning processing comprises outlier screening and missing value processing;
The characteristic data extraction module is used for determining a plurality of dimension curves corresponding to a plurality of charge-discharge cycle periods according to the target operation data, and extracting target characteristic data reflecting the capacity degradation of the lithium battery to be predicted from the plurality of dimension curves;
the prediction module is used for inputting the target characteristic data into a pre-trained lithium battery cycle life prediction model so as to determine a cycle life prediction value of the lithium battery to be predicted; the lithium battery cycle life prediction model is a Stacking integrated learning model combined by a target base learner and a target element learner.
Further, when the feature data extraction module is configured to extract target feature data reflecting capacity degradation of the lithium battery from the dimension curve, the feature data extraction module is further configured to:
extracting a plurality of original characteristic data reflecting the capacity degradation of the lithium battery to be predicted from a plurality of dimension curves;
determining a correlation coefficient and an XgBoost characteristic importance index of each original characteristic data;
screening a plurality of original characteristic data based on the correlation coefficient of each original characteristic data to determine a plurality of characteristic data to be screened;
And screening the plurality of feature data to be screened based on the XgBoost feature importance index of each feature data to be screened, and determining the target feature data.
Further, the dimension curves include a discharge capacity curve, a differential capacity curve, and a temperature curve.
In a third aspect, embodiments of the present application further provide an electronic device, including: the system comprises a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, the processor and the memory are communicated through the bus when the electronic device is running, and the machine-readable instructions are executed by the processor to execute the steps of the lithium battery cycle life prediction method.
In a fourth aspect, embodiments of the present application also provide a computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of a method for predicting cycle life of a lithium battery as described above.
According to the prediction method, the prediction device, the equipment and the medium for the cycle life of the lithium battery, firstly, real-time operation data of the lithium battery to be predicted in a plurality of charge and discharge cycle periods are obtained, and data cleaning processing is carried out on the real-time operation data to obtain target operation data; the data cleaning processing comprises outlier screening and missing value processing; then, determining a plurality of dimension curves corresponding to a plurality of charge-discharge cycle periods according to the target operation data, and extracting target characteristic data reflecting capacity degradation of the lithium battery to be predicted from the plurality of dimension curves; finally, inputting the target characteristic data into a pre-trained lithium battery cycle life prediction model to determine a cycle life prediction value of the lithium battery to be predicted; the lithium battery cycle life prediction model is a Stacking integrated learning model combined by a target base learner and a target element learner.
Therefore, compared with the lithium battery cycle life prediction method in the prior art, the method not only can improve the generalization capability of the model and avoid overfitting, but also can accurately predict the lithium battery cycle life, and improves the accuracy and efficiency of lithium battery cycle life prediction. The limitation of a single model in the aspects of generalization performance and the like is overcome, and the cycle life of the lithium battery is accurately predicted by using early cycle data.
In order to make the above objects, features and advantages of the present application more comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the embodiments will be briefly described below, it being understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered limiting the scope, and that other related drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a method for predicting cycle life of a lithium battery according to an embodiment of the present application;
fig. 2 is a schematic structural diagram of a device for predicting cycle life of a lithium battery according to an embodiment of the present application;
FIG. 3 is a second schematic structural diagram of a device for predicting cycle life of a lithium battery according to an embodiment of the present disclosure;
fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Detailed Description
For the purposes of making the objects, technical solutions and advantages of the embodiments of the present application more clear, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is apparent that the described embodiments are only some embodiments of the present application, but not all embodiments. The components of the embodiments of the present application, which are generally described and illustrated in the figures herein, may be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the present application, as provided in the accompanying drawings, is not intended to limit the scope of the application, as claimed, but is merely representative of selected embodiments of the application. Based on the embodiments of the present application, every other embodiment that a person skilled in the art would obtain without making any inventive effort is within the scope of protection of the present application.
First, application scenarios applicable to the present application will be described. The method and the device can be applied to the field of battery life prediction.
The state of health of lithium batteries is affected by various factors such as temperature, current multiplying power, cut-off voltage, etc., and aging is a long-term gradual change process. Therefore, feedback of the cycle life of a lithium battery is typically a time-lapse feedback, possibly months, years or even decades. If the early cycle data can be used for accurately predicting the service life of the lithium battery, the development period and the process optimization period of the lithium battery can be accelerated, and the service life expectation of the battery can be estimated for customers. Therefore, how to accurately predict the cycle life of a lithium battery based on less cycle data is important.
At present, the prediction of the cycle life of the lithium battery pays more attention to the residual service life, and the research for predicting the cycle life of the lithium battery by utilizing the early cycle data characteristics is less, so that the accelerated research and development of lithium battery enterprises and the early warning of the service cycle of a client are not facilitated.
According to research, the existing two methods for predicting the cycle life of the lithium battery by utilizing the early cycle data features are mainly based on a data driving method, firstly, characteristic indexes for reflecting the capacity degradation of the lithium battery are extracted from different charge and discharge cycle data of the lithium battery, then, the life prediction model (linear regression and elastic network) is trained by utilizing the data, and finally, the cycle life of the lithium battery is output. However, due to nonlinear degradation and wide variability in the cyclic process, the single model-based prediction method inevitably has some limitations, which are mainly represented by low generalization performance and prediction accuracy of the model.
Based on the above, the embodiment of the application provides a method, a device, equipment and a medium for predicting the cycle life of a lithium battery, which overcome the limitation of a single model in the aspects of generalization performance and the like and accurately predict the cycle life of the lithium battery by using early cycle data.
