CN114547969A - EMD-MRVR-based multi-stress battery life prediction method - Google Patents

EMD-MRVR-based multi-stress battery life prediction method Download PDF

Info

Publication number
CN114547969A
CN114547969A CN202210082592.0A CN202210082592A CN114547969A CN 114547969 A CN114547969 A CN 114547969A CN 202210082592 A CN202210082592 A CN 202210082592A CN 114547969 A CN114547969 A CN 114547969A
Authority
CN
China
Prior art keywords
battery
stress
model
under
mrvr
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
CN202210082592.0A
Other languages
Chinese (zh)
Inventor
杨宁
王艺澎
余涛
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.)
South China University of Technology SCUT
Original Assignee
South China University of Technology SCUT
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 South China University of Technology SCUT filed Critical South China University of Technology SCUT
Priority to CN202210082592.0A priority Critical patent/CN114547969A/en
Publication of CN114547969A publication Critical patent/CN114547969A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • 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]
    • G01R31/392—Determining battery ageing or deterioration, e.g. state of health
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/04—Ageing analysis or optimisation against ageing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/12—Timing analysis or timing optimisation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
    • Y02E60/10—Energy storage using batteries

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Computer Hardware Design (AREA)
  • Geometry (AREA)
  • Secondary Cells (AREA)

Abstract

The invention discloses an EMD-MRVR-based multi-stress battery life prediction method. The method comprises the following steps: inputting time sequence data which can reflect battery aging and is obtained by performing tests under different stress conditions and a unified failure threshold value which is applicable under all stress conditions; preprocessing input time sequence data; decomposing the preprocessed time sequence data into a high-frequency connotation modal component and a low-frequency residual component by empirical mode decomposition, and fitting the residual component to obtain a battery aging empirical model; training a multi-output correlation vector regression model; inputting a new stress condition to the trained multi-output correlation vector regression model, and obtaining model parameters of the battery aging empirical model under the stress condition, so as to obtain the battery aging empirical model under the stress condition, further realize the battery service life prediction under the stress condition, and assist battery manufacturers and maintainers in production and maintenance.

