CN113671394A - Lithium ion battery expected life prediction method and system - Google Patents
Lithium ion battery expected life prediction method and system Download PDFInfo
- Publication number
- CN113671394A CN113671394A CN202110938372.9A CN202110938372A CN113671394A CN 113671394 A CN113671394 A CN 113671394A CN 202110938372 A CN202110938372 A CN 202110938372A CN 113671394 A CN113671394 A CN 113671394A
- Authority
- CN
- China
- Prior art keywords
- data
- battery
- lithium ion
- ion battery
- model
- 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
- HBBGRARXTFLTSG-UHFFFAOYSA-N Lithium ion Chemical compound [Li+] HBBGRARXTFLTSG-UHFFFAOYSA-N 0.000 title claims abstract description 52
- 229910001416 lithium ion Inorganic materials 0.000 title claims abstract description 52
- 238000000034 method Methods 0.000 title claims abstract description 47
- 238000012545 processing Methods 0.000 claims abstract description 34
- 239000013598 vector Substances 0.000 claims abstract description 12
- 230000008569 process Effects 0.000 claims abstract description 9
- 238000002790 cross-validation Methods 0.000 claims abstract description 8
- 238000004364 calculation method Methods 0.000 claims abstract description 7
- 238000013501 data transformation Methods 0.000 claims description 15
- 238000012360 testing method Methods 0.000 claims description 15
- 238000012549 training Methods 0.000 claims description 13
- 238000004140 cleaning Methods 0.000 claims description 11
- 238000001035 drying Methods 0.000 claims description 11
- 230000000694 effects Effects 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 7
- 230000004927 fusion Effects 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 5
- 238000006243 chemical reaction Methods 0.000 claims description 4
- 238000002759 z-score normalization Methods 0.000 claims 1
- 238000007726 management method Methods 0.000 abstract description 7
- 238000012423 maintenance Methods 0.000 abstract description 4
- 238000010276 construction Methods 0.000 abstract description 3
- 238000013499 data model Methods 0.000 abstract description 3
- 238000013528 artificial neural network Methods 0.000 description 5
- 238000011156 evaluation Methods 0.000 description 4
- 230000002068 genetic effect Effects 0.000 description 4
- 238000005259 measurement Methods 0.000 description 4
- WHXSMMKQMYFTQS-UHFFFAOYSA-N Lithium Chemical compound [Li] WHXSMMKQMYFTQS-UHFFFAOYSA-N 0.000 description 3
- 230000015556 catabolic process Effects 0.000 description 3
- 238000006731 degradation reaction Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 229910052744 lithium Inorganic materials 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 239000000126 substance Substances 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000003487 electrochemical reaction Methods 0.000 description 2
- 238000005562 fading Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000007500 overflow downdraw method Methods 0.000 description 2
- 238000010200 validation analysis Methods 0.000 description 2
- 230000032683 aging Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000001311 chemical methods and process Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000005518 electrochemistry Effects 0.000 description 1
- 238000004146 energy storage Methods 0.000 description 1
- 230000001747 exhibiting effect Effects 0.000 description 1
- 238000007499 fusion processing Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
- 238000005406 washing Methods 0.000 description 1
Images
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]
- G01R31/385—Arrangements for measuring battery or accumulator variables
-
- 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]
- G01R31/385—Arrangements for measuring battery or accumulator variables
- G01R31/387—Determining ampere-hour charge capacity or SoC
- G01R31/388—Determining ampere-hour charge capacity or SoC involving voltage measurements
-
- 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
Abstract
The invention discloses a method and a system for predicting the expected life of a lithium ion battery, which are characterized in that the operation data of the lithium ion battery is collected, the data construction characteristics are transformed by constructing a relation curve between the discharge point capacity and the voltage through a machine, the initial characteristics and the label vector are respectively subjected to logarithmic processing, then the standardization processing is carried out, and the data after the standardization processing is selected to accord with the optimal parameters of the battery life prediction model parameters through setting grid search and cross validation, so that the lithium ion battery has excellent prediction performance based on a data model even before the capacity attenuation starts, the prediction result is more accurate, the calculation process is more efficient, the invention provides efficient and reliable decision reference for the operation management of the lithium ion battery, the management efficiency is improved, and the prediction maintenance and guarantee of the lithium ion battery are supported.