Referring to fig. 1, fig. 1 is a flowchart of a method for predicting cycle life of a lithium battery according to an embodiment of the present application. As shown in fig. 1, a method for predicting cycle life of a lithium battery according to an embodiment of the present application includes:
s101, acquiring real-time operation data of the lithium battery to be predicted in a plurality of charge-discharge cycle periods, and performing data cleaning processing on the real-time operation data to obtain target operation data.
It should be noted that, the lithium battery to be predicted refers to a lithium battery for performing cycle life prediction of the lithium battery. The lithium battery to be predicted is a charge-discharge cycle period after completing one charge-discharge process. The real-time operation data refers to the operation data of the lithium battery to be predicted in each charge-discharge cycle. The real-time operation data may include charge and discharge voltage, current, charge and discharge capacity, temperature, etc. of the lithium battery to be predicted, which is not particularly limited. The data cleaning process includes outlier screening and missing value processing. Here, the abnormal value refers to a case where an abnormal zero drop occurs in a data curve (for example, a curve in which the temperature changes with time in one cycle) in one charge and discharge cycle period (for example, a case where the temperature changes up and down at 25 ℃ in one cycle, and 0 ℃ suddenly occurs, and it can be regarded as an abnormal zero drop).
For the step S101, in implementation, real-time operation data of the lithium battery to be predicted in a plurality of charge-discharge cycle periods is obtained, and abnormal value screening and missing value processing are performed on the real-time operation data to obtain target operation data. Specifically, when abnormal value screening and missing value processing are performed, the specific steps are as follows: firstly screening out abnormal values, and then deleting the abnormal values to be regarded as missing values; then, for all missing values, backward selecting adjacent values for interpolation to replace the missing values.
S102, determining a plurality of dimension curves corresponding to the charge-discharge cycle periods according to the target operation data, and extracting target characteristic data reflecting the capacity degradation of the lithium battery to be predicted from the plurality of dimension curves.
The dimension curve refers to a curve in which real-time operation data changes with time in the charge-discharge cycle period. Each charge-discharge cycle corresponds to a dimension curve.
For the above step S102, in implementation, a plurality of dimension curves corresponding to a plurality of charge-discharge cycle periods are determined according to the target operation data. Then, target characteristic data reflecting the capacity degradation of the lithium battery to be predicted are extracted from the plurality of dimension curves.
Specifically, for the step S102, the extracting, from the dimension curve, the target feature data for reflecting the degradation of the capacity of the lithium battery includes:
and step 1021, extracting a plurality of original characteristic data reflecting the capacity degradation of the lithium battery to be predicted from a plurality of dimension curves.
It should be noted that, the original feature data refers to feature data determined directly according to the dimension curve.
For the above step 1021, in implementation, a plurality of raw feature data reflecting the degradation of the capacity of the lithium battery to be predicted are extracted from a plurality of dimension curves.
Here, according to the prediction method provided in the present application, the dimensional curve includes a discharge capacity curve, a differential capacity curve (dQ/dV curve), and a temperature curve.
Specifically, when the dimension curve is the discharge capacity curve, the raw characteristic data is determined by:
a: and determining a plurality of first target discharge capacity curves from the plurality of discharge capacity curves, and determining a target curve linearly fitted with the plurality of target discharge capacity curves based on the plurality of first target discharge capacity curves.
For the step a, in the implementation, a plurality of first target discharge capacity curves are determined from the plurality of discharge capacity curves, and a target curve linearly fitted to the plurality of target discharge capacity curves is determined based on the plurality of first target discharge capacity curves. For example, the dimension curve includes a discharge capacity curve corresponding to the first 100 charge-discharge cycle periods of the lithium battery to be predicted, and the discharge capacity curve corresponding to the 3 rd charge-discharge cycle period-the 100 th charge-discharge cycle period is determined as the first target discharge capacity curve. The method of how to perform the linear fitting is described in detail in the prior art and is not described in detail here.
B: and determining the first discharge capacity of the lithium battery to be predicted in the charge-discharge cycle period corresponding to the first target discharge capacity curve based on the first target discharge capacity curve aiming at each first target discharge capacity curve.
Since the discharge capacity curve represents the decrease of the electric capacity in the lithium battery to be predicted with the lapse of time, the discharge capacity of the lithium battery to be predicted in the charge-discharge cycle period corresponding to the discharge capacity curve can be determined according to the discharge capacity curve. For the step B, in implementation, for each first target discharge capacity curve, a first discharge capacity of the lithium battery to be predicted in a charge-discharge cycle period corresponding to the first target discharge capacity curve is determined based on the first target discharge capacity curve.
C: and determining the maximum discharge capacity from the plurality of first discharge capacities.
For the above step C, in the specific implementation, after determining the plurality of first discharge capacities in the step B, a maximum discharge capacity is determined from the plurality of first discharge capacities.
D: and determining two second target discharge capacity curves from the plurality of discharge capacity curves.
E: and determining the second discharge capacity of the lithium battery to be predicted in the charge-discharge cycle period corresponding to the second target discharge capacity curve based on the second target discharge capacity curve aiming at each second target discharge capacity curve.
For the above step E, in implementation, two second target discharge capacity curves are determined from the plurality of discharge capacity curves. For example, the discharge capacity curve corresponding to the third charge-discharge cycle period and the discharge capacity curve corresponding to the 100 th charge-discharge cycle period are determined as the second target discharge capacity curve, and the present application is not particularly limited. And determining the second discharge capacity of the lithium battery to be predicted in the charge-discharge cycle period corresponding to the second target discharge capacity curve based on the second target discharge capacity curve aiming at each second target discharge capacity curve.
