CN114578234A - Lithium ion battery degradation and capacity prediction model considering causality characteristics - Google Patents

Lithium ion battery degradation and capacity prediction model considering causality characteristics Download PDF

Info

Publication number
CN114578234A
CN114578234A CN202210276095.4A CN202210276095A CN114578234A CN 114578234 A CN114578234 A CN 114578234A CN 202210276095 A CN202210276095 A CN 202210276095A CN 114578234 A CN114578234 A CN 114578234A
Authority
CN
China
Prior art keywords
capacity
model
battery
causal
characteristic
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
Application number
CN202210276095.4A
Other languages
Chinese (zh)
Inventor
田毅
吴立锋
贺加贝
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Capital Normal University
Original Assignee
Capital Normal University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Capital Normal University filed Critical Capital Normal University
Priority to CN202210276095.4A priority Critical patent/CN114578234A/en
Publication of CN114578234A publication Critical patent/CN114578234A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01RMEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/36Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
    • G01R31/367Software therefor, e.g. for battery testing using modelling or look-up tables
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01RMEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/36Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
    • G01R31/392Determining battery ageing or deterioration, e.g. state of health
    • YGENERAL 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
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02EREDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E60/00Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
    • Y02E60/10Energy storage using batteries

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Charge And Discharge Circuits For Batteries Or The Like (AREA)
  • Secondary Cells (AREA)

Abstract

The invention provides a lithium ion battery degradation and capacity prediction model considering causal characteristics, which extracts and screens the causal characteristics, uses the causal characteristics as the input of the model to predict the available capacity of a battery, enhances the interpretability of the model in the aspects of prediction and characteristics, and improves the prediction precision of the capacity. The method specifically comprises the following steps: (1) aiming at the problem that the causal relationship between input parameters and output is not sufficiently explained by other neural network models, the capacity is predicted by considering causal characteristics, and the interpretability of the models is improved. (2) The influence of the characteristics on the capacity attenuation is analyzed by adopting an impulse response analysis method, and the exact influence and action mechanism of the selected characteristics on the capacity attenuation are determined by combining the battery aging mechanism analysis, so that the interpretability of the characteristics is improved. (3) The causal characteristics are used as input, a capacity prediction model based on the long-time memory LSTM network is constructed, and compared with a prediction result using original characteristics and prediction results of other methods, more accurate prediction of capacity is achieved.