Description
Technical Field
The invention belongs to the field of battery life prediction, and particularly relates to a method and a system for predicting the expected life of a lithium ion battery.
Background
With the rapid development of the industry in China, the current demand on various energy sources is still continuously increased, and due to the progress of energy storage equipment, some new energy sources such as wind energy and energy sources of photovoltaic power generation can be stored in an electric energy form. Lithium ion batteries have been greatly accepted by the market due to their numerous advantages, such as high energy density, stable electrochemical properties, low pollution, long cycle life, etc., and are widely used in various industries, exhibiting outstanding advantages. With the increasing application of the related fields to batteries, how to effectively evaluate and predict the expected life of the battery is not only an important basis for supporting the predictive maintenance and guarantee of the lithium ion battery, but also greatly guarantees the safe use of the battery. However, the decay process of the lithium battery is a typical nonlinear electrochemical system mechanism, and is easily affected by factors such as current, temperature, internal resistance and the like, so that aging occurs, and chemical internal characteristic parameters are not easily measured, so that the accurate evaluation of the battery life is difficult. Therefore, the service life of the lithium battery can be accurately and quickly predicted, so that the safety of the related field is improved, and a large amount of funds and time can be saved for the related field. Therefore, the research of the method capable of accurately predicting the residual life of the lithium ion battery has great significance for the practical application of the method.
The existing prediction method mainly comprises data driving and physical model driving, the degradation rule of the battery performance is obtained by mining failure data based on the data driving prediction method, fitting is completed by using a large amount of data, a priori lithium ion battery component degradation model is not needed, and the prediction method has great advantages in the aspect of prediction. The physical and chemical process in the lithium battery is analyzed by a method based on a physical model, so that the physical model reflecting the evolution process of an object is established, model parameters are adjusted through related data, and then the operation life is predicted, so that the internal condition of the battery can be truly reflected. The prior art needs to know the chemical and physical mechanisms of complicated electrochemical reactions in the electrochemistry and has harsh use conditions. The lithium ion battery life prediction needs to establish a prediction method which is real-time and can process a large amount of data, but the current prediction method cannot meet the requirements, and the current battery life prediction method has large prediction deviation, long prediction time and low prediction result precision, and cannot provide efficient and reliable lithium ion battery life prediction parameters for lithium ion battery management.
CN111426952A discloses a method for predicting the life of a lithium ion battery, in which a capacity fading model is established through two iterations, and then the actual working condition information of the battery is substituted into the capacity fading model, so as to obtain the predicted life of the battery. The advantages are that: compared with the method for simply predicting the storage life or the cycle life in the prior art, the method is more suitable for the actual service condition of the battery, and can quickly predict the actual service life of the battery. However, the conditions used by the method are limited, the chemical and physical mechanisms of the electrochemical internal complex electrochemical reaction need to be known, and once the model is established incorrectly, great deviation is brought to the subsequent prediction result.
CN107748936A discloses a BP neural network storage battery life prediction algorithm improved based on a genetic algorithm, which uses the genetic algorithm to optimize the BP neural network; and establishing a BP neural network model by using the existing service life prediction thought of the storage battery. The method combines the BP neural network and the genetic algorithm, seeks the optimal parameters for the BP neural network through the genetic algorithm, and can effectively avoid the algorithm from falling into local extreme points. However, although the neural network has good learning ability for historical data, the network structure is complicated and difficult to determine, the requirements on the sample size and quality of the data are high, and the uncertainty expression of the output is not provided.
Disclosure of Invention
The invention aims to provide a method and a system for predicting the expected life of a lithium ion battery, which are used for overcoming the defects of the prior art.