F: and determining the slope and intercept of the target curve, the second discharge capacity, and the difference value between the maximum discharge capacity and the second discharge capacity as the original characteristic data.
For the step F, after determining the target curve, the maximum discharge capacity and the second discharge capacity, determining the slope and intercept of the target curve, the second discharge capacity, and the difference between the maximum discharge capacity and the second discharge capacity as the original characteristic data.
When the dimension curve is the differential capacity curve, the raw feature data is determined by:
a: determining a first differential capacity curve and a second differential capacity curve from the differential capacity curves, and performing difference on the first differential capacity curve and the second differential capacity curve to obtain a difference curve.
b: and determining the maximum value, the minimum value, the variance, the skewness and the kurtosis in the difference curve as the original characteristic data.
For the above steps a-b, in implementation, the first differential capacity curve and the second differential capacity curve are determined from a plurality of differential capacity curves, for example, the differential capacity curve of the 100 th charge-discharge cycle is determined as the first differential capacity curve, and the differential capacity curve of the 3 rd charge-discharge cycle is determined as the second differential capacity curve. And the first differential capacity curve and the second differential capacity curve are subjected to difference to obtain a difference curve. The maximum, minimum, variance, bias and kurtosis in the difference curve are then determined as raw feature data.
When the dimension curve is the temperature curve, the raw characteristic data is determined by:
i: and determining a plurality of target temperature curves from a plurality of temperature curves.
And II, determining the highest temperature value, the lowest temperature value and the average temperature value of the lithium battery to be predicted in the charge-discharge cycle period corresponding to the target temperature curves from the target temperature curves.
III: and determining the difference between the highest temperature value and the lowest temperature value, the highest temperature value, the lowest temperature value and the average temperature value as the original characteristic data.
For the steps I to III, in implementation, a plurality of target temperature curves are determined from a plurality of temperature curves, for example, a dimension curve includes a temperature curve corresponding to the previous 100 charge-discharge cycle periods of the lithium battery to be predicted, and a temperature curve corresponding to the 3 rd charge-discharge cycle period to the 100 th charge-discharge cycle period is determined as the target temperature curve. Since the target temperature curve represents the condition that the temperature of the lithium battery to be predicted changes along with the time, the highest temperature value, the lowest temperature value and the average temperature value of the lithium battery to be predicted in the charge-discharge cycle period corresponding to the target temperature curves are determined from the target temperature curves. The difference between the highest temperature value and the lowest temperature value, the highest temperature value, the lowest temperature value, and the average temperature are then determined as raw characteristic data.
Step 1022 determines the correlation coefficient and XgBoost feature importance index for each raw feature data.
It should be noted that the correlation coefficient refers to the degree of correlation between the original characteristic data and the cycle life of the lithium battery to be predicted. The XgBoost feature importance is that which feature data is selected as a division point is calculated according to the gain condition of the structure score by constructing an XGBoost model, and the importance of certain feature data is the sum of the occurrence times of the feature data in all trees. In the XgBoost feature importance index calculation, the following three criteria can be selected: weight-the number of times that feature is used as a feature for a segmented sample in all trees; gain-average gain in all trees; cover-the average coverage when this feature is used in a tree. In the examples provided herein, weight was chosen as the XgBoost feature importance index.
For the above step 1022, in implementation, after the raw feature data is determined in step 1021, for each raw feature data, a correlation coefficient of the raw feature data feature and the battery cycle life is calculated, where the correlation coefficient varies from 0 to 1. How to calculate the correlation coefficient is described in detail in the prior art, and is not described here. And then carrying out data feature importance evaluation on the calculated original feature data, and calculating an XgBoost feature importance index of the original feature data.
Step 1023, screening the plurality of original characteristic data based on the correlation coefficient of each original characteristic data, and determining a plurality of characteristic data to be screened.
For the step 1023, in implementation, after the correlation coefficient of each original feature data is calculated, screening is performed on a plurality of original feature data based on the correlation coefficient of each original feature data, so as to determine a plurality of feature data to be screened. Specifically, the correlation coefficient of each original feature data may be used to sort the plurality of original feature data, and then the original feature data with the number of bits preset before the sorting is determined as the feature data to be screened.
And step 1024, screening a plurality of feature data to be screened based on the XgBoost feature importance index of each feature data to be screened, and determining the target feature data.
For the step 1023, in the specific implementation, after the feature data to be screened is determined, screening is performed on the plurality of feature data to be screened based on the XgBoost feature importance index of each feature data to be screened, so as to determine the target feature data. Specifically, the XgBoost feature importance indexes of the feature data to be screened can be utilized to sort the feature data to be screened, and then the feature data to be screened with the preset number of bits before sorting is determined as target feature data.
And S103, inputting the target characteristic data into a pre-trained lithium battery cycle life prediction model to determine a cycle life prediction value of the lithium battery to be predicted.
The lithium battery cycle life prediction model is a model for predicting the life of a lithium battery cell. Here, the lithium battery cycle life prediction model is a Stacking integrated learning model combined by a target base learner and a target element learner. The ensemble learning is a machine learning method based on multi-algorithm fusion of statistical learning theory, and in general, for a single prediction model, the prediction accuracy rate of the model is a trend of decreasing marginal utility, and the Stacking ensemble learning mode is a model integration technology for combining information from multiple prediction models to generate a new model. Different machine learning algorithms are combined together in different ways to achieve superior performance over a single algorithm. In the Stacking integrated learning model, the independent prediction capability of each base learner is analyzed, and meanwhile, the combined effect of each base learner is comprehensively compared, so that the Stacking integrated learning model obtains the optimal prediction effect. The cycle life prediction value refers to the number of times the battery can be fully charged and discharged before the nominal capacity of the battery reaches 80% of the initial rated capacity of the battery.