Description

Lithium ion battery degradation and capacity prediction model considering causality characteristics
Technical Field
The invention provides a lithium ion battery degradation and capacity prediction model considering causality characteristics, and belongs to the technical field of new energy batteries.
Background
The lithium ion battery has the advantages of large storage capacity, multiple charging and discharging times, recoverability, small pollution and the like, so that the lithium ion battery becomes a new energy source which is widely applied to multiple fields such as electric automobiles, aerospace, computers and the like. However, due to the influence of the working environment and the working conditions, a series of complicated changes occur inside the battery, causing various side reactions, and the capacity is gradually reduced, thereby causing some safety problems. Therefore, accurate life estimation is crucial to the safety and reliability of lithium ion batteries, and battery life prediction is an important research direction in this field. With the rise of artificial intelligence and machine learning, a lithium ion battery life prediction method based on data driving is becoming a mainstream and effective method. The prediction method based on data driving extracts features from the collected battery data as input variables, uses the battery life as a target variable, and learns related knowledge through a model, thereby predicting the battery life, such as soc (state Of charge), soh (state Of health), rul (remaining Useful life), capacity, and the like. Although the existing method obtains good prediction effect, the causal relationship between the input and the output of the model cannot be well explained, so how to enhance the interpretability of the model is a problem to be solved urgently.
Data-driven Machine Learning prediction methods, such as Support Vector Machines (SVMs), Extreme Learning Machines (ELMs), Gaussian regression processes (GPRs), Deep Neural Networks (DNNs), and wide Learning systems (BLSs), have achieved good prediction results. There are also some methods for setting and selecting the input features of the model, such as setting the variance and kurtosis of the raw data, Shannon entropy and Hausdorff distance of the charging curve, etc., and random segments of the charging curve as the input features of the model. In feature selection, features are screened using methods such as Pearson and Spearman correlation coefficients, Random Forest (RF), Gray correlation Analysis (GRA), or cross-validated recursive feature elimination.
Although the selection of input features by existing feature setting and extraction methods shows good results in terms of battery life prediction, whether there is a causal relationship between the input features and the predicted target variables is not well explained or verified, i.e., whether there is a causal relationship between the model "input and output" is not explained, and there is a problem that model interpretability is not sufficient.
Disclosure of Invention
Aiming at the problem that the existing prediction model can not be sufficiently explained, the method considers extracting and screening the causal characteristics which have influence on the battery capacity degradation, and uses the causal characteristics as the input of the model to predict the available capacity of the battery, so as to enhance the interpretability of the model in the aspects of prediction and characteristics and improve the prediction precision of the capacity. The core technical points of the invention are as follows:
(1) aiming at the problem that the causal relationship between input parameters (or characteristics) and output is not sufficiently explained by other neural network models, the invention provides the method for predicting the capacity by considering the causal characteristics, so that the interpretability of the model is improved.
(2) The influence of the characteristics on the capacity attenuation is analyzed by adopting an impulse response analysis method, and the exact influence and action mechanism of the selected characteristics on the capacity attenuation are determined by combining the battery aging mechanism analysis, so that the interpretability of the characteristics is improved.
(3) A capacity prediction model based on a Long Short-Term Memory (LSTM) network is constructed by taking causal characteristics as input, and more accurate prediction of capacity is realized compared with a prediction result using original characteristics and prediction results of other methods.
To further clarify the technical solution of the present invention, the following details are described below:
the present invention considers selection of causal features in the capacity degradation process to predict capacity. Specifically, the invention provides a lithium ion battery degradation and capacity prediction model based on a Vector Autoregressive (VAR) model and an LSTM, causal characteristics in the lithium ion battery capacity attenuation process can be selected, specific influences of the causal characteristics on the capacity attenuation are analyzed, the interpretability of the characteristics is enhanced, then the battery capacity is predicted by using the selected causal characteristics, the interpretability of the model in a prediction angle is enhanced, and the prediction accuracy of the capacity is improved. The technical scheme of the method provided by the invention comprises the following steps:
s1, extracting initial characteristics
First, voltage, current, temperature and time data obtained by the test device during each charge and discharge cycle of the battery is recorded as V, I, T, t. And then carrying out primary data processing (data cleaning, dimension reduction, normalization and standardization) on the initial data to construct initial characteristics of all cycles of the battery. Specifically, a total of 8 initial signatures, denoted F1-F8, were constructed by processing four physical data per cycle, so that the overall signature dimension was 8 × N, where N represents the total number of cycles.
S2, battery characteristic screening model