In order to achieve the purpose, the invention adopts the following technical scheme:
a method for predicting the expected life of a lithium ion battery comprises the following steps:
s1, collecting the measured data of the battery pack and cleaning and drying the collected data;
s2, analyzing the battery life of each battery in the battery pack according to the discharge capacity, charge capacity, voltage, current and temperature factors, analyzing the discharge capacity and voltage curve relation of the battery pack according to cleaned and dried measured data, and then performing data transformation on the discharge capacity and voltage curve relation to obtain model input initial characteristics and a label vector;
s3, respectively carrying out logarithm processing on the initial features and the label vectors, and then carrying out standardization processing;
s4, selecting the data after standardization processing to accord with the optimal parameters of the battery life prediction model parameters by setting grid searching and cross validation;
s5, selecting characteristic variables from the obtained optimal parameters as input data to be respectively brought into multiple models for training, searching for optimal model parameters through multiple loop iterations, and then selecting a model with the best fitting effect to perform result fusion, thereby obtaining a battery expected life prediction model;
and S6, inputting the time variable parameters of the battery to be tested and the data obtained by data transformation into a battery expected life prediction model as input characteristics, and predicting and acquiring the cycle period of the lithium ion battery through the battery expected life prediction model.
Further, the step of cleaning, washing and drying the measured data of the collected battery pack comprises data drying, missing value processing and format content processing.
Further, when data cleaning is carried out, high-frequency noise data are removed by adopting a box line graph and a wavelet threshold value method.
And further, carrying out data transformation on the relation between the discharge capacity and the voltage curve according to the relation between the discharge capacity and the voltage curve of the battery pack to be tested, selecting a time variable parameter and a parameter after the data transformation as input characteristics, inputting the input characteristics into a battery expected life prediction model, taking the cycle period of the lithium ion battery as output data, and predicting the expected life of the lithium ion battery by selecting and fusing a plurality of models.
Further, the time variable parameters include lithium ion battery discharge capacity, voltage, charging time, battery internal resistance and temperature.
Furthermore, the data set is normalized to be non-dimensionalized.
Further, the non-dimensionalization processing adopts a z-score standardization (zero-mean normalization) calculation method, and the conversion formula is as follows:
wherein，xiAs the original data, it is the original data,
is the mean of the raw data, wherein
n is the number of samples and s is the standard deviation of the original data, where
Furthermore, sample data is extracted from the data after the standardization processing and is divided into a training set, a test set and a second test set, and the second test data set is adopted to evaluate the expected life prediction model of the battery.
Further, the evaluation of the model selects root mean square error and percentage of average error to evaluate the performance of the model, the root mean square error is the square root of the ratio of the sum of squares of the deviations of the observed values and the true values to the observation times, and is used for measuring the deviation between the observed values and the true values, and the root mean square error RMSE is defined as:
the average percent error is defined by the formula:
where yi is the observed cycle life,
is the predicted cycle life and n is the total number of samples.
A lithium ion battery expected life prediction system comprises a data preprocessing module and a prediction module,
the data preprocessing module is used for cleaning and drying acquired data, then analyzing the service life of the battery for the discharge capacity, the charge capacity, the voltage, the current and the temperature factor of each battery in the battery pack, analyzing the relation between the discharge capacity and the voltage curve of the battery pack according to the cleaned and dried actually measured data, and then performing data transformation on the relation between the discharge capacity and the voltage curve to obtain the model input initial characteristic and the label vector; and then respectively carrying out logarithm processing on the initial characteristic and the label vector, then carrying out standardization processing, finally selecting the optimal parameters of which the data after the standardization processing accords with the parameters of the battery life prediction model by setting grid search and cross validation, selecting characteristic variables as input data by using the obtained optimal parameters to be respectively brought into various models for training, searching the optimal model parameters through repeated cycle iteration, then selecting the model with the best fitting effect to carry out result fusion, thereby obtaining the battery expected life prediction model, wherein the prediction module is used for storing the battery expected life prediction model, and predicting the life of the battery to be tested according to the time variable parameters of the battery to be tested and the data obtained by data transformation and outputting the prediction result.
Compared with the prior art, the invention has the following beneficial technical effects:
the invention relates to a lithium ion battery expected life prediction method, which comprises the steps of collecting lithium ion battery operation data, converting the data construction characteristics by constructing a relation curve machine between the discharge point capacity and the voltage, carrying out logarithm processing on initial characteristics and label vectors respectively, then carrying out standardization processing, and selecting the data after the standardization processing to accord with the optimal parameters of battery life prediction model parameters by setting grid search and cross validation, so that the lithium ion battery has excellent prediction performance based on a data model even before the capacity attenuation starts, the prediction result is more accurate, the calculation process is more efficient, the invention provides efficient and reliable decision reference for the operation management of the lithium ion battery, improves the management efficiency, and supports the prediction maintenance and guarantee of the lithium ion battery.