For the step S103, in the specific implementation, the target feature data determined in the step S102 is input into a pre-trained lithium battery cycle life prediction model, so as to determine a cycle life prediction value of the lithium battery to be predicted. Specifically, the lithium battery cycle life prediction model predicts the service life of the battery core of the lithium battery to be predicted through the first layer base learner, inputs a prediction result into the second layer element learner, and then discovers and corrects the deviation in the base learner through the second layer element learner to predict a final cycle life prediction value.
Specifically, according to the prediction method provided by the application, the lithium battery cycle life prediction model is trained through the following steps:
(1) And acquiring historical operation data of the sample lithium battery in a plurality of sample periods and actual values of cycle life in the plurality of sample periods, and performing data cleaning treatment on the historical operation data to obtain sample operation data.
(2) And determining a plurality of sample dimension curves corresponding to the sample periods according to the sample operation data, and extracting sample characteristic data reflecting the capacity degradation of the sample lithium battery from the sample dimension curves.
The sample lithium battery was a lithium battery used for model training. The sample lithium battery is one sample period after completing one charge and discharge process. Historical operating data refers to the operating data of the sample lithium battery during each sample period. The sample dimension curve refers to a curve in which historical operating data varies over time in a sample period. Each sample period corresponds to one sample dimension curve.
And (3) aiming at the steps (1) to (2), acquiring historical operation data and actual values of cycle life of the sample lithium battery in a plurality of sample periods when the method is specifically implemented, and performing data cleaning treatment on the historical operation data to obtain sample operation data. And then, determining a sample dimension curve corresponding to a plurality of sample periods according to the sample operation data, and extracting sample characteristic data for reflecting the capacity degradation of the sample lithium battery from the sample dimension curve. The method for performing data cleaning processing on the historical operation data is the same as the method provided in step S101, and the method for determining the sample dimension curve and the sample feature data is the same as the method provided in step S102, and the same technical effects can be achieved, which will not be described in detail.
(3) And determining a training sample set and a testing sample set according to the sample characteristic data and the actual value of the cycle life.
For the step (3), after the sample characteristic data is determined in the specific implementation, a training sample set and a test sample set are determined according to the sample characteristic data. As an alternative embodiment, the sample characteristic data may be scaled, for example, in a ratio of 8:2, i.e., 80% of the sample characteristic data and the same actual value of the cycle life as the sample period to which 80% of the sample characteristic data belongs are used as the training sample set, and 20% of the sample characteristic data and the same actual value of the cycle life as the sample period to which 20% of the sample characteristic data belongs are used as the test sample set.
(4) And combining different base learners and meta learners on the plurality of base models to obtain a plurality of initial Stacking integrated learning models.
The Stacking integrated learning model consists of the target base learner and an initial element learner.
And (3) aiming at the step (4), when the method is specifically implemented, combining different base learners and meta learners on a plurality of base models to obtain a plurality of Stacking integrated learning models. According to the embodiment provided by the application, in order to ensure the precision of Stacking integrated learning, the application selects a base model from the difference and independence of a single model, and selects 9 commonly used models in lithium battery life prediction, namely a Lasso regression model (Lasso), an elastic network model (EN), a support vector machine regression model (SVM), a gradient lifting regression model (GBR), a limiting gradient lifting regression model (XgBoost), an extreme random forest regression model (ET), a random forest regression model (RF), a general linear regression model (LR) and an adaptive enhancement regression model (AdaBoost), respectively. The 9 base models are respectively combined by different base learners and meta learners, and the Stacking integrated learning model has 466 combination modes. Wherein the target base learner is to arbitrarily select three base models from 9 base models and combine the three base models with each other. For example: 1. three base models combine: [ Lasso, EN, SVM ] or [ Lasso, EN, GBR ] or [ Lasso, EN, xgBoost ]; 2. four base models combine: [ Lasso, EN, xgBoost, SVM ]; 3. five base models are combined: [ Lasso, EN, SVM, GBR, xgBoost ]; 4. six base models combined: [ Lasso, EN, xgBoost, SVM, RF, LR ]; 5. seven base models combine: [ Lasso, EN, SVM, GBR, xgBoost, ET, RF ]; 6. eight base model combinations: [ Lasso, EN, SVM, GBR, xgBoost, ET, RF, LR ]; 7. nine base model combinations: [ Lasso, EN, SVM, GBR, xgBoost, ET, RF, LR, adaBoost ]. The initial element learner is to select one model from 9 base models, such as [ Lasso ] or [ EN ] or [ SVM ] or [ GBR ] or [ XgBoost ]. And then combining the target base learner with the initial meta learner to obtain the initial Stacking integrated learning model.
(5) And for each initial Stacking integrated learning model, performing ten-fold cross validation on the initial Stacking integrated learning model by using the training sample set, and determining the decision coefficient of the initial Stacking integrated learning model.
For the step (5), in the specific implementation, for each initial Stacking integrated learning model, ten-fold cross validation is performed on the initial Stacking integrated learning model by using a training sample set, and the decision coefficient of the initial Stacking integrated learning model is determined. Ten-fold cross-validation, namely dividing a training sample set into ten parts, taking 9 parts of the training sample set as training data and 1 part of the training sample set as validation data in turn, and performing validation. Evaluation index R of 10 total verification results 2 The mean value of (2) is used as the decision coefficient of the initial Stacking integrated learning model. Specifically, the evaluation index R 2 The calculation is performed by the following formula:
wherein y is i The actual value of the cycle life is indicated,
cycle life prediction value representing initial Stacking ensemble learning model prediction, +.>
N is the sample size, which is the cycle life average.