Constructing a feature screening system based on a VAR model and Granger Causality (GC) inspection, screening all initial features (F1-F8) of the battery constructed in S1, extracting the battery capacity as target variables, and respectively recording the features and the capacity as yiWherein i is 1, 2. The method comprises the following specific steps:
s2.1, assuming that all initial characteristics (F1-F8) and capacities are endogenous variables, constructing a multivariate VAR (p) system for the Glanker causal relationship test, wherein p represents the VAR model order, and the model is shown as follows:
Yt=Bt1Yt-12Yt-1+...+αpYt-p+et
wherein
Figure BDA0003556031360000041
Yt-pThen represents the variable YtThe lag value at time t-p. Specifically, yi,tRepresenting 8 signatures (F1-F8) and capacity data at time t (cycle), respectively. And α, b, e represent the weight parameter, constant term parameter and error term parameter in the model calculation process, respectively. The model performs regression analysis by using the hysteresis value of the capacity and the hysteresis value of the initial feature, and thus can be used to analyze the influence relationship between the endogenous variable and the target variable. The invention also utilizes the characteristic to analyze the initial characteristics of the batteryAnd the relationship between the characteristic and the battery capacity, and performing primary screening on the characteristics.
S2.2 taking the capacity as a target variable, carrying out the Glanker causal relationship test, and carrying out analysis and verification on the causal relationship between the assumed health factor and the capacity. The GC evaluates how past values of one variable result in another variable. Taking the characteristic F1 and capacity as an example, the characteristic F1 has a glange causal relationship with capacity (hysteresis order l, k) if and only if:
f(y9,t|y9,t-k,y1,t-l)≠f(y9,t|y9,t-k)
from the above definitions, to describe the screening principle more clearly, the original assumption that the feature F1 and the capacity have grave causality is that, according to the model formula in S2.1:
H0:α19,1=α19,2=…=α19,p=0
h1 of at least one alpha19,iIs not 0
If the original hypothesis is not rejected, y1,tFor y9,tThere is no granger causality, i.e., feature F1 has no predictive capability for battery capacity data; otherwise if the original hypothesis is rejected, y1,tFor y9,tThe data has granger causality, i.e., the characteristic F1 has predictive ability for battery capacity data. The hypothesis test can be completed by Chi's test, and the test statistic asymptotically obeys x2(p)。
S2.3, changing the order p, reconstructing a VAR model, repeating the steps S2.1-2.2, and completing all the checking processes of the single characteristic F1 of the single battery.
S2.4 steps S2.2-2.3 are repeated using the other characteristics of the monoblock battery (F2-F8) as explanatory variables until all characteristic checks of this battery are completed.
S2.5 repeats steps S2.2-2.4 until all feature tests for all cells are completed.
S2.6, all detection results are integrated for screening to obtain a screening result of causal characteristics. The invention carries out GC inspection on all the characteristics of a plurality of batteries for a plurality of times, thereby eliminating the randomness of single inspection results and single battery results, increasing the persuasion of the characteristic selection results and leading the characteristic selection results to have more objectivity.
S3, battery characteristic causality analysis model
According to the step S2, causal characteristic screening results can be obtained, so that the invention constructs a battery characteristic causal analysis system based on VAR system and system Impulse Response Functions (IRF) at this step, and analyzes the influence of the screened characteristics on the battery capacity, that is, the specific influence and action mechanism of the characteristics in the battery capacity degradation process, so as to further verify the selected causal characteristics of the battery capacity degradation. The method comprises the following steps:
s3.1, according to the preliminary screening result of the characteristics and the detection result in the S2, a plurality of multivariable VAR systems are reconstructed, and each system is used for analyzing the influence relationship of the battery causal characteristics on the battery capacity degradation.
S3.2 before the characteristic causality of the impulse response analysis is carried out, stability check is carried out on the constructed VAR model. The test method comprises the following steps:
by using the multivariate VAR (p) model formula in S2.1, k (from time t-k +1 to time t) multivariate VAR models can be obtained by setting p to t, t-1, …, and t-k + 1. If the block matrix is expressed, let:
Figure BDA0003556031360000061
the k VAR (p) models can be written as VAR (1) by friend matrix change, i.e.:
Zt=C+φZt-1t
this model is also the characteristic causality analysis model of the battery. Specifically, Yt8 signatures (F1-F8) representing time t (cycles) and capacity data, Yt-pThen represents the variable YtHysteresis value at time t-p, and Bt、etConstant term parameter and error term parameter, α, of the representative time-t VAR modeljThen, the weight parameter matrix of the jth var (p) model is represented, where j is 1, 2. Then, in the constructed VAR (1), C and θtA constant term parameter matrix and an error term parameter matrix, Z, respectively representing the analytical modeltAnd Zt-1Then the current and hysteresis values of the selected battery causal characteristic and capacity configuration matrix are represented.
For the integrated multivariate VAR model constructed currently, the condition of model stability is the total weight parameter matrix
Figure BDA0003556031360000062
All the eigenvalues of (A) are all within the unit circle, i.e. their eigen equation
Figure BDA0003556031360000063
Is within the unit circle. And if the VAR system is unstable, changing the model order and reconstructing the VAR system. If the model is stable, impulse response analysis can be carried out, and the specific influence of causal characteristics on the battery capacity attenuation is analyzed.