Furthermore, a multi-model fusion method is adopted, so that the robustness and the prediction precision of the model are improved.
Furthermore, the logarithm of the cycle life is predicted by adopting the linear combination of the feature subsets, and the domain specific features with different complexities can be obtained by adopting a regularized linear model, and meanwhile, the higher interpretability is kept. The linear model is also computationally inexpensive.
Drawings
Fig. 1 is a schematic flow chart of a method for predicting the expected life of a lithium ion battery in an embodiment of the present invention.
Fig. 2 is a schematic diagram of a model fusion process of a lithium ion battery life prediction method in an embodiment of the present invention.
Fig. 3 is a graph showing the difference between the 100 th and 10 th discharge capacities and the trend of the voltage curve of the lithium ion battery in the embodiment of the present invention.
Fig. 4 is a graph showing the variance of the difference between the 100 th and 10 th discharge capacities of a lithium ion battery and a log-log distribution of the cycle period according to an embodiment of the present invention.
Fig. 5 is a relationship diagram of the predicted value and the true value of the life of the lithium ion battery in the embodiment of the invention.
Detailed Description
The invention is described in further detail below with reference to the accompanying drawings:
as shown in fig. 1 and fig. 2, a method for predicting the expected life of a lithium ion battery includes the following steps:
s1, acquiring the actual measurement data of the battery pack by using a battery acquisition device, uploading the actual measurement data to a battery life prediction database, and cleaning and drying the acquired actual measurement data of the battery pack, wherein the cleaning and drying comprises data drying, missing value processing and format content processing; and when data cleaning is carried out, high-frequency noise data are removed by adopting a box line graph and a wavelet threshold value method.
S2, data preprocessing analysis: analyzing the service life of the battery for the discharge capacity, the charge capacity, the voltage, the current and the temperature factor of each battery in the battery pack, analyzing the relation between the discharge capacity and the voltage curve of the battery pack according to the cleaned and dried actually measured data, and then performing data transformation on the relation between the discharge capacity and the voltage curve to obtain the model input initial characteristic and the label vector;
s3, respectively carrying out logarithm processing on the initial features and the label vectors, and then carrying out standardization processing to eliminate dimensions among different features and improve the running speed;
then extracting sample data from the data after the standardization processing, and dividing the extracted sample data into a training set, a test set and a second test set;
s4, setting a plurality of groups of lithium ion battery life prediction model parameters, wherein the prediction model parameters comprise model punishment items, learning rates and loss functions, and selecting the optimal parameters by setting grid search and cross validation;
s5, predicting the logarithm of the cycle life by adopting the linear combination of the feature subsets consisting of the obtained optimal parameters, selecting feature variables from the obtained optimal parameters as input data to be respectively introduced into various models for training, searching the optimal model parameters through multiple cycle iterations, and then selecting the model with the best fitting effect to perform result fusion, thereby obtaining a prediction model of the expected life of the battery;
and S6, inputting the time variable parameters of the battery to be tested into the battery expected life prediction model as input characteristics, and predicting and acquiring the cycle period of the lithium ion battery through the battery expected life prediction model.
Performing data transformation on the discharge capacity and voltage curve relation according to the discharge capacity and voltage curve relation of the battery pack to be tested, selecting time variable parameters and parameters after data transformation as input characteristics to be input into a battery expected life prediction model, taking the cycle period of the lithium ion battery as output data, and predicting the expected life of the lithium ion battery by selecting and fusing a plurality of models; the time variable parameters comprise lithium ion battery discharge capacity, voltage, charging time, battery internal resistance and temperature.
Preferably, a multi-model fusion method is adopted, and stacking is to input output results of a series of models (also called base models) as new features into other models.
By selecting the regularized linear model, the domain specific features with different complexities can be obtained, meanwhile, higher interpretability is kept, and the calculation cost of the linear model is low. The model can be trained offline, and online prediction requires only a single dot product after data preprocessing.