(6) And determining a Stacking integrated learning target model with the largest decision coefficient from a plurality of initial Stacking integrated learning models, and determining the Stacking integrated learning target model as the lithium battery cycle life prediction model.
Aiming at the step (6), in the specific implementation, after determining the decision coefficient of each initial Stacking integrated learning model, determining a Stacking integrated learning target model with the largest decision coefficient from a plurality of initial Stacking integrated learning models, and determining the Stacking integrated learning target model as a lithium battery cycle life prediction model. According to the embodiment provided by the application, in the embodiment, the optimal lithium battery cycle life prediction model is that the first layer base learner comprises a Lasso regression model, a support vector machine regression model, a gradient lifting regression model and a limit gradient lifting regression model, and the second layer element learner is self-adaptive enhancement regression.
(7) And performing super-parameter learning on the initial element learner in the lithium battery cycle life prediction model according to the test sample set until the score of the initial element learner in the lithium battery cycle life prediction model reaches a score threshold value, and determining the initial element learner when the score reaches the score threshold value as a target element learner in the lithium battery cycle life prediction model.
It should be noted that the super-parameters are parameters whose values are set before the learning process is started, and are not parameter data obtained by training. In general, the super-parameters need to be optimized in the machine learning process, and a set of optimal super-parameters is selected for the learner to improve the learning performance and effect. For example, learning rate(s), regularization weights, adjusting sample equalization, etc. are super-parameters. Super-parameter learning refers to parameter optimization in the model training process, and is generally to effectively search possible values of parameters and then select optimal parameters by using an evaluation function. The evaluation index may be selected from base_counter, n_ estimators, learning _rate, loss, and the like as needed, and the present application is not particularly limited. Grid search refers to trying each possibility by looping through all candidate parameter choices, with the best performing parameter being the final result. The score threshold value refers to a value preset in advance and used for judging whether the initial meta-learner reaches the training standard or not.
And (3) aiming at the step (7), in the specific implementation, according to the test sample set, performing super-parameter learning on the initial element learner in the lithium battery cycle life prediction model until the score of the initial element learner in the lithium battery cycle life prediction model reaches a score threshold value, and determining the initial element learner when the score reaches the score threshold value as the target element learner in the lithium battery cycle life prediction model if the initial element learner is considered to reach the training standard. According to the embodiment provided in the application, the super parameters of the second layer element learner AdaBoost mainly include base_timer, n_ estimators, learning _rate and loss, and the optimal super parameters corresponding to the base_timer, n_ estimators, learning _rate and loss are 5, 40, 1 and 'linear', respectively.
The embodiment of the application provides a prediction method for the cycle life of a lithium battery, which comprises the following steps: firstly, acquiring real-time operation data of a lithium battery to be predicted in a plurality of charge-discharge cycle periods, and performing data cleaning treatment on the real-time operation data to obtain target operation data; the data cleaning processing comprises outlier screening and missing value processing; then, determining a plurality of dimension curves corresponding to a plurality of charge-discharge cycle periods according to the target operation data, and extracting target characteristic data reflecting capacity degradation of the lithium battery to be predicted from the plurality of dimension curves; finally, inputting the target characteristic data into a pre-trained lithium battery cycle life prediction model to determine a cycle life prediction value of the lithium battery to be predicted; the lithium battery cycle life prediction model is a Stacking integrated learning model combined by a target base learner and a target element learner.
Therefore, compared with the lithium battery cycle life prediction method in the prior art, the method not only can improve the generalization capability of the model and avoid overfitting, but also can accurately predict the lithium battery cycle life, and improves the accuracy and efficiency of lithium battery cycle life prediction. The limitation of a single model in the aspects of generalization performance and the like is overcome, and the cycle life of the lithium battery is accurately predicted by using early cycle data.
Referring to fig. 2 and 3, fig. 2 is a schematic structural diagram of a device for predicting cycle life of a lithium battery according to an embodiment of the present application, and fig. 3 is a schematic structural diagram of a second device for predicting cycle life of a lithium battery according to an embodiment of the present application. As shown in fig. 3, the prediction apparatus 200 includes:
the data acquisition module 201 is configured to acquire real-time operation data of the lithium battery to be predicted in a plurality of charge-discharge cycle periods, and perform data cleaning processing on the real-time operation data to obtain target operation data; the data cleaning processing comprises outlier screening and missing value processing;
the feature data extraction module 202 is configured to determine a plurality of dimension curves corresponding to the charge-discharge cycle periods according to the target operation data, and extract target feature data reflecting the degradation of the capacity of the lithium battery to be predicted from the plurality of dimension curves;
The prediction module 203 is configured to input the target feature data into a pre-trained lithium battery cycle life prediction model, so as to determine a cycle life prediction value of the lithium battery to be predicted; the lithium battery cycle life prediction model is a Stacking integrated learning model combined by a target base learner and a target element learner.
Further, when the feature data extracting module 202 is configured to extract target feature data reflecting degradation of the capacity of the lithium battery from the dimension curve, the feature data extracting module 202 is further configured to:
extracting a plurality of original characteristic data reflecting the capacity degradation of the lithium battery to be predicted from a plurality of dimension curves;
determining a correlation coefficient and an XgBoost characteristic importance index of each original characteristic data;
screening a plurality of original characteristic data based on the correlation coefficient of each original characteristic data to determine a plurality of characteristic data to be screened;
and screening the plurality of feature data to be screened based on the XgBoost feature importance index of each feature data to be screened, and determining the target feature data.