S3.3 the specific impact of each causal feature on capacity in the model can be analyzed by the impulse response function of the battery feature causal analysis system VAR (1) constructed in S3.2. The impulse response function describes the influence of a standard deviation impact on the current value and the future value of an endogenous variable after a random error term is applied, so that the impulse response function can be used for analyzing the influence relationship of a time series variable on another time series variable in time sequence. First available from VAR (1):
Figure BDA0003556031360000071
order to
Figure BDA0003556031360000072
The following can be obtained:
Zt+r=θt+r1θt+r-12θt+r-2+…+ψrθt+…
then there are:
Figure BDA0003556031360000073
wherein ZtAnd Zt+rThen the values at times t and t + r of the battery causal characteristics and capacity matrix are represented and theta represents the error term parameter matrix of the model. E represents an identity matrix, L and
Figure BDA0003556031360000074
representing the parameter matrix in the model transformation and calculation process. PsirAn impulse response function representing the causality analysis system model of the whole battery characteristic, wherein the element of the ith row and the jth column is regarded as a function of the lag period number r, and the element represents that other error terms are not changed in any period when the jth variable y isj,tError term e oftAfter receiving a unit of impact at time t, for the ith endogenous variable yi,tThe effect caused during the t + r period. In the invention, the battery capacity is taken as a target variable, and the impulse response result of the capacity to the selected causal characteristic is obtained through the IRF of the battery causal analysis system.
S3.4 the pulse response analysis result is explained by combining the mechanism of the internal capacity degradation of the lithium ion battery, and the characteristic interpretability is further enhanced.
S4, estimating battery capacity
S4.1, constructing a capacity prediction model. The invention constructs a stacked LSTM network for capacity prediction. The network structure has multiple levels: input layer, LSTM layer, Dropout layer, Dense layer and output layer, and the network structure is shown in FIG. 3. The input layer is responsible for receiving input data and converting the data into a format acceptable by the LSTM layer, and a sliding window technology with a 'many-to-one' structure is used in the layer, so that the output and the input of the network are related at multiple moments to better learn the time information of the sequence. The LSTM layer is responsible for learning the characteristic information of the data transmitted by the input layer, and the non-linear relationship of this information to the predictor variables (capacities). The Dropout layer is responsible for carrying out random inactivation treatment on the neurons of the LSTM layer, and prevents the model from generating an overfitting phenomenon to reduce the prediction effect. The Dense layer is responsible for performing dimension transformation on the output data of the LSTM layer-Dropout layer. And finally, receiving the transformation data of the Dense layer by an output layer, and outputting a prediction result, wherein the output of the network is the predicted capacity of the corresponding life cycle of the battery.
S4.2 the battery capacity is predicted according to S3 using the LSTM prediction model constructed in S4.1 with the determined causal characteristics as input data. Specifically, firstly, input data is converted into a format which is suitable for being accepted by an LSTM layer through a sliding window technology at an input layer; secondly, learning characteristic information of data transmitted by an input layer on an LSTM layer, and keeping weight parameters of the LSTM layer; then, discarding a part of the calculation results of the LSTM neurons by using Dropout technology in a Dropout layer; then, carrying out dimension reduction on the output result of the LSTM layer subjected to Dropout processing in the Dense layer, and transmitting the conversion result to the output layer; finally, the output layer outputs the capacity prediction result
Figure BDA0003556031360000081
And S4.3, obtaining a capacity prediction result.
The invention has the advantages and beneficial effects that: the method uses the causality characteristics in the battery degradation process to predict the available capacity of the battery, not only realizes more accurate prediction of the battery capacity, but also improves the interpretability of the model in the aspects of prediction and characteristics, which is an important research field of machine learning at present, namely the interpretability problem. In addition, the method can be popularized to the field of other types of batteries or service life prediction, so that the method has important academic significance and has potential engineering application value for improving the reliability and safety of new energy electronic equipment.
Drawings
Fig. 1 is a flow chart of an overall battery degradation and capacity prediction model.
Fig. 2 is a schematic diagram of the model.
Fig. 3 is a block diagram of a capacity prediction model constructed based on LSTM.
Fig. 4 is a characteristic graph extracted from the test data of a certain cell in the NASA public data set used in the examples.
Fig. 5 is a graph of the degradation of a certain battery capacity in the NASA-published data set used in the examples.
Fig. 6 is a diagram illustrating an example of a stability check result of a certain VAR system.
Fig. 7 is an exemplary graph of the results of the characteristic impulse response analysis.
Fig. 8a is a diagram of the results of predicting the capacity of a certain battery using causal characteristics and primitive characteristics as model input characteristics.
Fig. 8b is a prediction error map of the result of predicting the capacity of a certain battery using the causal characteristics and the raw characteristics as model input characteristics.