As shown in fig. 3 to 5, the period evolution of q (v), i.e. the discharge capacity versus voltage curve for a given period, is considered. Since the voltage range is the same for each cycle, the capacitance is treated as a function of voltage rather than as a function of capacitance to maintain uniformity of the comparison cycle. And a discharge voltage curve between periods 10 and 100 was found to vary, noted as Δ dQ100-10(V) ═ dQ100(V) -dQ10(V), where the subscripts indicate the number of periods. This conversion Δ dq (v) is of particular concern because the voltage curve and its derivatives are a rich source of data that can be used to make degradation diagnostics effective.
Since the statistical features of Δ Q100-10(V) have high predictive power, the model takes into account all available characteristic transformations from the data, including variance, minimum, logarithm, skewness, kurtosis, slope, and intercept. Each summary statistic is a scalar that captures the change in the voltage curve between two cycles. These summary statistics were chosen because of their predictive power, not physical significance, and a clear trend between the cycle life and the summary statistics (particularly the variance) of Δ Q100-10 (V).
Because the dimensions of the variables are different, the variable values with smaller dimensions of part of the variables are larger, the effect of the small-value variables with larger dimensions can be weakened, in the preprocessing process, the data set is subjected to non-dimensionalization processing by adopting the standardized processing of the data set, and meanwhile, the comparison and weighting among different variables can be facilitated;
the non-dimensionalization processing adopts a z-score standardization (zero-mean normalization) calculation method, and the conversion formula is as follows:
wherein x isiAs the original data, it is the original data,
is the mean of the raw data, wherein
n is the number of samples and s is the standard deviation of the original data, where
The standardization not only avoids the weighting problem of different dimensional variables, but also can improve the training speed of the model.
Extracting sample data from the data after the standardization processing, and dividing the extracted sample data into three parts, namely a training set, a test set and a second test set; data was taken from only the first 100 cycles. Where the training set and test set each span a range of cycle life. The training set data is used to select model hyper-parameters by grid search. And the training set data is further subdivided into calibration and validation sets for cross-validation, with the test set used as an accuracy validation of the model. The model is then evaluated on a second test data set generated after model development. The importance of the second test data set is emphasized here, since these data were not generated at model development and are therefore a rigorous test of the model performance, which makes more sense for the evaluation of the model.
Evaluation of the model the Root Mean Square Error (RMSE) and the percentage of mean error were selected to evaluate the performance of the model. The root mean square error is the square root of the ratio of the sum of the squares of the deviations of the observations from the true values to the number of observations, and is a measure of the deviation of the observations from the true values. RMSE is defined by the formula:
the average percent error is defined by the formula:
where yi is the observed cycle life,
is the predicted cycle life and n is the total number of samples.
Finally, according to the root mean square error and the average error percentage, ElasticNet Regression, Kernel Ridge Regression Ridge Regression and Xgboost ensemble learning are selected as a fusion model. Among them, the ElasticNet regression was trained using L1 and L2 is preferred as the regularizer. The ElasticNet is useful when there are multiple related features. Lasso will choose one of them randomly, while ElasticNet will choose two, which effectively avoids overfitting. The robustness and the accuracy of the fused model are correspondingly improved.
The invention applies the machine learning algorithm principle to process the actual measurement parameters of the object battery pack acquired in mass, and can gradually evolve into an intelligent model which can comprehensively reflect the full-cycle service life of the battery through the big data modeling and the automatic real-time correction of the model by the machine learning algorithm, thereby improving the prediction precision of the service life of the lithium ion battery.
According to the method, the lithium ion battery operation data is collected, the data construction characteristics are transformed by a machine through constructing the relation curve between the discharge point capacity and the voltage, and the excellent algorithm results are continuously updated, iterated and fused, so that the lithium ion battery has excellent prediction performance based on a data model even before the capacity attenuation begins, the prediction result is more accurate, and the calculation process is more efficient. The invention provides efficient and reliable decision reference for the operation management of the lithium ion battery, improves the management efficiency and supports the prediction maintenance and guarantee of the lithium ion battery.