Further, the dimension curves include a discharge capacity curve, a differential capacity curve, and a temperature curve.
Further, when the dimension curve is the discharge capacity curve, the feature data extraction module 202 is further configured to determine the raw feature data by:
determining a plurality of first target discharge capacity curves from the plurality of discharge capacity curves, and determining a target curve linearly fitted with the plurality of target discharge capacity curves based on the plurality of first target discharge capacity curves;
for each first target discharge capacity curve, determining the first discharge capacity of the lithium battery to be predicted in a charge-discharge cycle period corresponding to the first target discharge capacity curve based on the first target discharge capacity curve;
determining a maximum discharge capacity from a plurality of first discharge capacities;
determining two second target discharge capacity curves from the plurality of discharge capacity curves;
for each second target discharge capacity curve, determining the second discharge capacity of the lithium battery to be predicted in the charge-discharge cycle period corresponding to the second target discharge capacity curve based on the second target discharge capacity curve;
determining the slope and intercept of the target curve, the second discharge capacity, the difference between the maximum discharge capacity and the second discharge capacity as the original characteristic data;
When the dimension curve is the differential capacity curve, the feature data extraction module 202 is further configured to determine the raw feature data by:
determining a first differential capacity curve and a second differential capacity curve from a plurality of differential capacity curves, and performing difference on the first differential capacity curve and the second differential capacity curve to obtain a difference value curve;
determining the maximum value, the minimum value, the variance, the bias state and the kurtosis in the difference curve as the original characteristic data;
when the dimension curve is the temperature curve, the feature data extraction module 202 is further configured to determine the raw feature data by:
determining a plurality of target temperature curves from a plurality of temperature curves;
determining the highest temperature value, the lowest temperature value and the average temperature value of the lithium battery to be predicted in the charge-discharge cycle period corresponding to the target temperature curves from the target temperature curves;
and determining the difference between the highest temperature value and the lowest temperature value, the highest temperature value, the lowest temperature value and the average temperature value as the original characteristic data.
Further, referring to fig. 3, the prediction apparatus 200 further includes a model training module 204, where the model training module 204 is configured to train the lithium battery cycle life prediction model by:
Acquiring historical operation data of a sample lithium battery in a plurality of sample periods and actual values of cycle life in the plurality of sample periods, and performing data cleaning treatment on the historical operation data to obtain sample operation data;
determining sample dimension curves corresponding to the sample periods according to the sample operation data, and extracting sample characteristic data reflecting capacity degradation of the sample lithium battery from the sample dimension curves;
determining a training sample set and a testing sample set according to the sample characteristic data and the actual value of the cycle life;
combining different base learners and meta learners on the plurality of base models to obtain a plurality of initial Stacking integrated learning models; the Stacking integrated learning model consists of the target base learner and an initial element learner;
for each initial Stacking integrated learning model, performing ten-fold cross validation on the initial Stacking integrated learning model by using the training sample set, and determining a decision coefficient of the initial Stacking integrated learning model;
determining a Stacking integrated learning target model with the largest decision coefficient from a plurality of initial Stacking integrated learning models, and determining the Stacking integrated learning target model as the lithium battery cycle life prediction model;
And performing super-parameter learning on the initial element learner in the lithium battery cycle life prediction model according to the test sample set until the score of the initial element learner in the lithium battery cycle life prediction model reaches a score threshold value, and determining the initial element learner when the score reaches the score threshold value as a target element learner in the lithium battery cycle life prediction model.
Referring to fig. 4, fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present application. As shown in fig. 4, the electronic device 400 includes a processor 410, a memory 420, and a bus 430.
The memory 420 stores machine-readable instructions executable by the processor 410, when the electronic device 400 is running, the processor 410 communicates with the memory 420 through the bus 430, and when the machine-readable instructions are executed by the processor 410, the steps of the method for predicting the cycle life of the lithium battery in the method embodiment shown in fig. 1 can be executed, and the specific implementation is referred to the method embodiment and will not be described herein.
The embodiment of the present application further provides a computer readable storage medium, where a computer program is stored on the computer readable storage medium, and the computer program may execute the steps of the method for predicting the cycle life of a lithium battery in the method embodiment shown in fig. 1 when the computer program is executed by a processor, and the specific implementation manner may refer to the method embodiment and will not be described herein.
It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, and are not repeated herein.
In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods may be implemented in other manners. The above-described apparatus embodiments are merely illustrative, for example, the division of the units is merely a logical function division, and there may be other manners of division in actual implementation, and for example, multiple units or components may be combined or integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be through some communication interface, device or unit indirect coupling or communication connection, which may be in electrical, mechanical or other form.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in each embodiment of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit.
The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a non-volatile computer readable storage medium executable by a processor. Based on such understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
It should be noted that: like reference numerals and letters in the following figures denote like items, and thus once an item is defined in one figure, no further definition or explanation of it is required in the following figures, and furthermore, the terms "first," "second," "third," etc. are used merely to distinguish one description from another and are not to be construed as indicating or implying relative importance.