Detailed Description
The invention is further described below with reference to the figures and examples. As shown in fig. 1-2, the present invention provides a model for predicting lithium ion battery degradation and capacity that takes causal characteristics into account. The present invention uses the public battery degradation dataset of NASA Ames Center of Excellence as an example dataset. The data set contained 4 18650 batteries B5, B6, B7, and B18, with three different operations of charging, discharging, and impedance performed at room temperature, and data collected. The experiment was stopped when the rated capacity of the battery decreased by 30% (from 2Ah to 1.4 Ah). The method comprises the following specific steps:
s1, extracting initial characteristics. First, the initial characteristics of all cycles of the battery are constructed based on the battery data obtained by the monitoring device. The characteristic curve of the discharge voltage data of the battery No. B5 in the data set is shown in fig. 4.
And S2, primarily screening characteristics.
S2.1 first, battery capacity data is extracted, and a capacity degradation curve of the battery No. B5 is shown in fig. 5.
S2.2, with capacity as a target variable and all characteristic factors and capacity as endogenous variables, constructing a multivariate VAR (p) system to carry out the Greenjek causal relationship test. Certain characteristics of the B5 size battery the results of one test are shown in the following table:
Feature Chi-sq df Prob.
discharge current 33.1035 10 0.0003
Wherein Chi-sq represents Chi-squared test value of the original hypothesis, df represents VAR model order, and Prob represents probability of rejecting the original hypothesis. The present invention sets the test significance level to 0.05, indicating that the probability of the selected variable from the target variable being the endogenous variable is at least 95%, i.e., there is a granger causal relationship. In this test, the discharge current was initially selected as a causal characteristic.
S2.3, changing the order, reconstructing a VAR model, repeating the steps S2.1-2.2, and completing the single characteristic inspection process of the single battery.
S2.4 repeats steps S2.1-2.3 until all feature tests for this cell are completed.
S2.5 repeats steps S2.2-2.4 until all feature tests for all cells are completed.
S2.6, all the detection results are integrated for screening to obtain a primary screening result.
And S3, characteristic causality analysis.
S3.1 reconstructing a plurality of multivariate VAR (p) systems according to the primary screening of the characteristics and the result combined with the test result in S2, and analyzing the influence of the screened characteristics on the capacity degradation.
S3.2, before the impulse response is carried out, stability test needs to be carried out on the constructed VAR model, the stability test result of a certain VAR system is shown in figure 6, the system is stable, and the impulse response analysis can be carried out.
S3.3, analyzing the specific influence of the screening characteristics on the capacity in the model by combining the impulse response results of a plurality of systems through the impulse response function of the systems, wherein an impulse response graph of the capacity on the discharge voltage in a certain system is shown in figure 7. From this result, it can be seen that the influence of the discharge voltage on the degradation of the capacity, although fluctuating, is substantially negative, and therefore, the discharge voltage is preliminarily selected as a characteristic having an influence on the degradation of the capacity according to the result of the impulse response analysis.
S3.4 explains the results of impulse response analysis in conjunction with the mechanism of internal capacity degradation in lithium ion batteries. As analyzed for the impulse response results in S3.3, the discharge cutoff voltage of the present cell was greater than the maximum discharge voltage (2.75V), and the cell was in an "over-discharge" state in each cycle. Such operating conditions can cause negative effects such as corrosion of the cell's collector and loss of active material, which can lead to degradation of the cell's capacity. Therefore, in conjunction with the impulse response results, the discharge voltage is considered to be one of the causes of the degradation of the battery capacity.
S3.5, determining the final causal characteristics. In the present embodiment, the final charge-discharge temperature, discharge voltage, current, time are preliminarily screened as characteristics having an influence on the capacity degradation.
S4. Capacity estimation
S4.1, constructing a capacity prediction model.
And S4.2, processing the data and rewriting the data dimension. At S3, the battery capacity is predicted using the causal characteristic as an input characteristic of the prediction model.
And S4.3, obtaining a capacity prediction result, wherein the prediction result of a certain battery is shown in FIGS. 8(a) and 8(b), and FIGS. 8(a) and 8(b) show the prediction result and the corresponding prediction error by using the original characteristic and the causal characteristic. It can be seen that the capacity prediction result using causal features is closer to the true degradation curve and the prediction error is closer to 0 compared to the original features, verifying the effectiveness of the picking method and picking features. Comparison with other methods is shown in the following table:
Figure BDA0003556031360000121
wherein MAE (mean absolute error) and RMSE (root mean square error) are common evaluation indexes of regression prediction problems. m is the predicted cycle number, Qi and
Figure BDA0003556031360000122
for the actual capacity and predicted capacity of the battery at the ith cycle, the calculation method is as follows:
Figure BDA0003556031360000123
Figure BDA0003556031360000124
it can be seen that the MAE, RMSE of the method of the invention are minimal, demonstrating the effectiveness of the proposed method.