Claims (10)
1. A method for predicting the expected life of a lithium ion battery is characterized by comprising the following steps:
s1, collecting the measured data of the battery pack and cleaning and drying the collected data;
s2, analyzing the battery life of each battery in the battery pack according to the discharge capacity, charge capacity, voltage, current and temperature factors, analyzing the discharge capacity and voltage curve relation of the battery pack according to cleaned and dried measured data, and then performing data transformation on the discharge capacity and voltage curve relation to obtain model input initial characteristics and a label vector;
s3, respectively carrying out logarithm processing on the initial features and the label vectors, and then carrying out standardization processing;
s4, selecting the data after standardization processing to accord with the optimal parameters of the battery life prediction model parameters by setting grid searching and cross validation;
s5, selecting characteristic variables from the obtained optimal parameters as input data to be respectively brought into multiple models for training, searching for optimal model parameters through multiple loop iterations, and then selecting a model with the best fitting effect to perform result fusion, thereby obtaining a battery expected life prediction model;
and S6, inputting the time variable parameters of the battery to be tested and the data obtained by data transformation into a battery expected life prediction model as input characteristics, and predicting and acquiring the cycle period of the lithium ion battery through the battery expected life prediction model.
2. The method of claim 1, wherein the step of cleaning and drying the collected measured data of the battery pack comprises data drying, missing value processing and format content processing.
3. The method for predicting the expected life of the lithium ion battery according to claim 1, wherein high-frequency noise data are removed by a box line graph and a wavelet threshold method during data cleaning.
4. The method for predicting the expected life of the lithium ion battery according to claim 1, wherein the relation between the discharge capacity and the voltage curve is subjected to data transformation according to the relation between the discharge capacity and the voltage curve of the battery pack to be tested, time variable parameters and parameters after data transformation are selected as input characteristics and input into a battery expected life prediction model, the cycle period of the lithium ion battery is used as output data, and the expected life of the lithium ion battery is predicted by selecting and fusing a plurality of models.
5. The method of claim 4, wherein the time variable parameters include lithium ion battery discharge capacity, voltage, charge time, battery internal resistance and temperature.
6. The method according to claim 1, wherein the normalization of the data set is performed in a non-dimensionalization manner.
7. The method of claim 6, wherein the non-dimensionalization process adopts a z-score normalization (zero-mean normalization) calculation method, and the conversion formula is as follows:
wherein x isiAs the original data, it is the original data,
is the mean of the raw data, wherein
n is the number of samples and s is the standard deviation of the original data, where
8. The method according to claim 1, wherein the sample data is extracted from the data after the normalization processing, and is divided into a training set, a test set and a second test set, and the battery life expectancy prediction model is evaluated by using the second test data set.
9. The method of claim 8, wherein the model is evaluated by selecting a root mean square error and a percentage of the average error, wherein the root mean square error is a square root of a ratio of a sum of squares of deviations between observed values and true values to observation times, and is used for measuring a deviation between the observed values and the true values, and wherein the root mean square error RMSE is defined as:
the average percent error is defined by the formula:
where yi is the observed cycle life,
is the predicted cycle life and n is the total number of samples.