Finally, it should be noted that: the foregoing examples are merely specific embodiments of the present application, and are not intended to limit the scope of the present application, but the present application is not limited thereto, and those skilled in the art will appreciate that while the foregoing examples are described in detail, the present application is not limited thereto. Any person skilled in the art may modify or easily conceive of the technical solution described in the foregoing embodiments, or make equivalent substitutions for some of the technical features within the technical scope of the disclosure of the present application; such modifications, changes or substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present application, and are intended to be included in the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (10)
1. A method for predicting cycle life of a lithium battery, the method comprising:
acquiring real-time operation data of the lithium battery to be predicted in a plurality of charge-discharge cycle periods, and performing data cleaning treatment on the real-time operation data to obtain target operation data; the data cleaning processing comprises outlier screening and missing value processing;
Determining a plurality of dimension curves corresponding to a plurality of charge-discharge cycle periods according to the target operation data, and extracting target characteristic data reflecting capacity degradation of the lithium battery to be predicted from the plurality of dimension curves;
inputting the target characteristic data into a pre-trained lithium battery cycle life prediction model to determine a cycle life prediction value of the lithium battery to be predicted; the lithium battery cycle life prediction model is a Stacking integrated learning model combined by a target base learner and a target element learner.
2. The method of claim 1, wherein extracting target feature data from the dimensional curve that reflects degradation of lithium battery capacity comprises:
extracting a plurality of original characteristic data reflecting the capacity degradation of the lithium battery to be predicted from a plurality of dimension curves;
determining a correlation coefficient and an XgBoost characteristic importance index of each original characteristic data;
screening a plurality of original characteristic data based on the correlation coefficient of each original characteristic data to determine a plurality of characteristic data to be screened;
and screening the plurality of feature data to be screened based on the XgBoost feature importance index of each feature data to be screened, and determining the target feature data.
3. The prediction method according to claim 2, wherein the dimensional curve includes a discharge capacity curve, a differential capacity curve, and a temperature curve.
4. A prediction method according to claim 3, wherein when the dimension curve is the discharge capacity curve, the raw characteristic data is determined by:
determining a plurality of first target discharge capacity curves from the plurality of discharge capacity curves, and determining a target curve linearly fitted with the plurality of target discharge capacity curves based on the plurality of first target discharge capacity curves;
for each first target discharge capacity curve, determining the first discharge capacity of the lithium battery to be predicted in a charge-discharge cycle period corresponding to the first target discharge capacity curve based on the first target discharge capacity curve;
determining a maximum discharge capacity from a plurality of first discharge capacities;
determining two second target discharge capacity curves from the plurality of discharge capacity curves;
for each second target discharge capacity curve, determining the second discharge capacity of the lithium battery to be predicted in the charge-discharge cycle period corresponding to the second target discharge capacity curve based on the second target discharge capacity curve;
Determining the slope and intercept of the target curve, the second discharge capacity, the difference between the maximum discharge capacity and the second discharge capacity as the original characteristic data;
when the dimension curve is the differential capacity curve, the raw feature data is determined by:
determining a first differential capacity curve and a second differential capacity curve from a plurality of differential capacity curves, and performing difference on the first differential capacity curve and the second differential capacity curve to obtain a difference value curve;
determining the maximum value, the minimum value, the variance, the bias state and the kurtosis in the difference curve as the original characteristic data;
when the dimension curve is the temperature curve, the raw characteristic data is determined by:
determining a plurality of target temperature curves from a plurality of temperature curves;
determining the highest temperature value, the lowest temperature value and the average temperature value of the lithium battery to be predicted in the charge-discharge cycle period corresponding to the target temperature curves from the target temperature curves;
and determining the difference between the highest temperature value and the lowest temperature value, the highest temperature value, the lowest temperature value and the average temperature value as the original characteristic data.
5. The prediction method according to claim 1, wherein the lithium battery cycle life prediction model is trained by:
acquiring historical operation data of a sample lithium battery in a plurality of sample periods and actual values of cycle life in the plurality of sample periods, and performing data cleaning treatment on the historical operation data to obtain sample operation data;
determining sample dimension curves corresponding to the sample periods according to the sample operation data, and extracting sample characteristic data reflecting capacity degradation of the sample lithium battery from the sample dimension curves;
determining a training sample set and a testing sample set according to the sample characteristic data and the actual value of the cycle life;
combining different base learners and meta learners on the plurality of base models to obtain a plurality of initial Stacking integrated learning models; the Stacking integrated learning model consists of the target base learner and an initial element learner;
for each initial Stacking integrated learning model, performing ten-fold cross validation on the initial Stacking integrated learning model by using the training sample set, and determining a decision coefficient of the initial Stacking integrated learning model;
Determining a Stacking integrated learning target model with the largest decision coefficient from a plurality of initial Stacking integrated learning models, and determining the Stacking integrated learning target model as the lithium battery cycle life prediction model;
and performing super-parameter learning on the initial element learner in the lithium battery cycle life prediction model according to the test sample set until the score of the initial element learner in the lithium battery cycle life prediction model reaches a score threshold value, and determining the initial element learner when the score reaches the score threshold value as a target element learner in the lithium battery cycle life prediction model.
6. A prediction apparatus for cycle life of a lithium battery, the prediction apparatus comprising:
the data acquisition module is used for acquiring real-time operation data of the lithium battery to be predicted in a plurality of charge-discharge cycle periods, and performing data cleaning treatment on the real-time operation data to obtain target operation data; the data cleaning processing comprises outlier screening and missing value processing;
the characteristic data extraction module is used for determining a plurality of dimension curves corresponding to a plurality of charge-discharge cycle periods according to the target operation data, and extracting target characteristic data reflecting the capacity degradation of the lithium battery to be predicted from the plurality of dimension curves;
The prediction module is used for inputting the target characteristic data into a pre-trained lithium battery cycle life prediction model so as to determine a cycle life prediction value of the lithium battery to be predicted; the lithium battery cycle life prediction model is a Stacking integrated learning model combined by a target base learner and a target element learner.