Claims (6)

1. A lithium ion battery degradation and capacity prediction model considering causality characteristics is characterized by comprising the following specific steps:
s1, extracting initial characteristics:
firstly, voltage, current, temperature and time data obtained by detection equipment in each charge-discharge cycle process of the battery are recorded as V, I, T, t; then, carrying out primary data processing on the initial data to construct initial characteristics of all cycles of the battery;
s2, battery characteristic screening model:
constructing a characteristic screening system based on a VAR model and a Glankey causal relationship GC test, and screening all initial characteristics of the battery constructed in S1;
s3, a battery characteristic causality analysis model:
obtaining a causal characteristic screening result according to S2, constructing a battery characteristic causal analysis system based on a VAR system and a system Impulse Response Function (IRF), analyzing the influence of the screened characteristics on the battery capacity, namely the specific influence and action mechanism of the characteristics in the battery capacity degradation process, and further verifying the selected causal characteristics of the battery capacity degradation;
and S4, estimating the capacity of the battery.
2. The model of claim 1, wherein the model is for predicting lithium ion battery degradation and capacity based on causal characteristics, and comprises: in step S1, a total of 8 initial signatures, denoted as F1-F8, are constructed by processing the four types of physical data for each cycle, so that the overall signature dimension is 8 × N, where N represents the total number of cycles.
3. The model of claim 1, wherein the model is for predicting lithium ion battery degradation and capacity based on causal characteristics, and comprises: in step S2, the battery capacity is extracted as a target variable, and the characteristics and the capacity are respectively denoted as yiWherein i is 1, 2.
4. The model of claim 1,2 or 3 for predicting lithium ion battery degradation and capacity considering causal characteristics, wherein: in step S2, the method further comprises the following steps:
s2.1: setting all initial characteristics F1-F8 and capacity as endogenous variables, constructing a multivariate VAR (p) system for carrying out Glanker causal relationship test, wherein p represents the VAR model order, and the model is shown as the following formula:
Yt=Bt1Yt-12Yt-1+…+αpYt-p+et
wherein
Figure FDA0003556031350000021
Yt-pRepresents the variable YtA hysteresis value at time t-p; y isi,tRespectively representing 8 characteristics F1-F8 and capacity data at the time t; and alpha, b, e are respectively substitutedThe table represents a weight parameter, a constant term parameter and an error term parameter in the calculation process of the model; the battery capacity is subjected to regression analysis by using the hysteresis value of the capacity and the hysteresis value of the initial characteristic, so that the battery capacity is used for analyzing the influence relationship between the interpretation variable and the target variable and primarily screening the characteristic;
s2.2: carrying out Glangel causal relationship test by taking the capacity as a target variable, and carrying out analysis and verification on the causal relationship between the assumed health factor and the capacity; the GC evaluates how past values of one variable result in another variable; characteristic F1 has a grand causal relationship with capacity if and only if:
f(y9,t|y9,t-k,y1,t-l)≠f(y9,t|y9,t-k)
from the model formula in S2.1, the primary assumption that the feature F1 and capacity have grand causality is:
H0:α19,1=α19,2=…=α19,p=0
h1 having at least one alpha19,iIs not 0
If the original hypothesis is not rejected, y1,tFor y9,tThere is no granger causality, i.e., feature F1 has no predictive power for battery capacity data; otherwise if the original hypothesis is rejected, y1,tFor y9,tThe characteristics F1 have the causality of Glankey, namely the characteristics F1 have the prediction capability on the battery capacity data; the hypothesis test is completed by chi-square test, and the test statistic asymptotically obeys chi2(p);
S2.3: changing the order p, reconstructing a VAR model, repeating the steps S2.1-2.2, and completing all the inspection processes of the single characteristic F1 of the single battery;
s2.4: repeating steps S2.2-2.3 using the other characteristics F2-F8 of the monoblock battery as explanatory variables until all characteristic checks of the battery are completed;
s2.5: repeating the step S2.2-2.4 until all the characteristics of all the batteries are checked;
s2.6: screening by integrating all detection results to obtain a screening result with causal characteristics; and performing GC (gas chromatography) inspection on all the characteristics of the plurality of batteries for multiple times, thereby eliminating the randomness of a single inspection result and a single battery result, increasing the persuasion of a characteristic selection result and enabling the characteristic selection result to have more objectivity.
5. The model of claim 4, wherein the model is used for predicting lithium ion battery degradation and capacity by considering causality characteristics, and comprises: in step S3, the method further comprises the following steps:
s3.1: reconstructing a plurality of multivariable VAR systems according to the preliminary screening results of the characteristics and combined with the detection results in S2, wherein each system is used for analyzing the influence relationship of the battery causal characteristics on the battery capacity degradation;
s3.2: before the characteristic causality of impulse response analysis is carried out, stability inspection needs to be carried out on the constructed VAR model; the test method comprises the following steps:
obtaining k multivariate VAR models by using a multivariate VAR (p) model formula in S2.1, wherein if p is t, t-1, …, t-k + 1; if the block matrix is expressed, let:
Figure FDA0003556031350000041