10. A lithium ion battery expected life prediction system based on the prediction method of claim 1 is characterized by comprising a data preprocessing module and a prediction module,
the data preprocessing module is used for cleaning and drying acquired data, then analyzing the service life of the battery for the discharge capacity, the charge capacity, the voltage, the current and the temperature factor of each battery in the battery pack, analyzing the relation between the discharge capacity and the voltage curve of the battery pack according to the cleaned and dried actually measured data, and then performing data transformation on the relation between the discharge capacity and the voltage curve to obtain the model input initial characteristic and the label vector; and then respectively carrying out logarithm processing on the initial characteristic and the label vector, then carrying out standardization processing, finally selecting the optimal parameters of which the data after the standardization processing accords with the parameters of the battery life prediction model by setting grid search and cross validation, selecting characteristic variables as input data by using the obtained optimal parameters to be respectively brought into various models for training, searching the optimal model parameters through repeated cycle iteration, then selecting the model with the best fitting effect to carry out result fusion, thereby obtaining the battery expected life prediction model, wherein the prediction module is used for storing the battery expected life prediction model, and predicting the life of the battery to be tested according to the time variable parameters of the battery to be tested and the data obtained by data transformation and outputting the prediction result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110938372.9A CN113671394A (en) | 2021-08-16 | 2021-08-16 | Lithium ion battery expected life prediction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110938372.9A CN113671394A (en) | 2021-08-16 | 2021-08-16 | Lithium ion battery expected life prediction method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113671394A true CN113671394A (en) | 2021-11-19 |
Family
ID=78543091
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110938372.9A Pending CN113671394A (en) | 2021-08-16 | 2021-08-16 | Lithium ion battery expected life prediction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113671394A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114047452A (en) * | 2022-01-13 | 2022-02-15 | 浙江玥视科技有限公司 | Method and device for determining cycle life of battery |
| CN114462309A (en) * | 2022-01-19 | 2022-05-10 | 苏州浪潮智能科技有限公司 | Physical simulation model training method, device, equipment and storage medium |
| CN115453399A (en) * | 2022-08-26 | 2022-12-09 | 广东工业大学 | Battery pack SOH estimation method considering inconsistency |
| CN115544903A (en) * | 2022-11-29 | 2022-12-30 | 深圳市菲尼基科技有限公司 | Battery data processing method, device and equipment based on big data and storage medium |
| CN115619106A (en) * | 2022-12-19 | 2023-01-17 | 中国人民解放军火箭军工程大学 | Method and system for determining quantity of spare parts of laser gyroscope in consideration of performance degradation |
| CN115856646A (en) * | 2023-03-03 | 2023-03-28 | 中山大学 | Lithium ion battery early life prediction method based on three-dimensional voltage characteristics |
| CN117452236A (en) * | 2023-12-26 | 2024-01-26 | 深圳市北测检测技术有限公司 | Method and system for detecting service life of battery of new energy automobile |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103344923A (en) * | 2013-08-01 | 2013-10-09 | 哈尔滨工业大学 | EKF (extended Kalmar filter)-method and NSDP-AR-model-based method for predicting cycling life of fusion-type lithium ion battery |
| US20160239586A1 (en) * | 2013-08-28 | 2016-08-18 | Hitachi, Ltd. | Lifetime prediction method and system of lithium-ion battery |
| CN107403242A (en) * | 2017-08-10 | 2017-11-28 | 四川大学 | Based on the equipment method for predicting residual useful life for improving tasteless particle filter |
| US20190113577A1 (en) * | 2017-10-17 | 2019-04-18 | The Board Of Trustees Of The Leland Stanford Junior University | Data-driven Model for Lithium-ion Battery Capacity Fade and Lifetime Prediction |
| CN112765772A (en) * | 2020-12-25 | 2021-05-07 | 武汉理工大学 | Power battery residual life prediction method based on data driving |
-
2021
- 2021-08-16 CN CN202110938372.9A patent/CN113671394A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103344923A (en) * | 2013-08-01 | 2013-10-09 | 哈尔滨工业大学 | EKF (extended Kalmar filter)-method and NSDP-AR-model-based method for predicting cycling life of fusion-type lithium ion battery |
| US20160239586A1 (en) * | 2013-08-28 | 2016-08-18 | Hitachi, Ltd. | Lifetime prediction method and system of lithium-ion battery |
| CN107403242A (en) * | 2017-08-10 | 2017-11-28 | 四川大学 | Based on the equipment method for predicting residual useful life for improving tasteless particle filter |