7. The prediction apparatus of claim 6, wherein the feature data extraction module, when configured to extract target feature data reflecting degradation of lithium battery capacity from the dimensional curve, is further configured to:
extracting a plurality of original characteristic data reflecting the capacity degradation of the lithium battery to be predicted from a plurality of dimension curves;
determining a correlation coefficient and an XgBoost characteristic importance index of each original characteristic data;
screening a plurality of original characteristic data based on the correlation coefficient of each original characteristic data to determine a plurality of characteristic data to be screened;
and screening the plurality of feature data to be screened based on the XgBoost feature importance index of each feature data to be screened, and determining the target feature data.
8. The predictive device of claim 7, wherein the dimensional curve includes a discharge capacity curve, a differential capacity curve, and a temperature curve.
9. An electronic device, comprising: a processor, a memory and a bus, said memory storing machine readable instructions executable by said processor, said processor and said memory communicating via said bus when the electronic device is running, said machine readable instructions when executed by said processor performing the steps of the method of predicting the cycle life of a lithium battery as claimed in any one of claims 1 to 5.
10. A computer-readable storage medium, characterized in that the computer-readable storage medium has stored thereon a computer program which, when executed by a processor, performs the steps of the method for predicting the cycle life of a lithium battery according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310439970.0A CN116401954A (en) | 2023-04-23 | 2023-04-23 | Prediction method, prediction device, equipment and medium for cycle life of lithium battery |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310439970.0A CN116401954A (en) | 2023-04-23 | 2023-04-23 | Prediction method, prediction device, equipment and medium for cycle life of lithium battery |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116401954A true CN116401954A (en) | 2023-07-07 |
Family
ID=87015981
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310439970.0A Pending CN116401954A (en) | 2023-04-23 | 2023-04-23 | Prediction method, prediction device, equipment and medium for cycle life of lithium battery |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116401954A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116953554A (en) * | 2023-07-25 | 2023-10-27 | 国网江苏省电力有限公司镇江供电分公司 | Multi-fragment data-based method and device for estimating SOH of lithium battery of energy storage power station |
| CN117233630A (en) * | 2023-11-16 | 2023-12-15 | 深圳屹艮科技有限公司 | Method and device for predicting service life of lithium ion battery and computer equipment |
-
2023
- 2023-04-23 CN CN202310439970.0A patent/CN116401954A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116953554A (en) * | 2023-07-25 | 2023-10-27 | 国网江苏省电力有限公司镇江供电分公司 | Multi-fragment data-based method and device for estimating SOH of lithium battery of energy storage power station |
| CN117233630A (en) * | 2023-11-16 | 2023-12-15 | 深圳屹艮科技有限公司 | Method and device for predicting service life of lithium ion battery and computer equipment |
| CN117233630B (en) * | 2023-11-16 | 2024-03-15 | 深圳屹艮科技有限公司 | Method and device for predicting service life of lithium ion battery and computer equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116401954A (en) | Prediction method, prediction device, equipment and medium for cycle life of lithium battery | |
| CN109242135B (en) | Model operation method, device and business server | |
| JP2023037027A (en) | Life prediction method and life prediction program for lithium ion battery, and information processing apparatus | |
| KR20210016154A (en) | Battery diagnostic methods using machine learning | |
| CN112327191B (en) | Battery diving probability evaluation method based on geometric feature fusion decision | |
| CN116298906B (en) | Battery capacity prediction model training method, prediction method, device and medium | |
| CN116683588B (en) | Lithium ion battery charge and discharge control method and system | |
| CN103399280A (en) | Method for predicting cycle life of lithium ion battery based on NSDP-AR (AutoRegressive) model | |
| CN109768340B (en) | Method and device for estimating voltage inconsistency in battery discharge process | |
| CN112834927A (en) | Lithium battery residual life prediction method, system, device and medium | |
| CN114660497A (en) | Lithium ion battery service life prediction method aiming at capacity regeneration phenomenon | |
| CN117031310A (en) | Method for predicting residual service life of power battery of electric automobile | |
| CN117674119A (en) | Power grid operation risk assessment method, device, computer equipment and storage medium | |
| CN117110884A (en) | Lithium battery remaining service life prediction method based on multi-core correlation vector machine | |
| CN109884548B (en) | Method for predicting remaining life of lithium battery based on GASVM-AUKF algorithm | |
| CN113884936B (en) | ISSA coupling DELM-based lithium ion battery health state prediction method | |
| CN116413609B (en) | Battery diving identification method and device, electronic equipment and storage medium | |
| CN117723999B (en) | Battery service life prediction method, device, equipment and storage medium | |
| Hu et al. | Multi-objective optimization estimation of state of health for lithium-ion battery based on constant current charging profile | |
| CN118169582A (en) | Lithium ion battery health state and residual life prediction method | |
| CN117556753B (en) | Method, device, equipment and storage medium for analyzing energy consumption of storage chip | |
| Li et al. | Lithium-ion Battery SOH Estimation with Varying Amount of Battery Operation Data | |
| CN117723999A (en) | Battery service life prediction method, device, equipment and storage medium | |
| CN117234251A (en) | Electrical appliance pressure control method and system based on artificial intelligence | |
| CN117371300A (en) | Method and device for testing reliability of power system containing wind power and electronic equipment |
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 |