the k VAR (p) models are written as VAR (1) by friend matrix change, i.e.:
Zt=C+φZt-1t
this model is a characteristic causality analysis model of the battery; y ist8 features F1-F8 representing time t and volume data, Yt-pThen represents the variable YtHysteresis value at time t-p, and Bt、etConstant term parameter and error term parameter, α, of the representative time-t VAR modeljThen, a weight parameter matrix representing the jth var (p) model, where j is 1, 2.., k, E represents an identity matrix; then, in the constructed VAR (1), C and θtA constant term parameter matrix and an error term parameter matrix, Z, respectively representing the analytical modeltAnd Zt-1Then the current and hysteresis values of the selected battery causal characteristic and capacity construction matrix are represented;
for the integrated multivariate VAR model currently constructed, the condition of model stability is the total weight parameter matrix
Figure FDA0003556031350000042
All the eigenvalues of (A) are all within the unit circle, i.e. the eigenequation
Figure FDA0003556031350000043
Is within the unit circle; if the model is unstable, changing the order of the model, and reconstructing a VAR system; if the model is stable, performing impulse response analysis, and analyzing the specific influence of causal characteristics on the battery capacity attenuation;
s3.3: analyzing the specific influence of each causal feature in the model on the capacity through an impulse response function of a battery feature causal analysis system VAR (1) constructed in S3.2; the impulse response function describes the influence on the current value and the future value of an endogenous variable after a standard deviation impact is applied to a random error term, so that the impulse response function is used for analyzing the influence relation of a certain time series variable on another time series variable in time sequence; first obtained from VAR (1):
Figure FDA0003556031350000044
order to
Figure FDA0003556031350000045
Obtaining:
Zt+r=θt+r1θt+r-12θt+r-2+…+ψrθt+...
then there are:
Figure FDA0003556031350000051
wherein Z istAnd Zt+rThen represents electricityValues of the pool causality characteristic and the capacity matrix at t and t + r moments, and theta represents an error item parameter matrix of the model; e represents an identity matrix, L and
Figure FDA0003556031350000052
representing a parameter matrix in the process of model transformation and calculation; psirAn impulse response function representing the causality analysis system model of the whole battery characteristic, wherein the element of the ith row and the jth column is regarded as a function of the lag period number r, and the element represents that other error terms are not changed in any period when the jth variable y isj,tError term e oftAfter receiving a unit of impact at time t, for the ith endogenous variable yi,tThe effect caused during the t + r period; acquiring an impulse response result of the capacity to the selected causal characteristic through an IRF (interference rejection filter) of a battery causal analysis system by taking the battery capacity as a target variable;
s3.4: the interpretability of the characteristics is further enhanced by interpreting the results of the impulse response analysis in conjunction with the mechanism of internal capacity degradation in the lithium ion battery.
6. The model of claim 5, wherein the model is used for predicting lithium ion battery degradation and capacity by considering causality characteristics, and comprises: in step S4, the method further comprises the following steps:
s4.1: constructing a capacity prediction model; constructing a stacked LSTM network for capacity prediction; the network architecture has multiple levels: an input layer, an LSTM layer, a Dropout layer, a Dense layer and an output layer; the input layer is responsible for receiving input data and converting the data into a format acceptable by the LSTM layer, and a sliding window technology with a 'many-to-one' structure is used in the input layer, so that the output and the input of a network are related at a plurality of moments, and the time information of a sequence is learned; the LSTM layer is responsible for learning characteristic information of data transmitted by the input layer and the nonlinear relation between the information and a prediction variable; the Dropout layer is responsible for carrying out random inactivation treatment on the neurons of the LSTM layer, and the phenomenon that the model is over-fitted to reduce the prediction effect is prevented; the Dense layer is responsible for carrying out dimension transformation on output data of the LSTM layer-Dropout layer; finally, an output layer receives the transformed data of the Dense layer and outputs a prediction result, wherein the output of the network is the predicted capacity of the corresponding life cycle of the battery;
s4.2: according to step S3, the battery capacity is predicted using the LSTM prediction model constructed in S4.1, with the determined causal characteristics as input data; firstly, converting input data into a format which is suitable for an LSTM layer to be accepted by an input layer through a sliding window technology; secondly, learning characteristic information of data transmitted by an input layer on an LSTM layer, and keeping weight parameters of the LSTM layer; then, discarding a part of the calculation results of the LSTM neurons by using Dropout technology in a Dropout layer; then, carrying out dimension reduction on the output result of the LSTM layer subjected to Dropout processing in the Dense layer, and transmitting the conversion result to the output layer; finally, the output layer outputs the capacity prediction result
Figure FDA0003556031350000061
S4.3: and obtaining a capacity prediction result.
CN202210276095.4A 2022-03-21 2022-03-21 Lithium ion battery degradation and capacity prediction model considering causality characteristics Pending CN114578234A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210276095.4A CN114578234A (en) 2022-03-21 2022-03-21 Lithium ion battery degradation and capacity prediction model considering causality characteristics