| US20190113577A1 (en) * | 2017-10-17 | 2019-04-18 | The Board Of Trustees Of The Leland Stanford Junior University | Data-driven Model for Lithium-ion Battery Capacity Fade and Lifetime Prediction |
| CN112765772A (en) * | 2020-12-25 | 2021-05-07 | 武汉理工大学 | Power battery residual life prediction method based on data driving |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114047452A (en) * | 2022-01-13 | 2022-02-15 | 浙江玥视科技有限公司 | Method and device for determining cycle life of battery |
| CN114047452B (en) * | 2022-01-13 | 2022-05-13 | 浙江玥视科技有限公司 | Method and device for determining cycle life of battery |
| CN114462309A (en) * | 2022-01-19 | 2022-05-10 | 苏州浪潮智能科技有限公司 | Physical simulation model training method, device, equipment and storage medium |
| CN114462309B (en) * | 2022-01-19 | 2023-11-17 | 苏州浪潮智能科技有限公司 | Physical simulation model training method, device, equipment and storage medium |
| CN115453399A (en) * | 2022-08-26 | 2022-12-09 | 广东工业大学 | Battery pack SOH estimation method considering inconsistency |
| CN115544903A (en) * | 2022-11-29 | 2022-12-30 | 深圳市菲尼基科技有限公司 | Battery data processing method, device and equipment based on big data and storage medium |
| CN115619106A (en) * | 2022-12-19 | 2023-01-17 | 中国人民解放军火箭军工程大学 | Method and system for determining quantity of spare parts of laser gyroscope in consideration of performance degradation |
| CN115856646A (en) * | 2023-03-03 | 2023-03-28 | 中山大学 | Lithium ion battery early life prediction method based on three-dimensional voltage characteristics |
| CN117452236A (en) * | 2023-12-26 | 2024-01-26 | 深圳市北测检测技术有限公司 | Method and system for detecting service life of battery of new energy automobile |
| CN117452236B (en) * | 2023-12-26 | 2024-03-15 | 深圳市北测检测技术有限公司 | Method and system for detecting service life of battery of new energy automobile |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113671394A (en) | Lithium ion battery expected life prediction method and system | |
| Chang et al. | Lithium-ion battery state of health estimation using the incremental capacity and wavelet neural networks with genetic algorithm | |
| CN112904219B (en) | Big data-based power battery health state prediction method | |
| Song et al. | Remaining life prediction of lithium-ion batteries based on health management: A review | |
| CN111832221B (en) | Lithium battery life prediction method based on feature screening | |
| CN113065283A (en) | Battery life prediction method, system, electronic device and storage medium | |
| Zhou et al. | Research on state of health prediction model for lithium batteries based on actual diverse data | |
| CN111983474A (en) | Lithium ion battery life prediction method and system based on capacity decline model | |
| CN116298934B (en) | Modeling method of prediction network for lithium battery health state estimation | |
| CN113484774A (en) | Lithium battery pack capacity estimation method based on temperature calibration and neural network | |
| Yao et al. | Novel lithium-ion battery state-of-health estimation method using a genetic programming model | |
| CN114167284A (en) | Lithium battery RUL prediction method and device based on BMS big data and ensemble learning | |
| Tan et al. | Intelligent online health estimation for lithium-ion batteries based on a parallel attention network combining multivariate time series | |
| Yang et al. | Critical summary and perspectives on state-of-health of lithium-ion battery | |
| Wang et al. | Dynamic early recognition of abnormal lithium-ion batteries before capacity drops using self-adaptive quantum clustering | |
| Zhang et al. | Remaining useful life and state of health prediction for lithium batteries based on differential thermal voltammetry and a deep learning model | |
| CN112287597B (en) | Lead-acid storage battery SOH estimation method based on VPGA-GPR algorithm | |
| Mazzi et al. | Lithium-ion battery state of health estimation using a hybrid model based on a convolutional neural network and bidirectional gated recurrent unit | |
| CN113657030A (en) | Method for predicting remaining service life of lithium battery based on Gaussian process regression | |
| CN116736133A (en) | Early prediction method for capacity degradation track of lithium ion battery in full life cycle | |
| CN115219907A (en) | Lithium battery SOC estimation method, system, medium, equipment and terminal | |
| Guo et al. | Health prognostics of lithium-ion batteries based on universal voltage range features mining and adaptive multi-Gaussian process regression with Harris Hawks optimization algorithm | |
| Zhao et al. | Lithium-ion battery state-of-health estimation method using isobaric energy analysis and pso-lstm | |
| Sui | Robust State of Health Estimation for Lithium-Ion Batteries Using Machines Learning | |
| Hassanieh et al. | Orthogonal autoencoders for long-term state-of-charge forecasting of Li-ion battery cells |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20211119 |
|
| RJ01 | Rejection of invention patent application after publication |