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210276095.4A CN114578234A (en) 2022-03-21 2022-03-21 Lithium ion battery degradation and capacity prediction model considering causality characteristics

Publications (1)

Publication Number Publication Date
CN114578234A true CN114578234A (en) 2022-06-03

Family

ID=81783388

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210276095.4A Pending CN114578234A (en) 2022-03-21 2022-03-21 Lithium ion battery degradation and capacity prediction model considering causality characteristics

Country Status (1)

Country Link
CN (1) CN114578234A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116298906A (en) * 2023-01-19 2023-06-23 四川新能源汽车创新中心有限公司 Battery capacity prediction model training method, prediction method, device and medium
CN116449209A (en) * 2023-01-12 2023-07-18 帕诺(常熟)新能源科技有限公司 Actual operation energy storage lithium capacitance prediction method based on LSTM
CN116804706A (en) * 2023-06-06 2023-09-26 淮阴工学院 Temperature prediction method and device for lithium battery of electric automobile

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116449209A (en) * 2023-01-12 2023-07-18 帕诺(常熟)新能源科技有限公司 Actual operation energy storage lithium capacitance prediction method based on LSTM
CN116298906A (en) * 2023-01-19 2023-06-23 四川新能源汽车创新中心有限公司 Battery capacity prediction model training method, prediction method, device and medium
CN116298906B (en) * 2023-01-19 2024-04-16 四川新能源汽车创新中心有限公司 Battery capacity prediction model training method, prediction method, device and medium
CN116804706A (en) * 2023-06-06 2023-09-26 淮阴工学院 Temperature prediction method and device for lithium battery of electric automobile

Similar Documents

Publication Publication Date Title
Sui et al. A review of non-probabilistic machine learning-based state of health estimation techniques for Lithium-ion battery
Khan et al. Batteries state of health estimation via efficient neural networks with multiple channel charging profiles
Tao et al. A novel support vector regression method for online reliability prediction under multi-state varying operating conditions
Li et al. Remaining useful life prediction for lithium-ion batteries with a hybrid model based on TCN-GRU-DNN and dual attention mechanism
Cui et al. A dynamic spatial-temporal attention-based GRU model with healthy features for state-of-health estimation of lithium-ion batteries
Fan et al. A novel machine learning method based approach for Li-ion battery prognostic and health management
Zhou et al. Battery health prognosis using improved temporal convolutional network modeling
CN114578234A (en) Lithium ion battery degradation and capacity prediction model considering causality characteristics
CN113065283A (en) Battery life prediction method, system, electronic device and storage medium
CN112834927A (en) Lithium battery residual life prediction method, system, device and medium
Su et al. Prediction of remaining useful life for lithium-ion battery with multiple health indicators
CN111707956B (en) Method for predicting health state and residual life of multi-type lithium ion battery pack management system
CN115201686B (en) Lithium ion battery health state assessment method under incomplete charge and discharge data
AU2021101964A4 (en) Artificial intelligence based smart electric vehicle battery management system
CN114966436A (en) Lithium battery state of charge prediction method, device, equipment and readable storage medium
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
Shen et al. Transfer learning-based state of charge and state of health estimation for Li-ion batteries: A review
Lin et al. Physics-informed deep learning for lithium-ion battery diagnostics using electrochemical impedance spectroscopy
CN117825965A (en) State prediction method and system for lithium ion battery
CN116359773A (en) Lithium ion battery remaining life prediction method and system
Zou et al. State of Health prediction of lithium-ion batteries based on temporal degeneration feature extraction with Deep Cycle Attention Network
CN113469013B (en) Motor fault prediction method and system based on transfer learning and time sequence
Zhang et al. Remaining useful life prediction of lithium-ion batteries based on TCN-DCN fusion model combined with IRRS filtering
Wang et al. A conditional random field based feature learning framework for battery capacity prediction
Wang et al. Assessing the Performance Degradation of Lithium‐Ion Batteries Using an Approach Based on Fusion of Multiple Feature Parameters

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