Description

EMD-MRVR-based multi-stress battery life prediction method
Technical Field
The invention relates to the field of battery life prediction, in particular to a method for predicting battery life under multiple stresses based on EMD-MRVR.
Background
In order to better realize the management of the battery, the service life prediction of the battery needs to be researched, and because the battery under the actual operation working condition is not easy to obtain, a natural aging test is carried out by related organizations, so that the research on the aging rule of the battery is beneficially explored. However, since the natural aging test requires a long test time and is inefficient, the accelerated aging test (or the aging test of the battery performed under a multi-stress condition) is thought by researchers.
At present, batteries in the market are various in types, and working conditions (discharge depth, temperature, charge rate, discharge rate, cycle interval and charge-discharge cutoff voltage) are also different, so that aging tests cannot be carried out under all stress conditions, and because the aging of the batteries presents a nonlinear law, how to apply the aging law to different stress combination working conditions and further carry out service life prediction has certain difficulty.
The literature, "lithium power battery life research based on working condition simulation" introduces empirical formulas of equivalent capacity attenuation coefficients at 6 temperatures and 5 discharge rates, and on this basis, the temperatures and the discharge rates are considered to be independent and irrelevant to battery aging, but in fact, the temperatures, the discharge rates, even the charge rates, the discharge depths, the cycle intervals and the charge-discharge cutoff voltages have certain coupling relations to the battery aging, and are not independent and linearly characterizable, and the capacity attenuation coefficients at different stages are not equivalent capacity coefficients, so that the curve of the battery aging cannot be well described. In the invention, the coupling relation of multiple factors is considered, and the accuracy cannot be ensured by a common method with fewer test samples, so that the method of the multi-output relevance vector regression machine which is applicable to small samples and has multiple outputs is adopted, and the practicability and the good generalization of the battery life prediction are improved.
Disclosure of Invention
In order to solve the problems, the invention provides a method for predicting the service life of a battery under multi-stress based on Empirical Mode Decomposition (EMD) and a multi-output relevance vector Regression (MRVR). firstly, EMD is utilized to decompose high-frequency content modal components and low-frequency residual components from battery accelerated aging data under different working conditions so as to realize high-frequency de-noising, and then the residual components are fitted to obtain an Empirical model and corresponding model parameters thereof. And then, taking different working conditions of the developed test as input, taking each parameter of the obtained experience model under the corresponding working condition as output, learning by adopting MRVR, finally training the MRVR model, and finally inputting the working condition of the service life to be solved into the trained MRVR model, thereby realizing the service life prediction under different stress conditions.
The purpose of the invention is realized by at least one of the following technical solutions.
An EMD-MRVR-based multi-stress battery life prediction method comprises the following steps:
s1, inputting time sequence data which can reflect battery aging and is obtained by performing tests under different stress conditions and unified failure threshold values which are applicable under all stress conditions;
s2, preprocessing the input time sequence data to obtain preprocessed time sequence data;
s3, decomposing the preprocessed time sequence data into a high-frequency content modal component and a low-frequency residual component by empirical mode decomposition, and fitting the residual components to obtain a battery aging empirical model;
s4, taking the different stress conditions mentioned in the step S1 as input, taking the model parameters of the corresponding battery aging empirical model as output, and training a multi-output relevance vector regression (MRVR) model;
and S5, inputting a new stress condition different from the step S1 into the trained multi-output relevance vector regression (MRVR) model, and obtaining model parameters of the battery aging empirical model under the stress condition, so that the battery aging empirical model under the stress condition is obtained, the battery service life prediction under the stress condition is further realized, and a battery manufacturer and a maintenance worker are assisted to carry out production and maintenance work.
Further, in step S1, the stress conditions are combinations of different discharge depths, temperatures, charge rates, discharge rates, cycle intervals, and charge/discharge cutoff voltages.
Further, in step S1, the time series data includes battery capacity or impedance;
the unified failure threshold comprises a battery capacity failure threshold or a battery impedance failure threshold, when the battery capacity is smaller than the battery capacity failure threshold or the battery impedance is larger than the battery impedance failure threshold, the battery is determined to be failed, and for the battery capacity, the battery capacity failure threshold is 70% -80% of the initial capacity of the battery; for impedance, the battery impedance failure threshold is 2 times the initial impedance of the battery.
Further, in step S2, the missing value of the time series data and the abnormal value of the time series data are filled in by linear interpolation for the input time series data, and the last valid data y before the missing value or the abnormal value is taken0And corresponding time t0Taking the first valid data y after the missing value or the abnormal value1And corresponding time t1And obtaining a filling value of the missing value or a replacement value of the abnormal value according to the time t of the missing value or the abnormal value, which is specifically as follows:
Figure BDA0003486489240000031
and intercepting the time sequence data before the battery fails according to the unified failure threshold value for the time sequence data x after filling or replacement, and carrying out data normalization to obtain the time sequence data x after final data preprocessing*The method comprises the following steps:
Figure BDA0003486489240000032
wherein max is the maximum value of the time sequence data before the battery failure, and min is the minimum value of the time sequence data before the battery failure.
Further, in step S3, decomposing the preprocessed time series data into a high-frequency content modal component and a low-frequency residual component by using Empirical Mode Decomposition (EMD), so as to remove high-frequency noise in the preprocessed time series data, make the health indicator sequence smoother, and facilitate data fitting and establishment of a battery aging empirical model;
fitting the residual error component to obtain a battery aging empirical model;
the fitting to the residual component is an empirical fit, including a fourth, fifth or sixth order polynomial or exponential form.
Further, in step S4, a multiple output correlation vector regression (MRVR) machine is trained by learning the different stress conditions and corresponding model parameters of the battery aging empirical model in step S1 using the MRVR machine.
The nature of MRVR is a correlation Vector Machine, which is a sparse probability model similar to a Support Vector Machine (SVM) proposed by Micnacl e.tilting in 2000, and is a new supervised learning method. The method converts the nonlinear problem of the low-dimensional space into the linear problem of the high-order space based on the kernel function mapping like a support vector machine, but greatly reduces the calculation amount of the kernel function, has shorter test time, supports small sample data, and has better generalization capability. Because the model parameters of the battery aging empirical model have a coupling relation, when a plurality of single-output correlation vector machines are adopted, model parameter predictions are mutually separated, and the prediction accuracy is reduced, so that the multiple-output correlation vector machines are adopted for carrying out regression analysis.
Further, in step S5, inputting a stress condition different from step S1 by using a trained multiple output correlation vector regression (MRVR) model, predicting model parameters of the battery aging empirical model under the stress condition by using the trained multiple output correlation vector regression (MRVR) model, so as to obtain a battery aging empirical model under the stress condition, that is, a complete aging curve, and obtaining a battery life value under the stress condition by taking a time when the aging curve reaches a uniform failure threshold from the beginning, thereby realizing battery life prediction under each stress condition; the regression model is suitable for small-batch training samples, so that the life prediction is generalized to more stress conditions without tests on the premise of less stress condition combination tests, the battery life prediction is carried out, operation and maintenance personnel are assisted to overhaul and replace, meanwhile, the test cost is saved, and the complete aging curve or the complete aging process under the stress conditions without tests is made up.
Compared with the prior art, the invention has the following advantages and effects:
(1) the current operation state of the battery can be well judged according to the prediction curve and by combining with the state of health evaluation or reliability.
(2) The method is suitable for small samples and has good generalization capability.
(3) The method of empirical mode decomposition and a related vector machine is applied to the service life prediction of the battery under the multi-stress condition for the first time, and a new research idea is provided for the service life prediction of the batteries with different types and different working conditions.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below.
FIG. 1 is a flow chart of an embodiment of the present invention;
FIG. 2 is a graph of empirical mode decomposition results of capacity time series data in example 1 of the present invention;
FIG. 3 is a graph showing the effect of comparing the residual error component with the original data point under a certain stress condition in example 1 of the present invention;
fig. 4 is a graph showing the fitting effect of the fourth-order polynomial battery aging empirical model under a certain stress condition in example 1 of the present invention.
FIG. 5 is a graph of empirical mode decomposition results of time series data of maximum temperature time in example 2 of the present invention;
FIG. 6 is a graph showing the effect of comparing residual components with original data points under a certain stress condition in example 2 of the present invention;
fig. 7 is a graph showing the fitting effect of an eighth order polynomial battery aging empirical model under a certain stress condition in example 2 of the present invention.
Detailed Description
The core of the invention is to provide a method for predicting the service life of the battery under the multi-stress based on EMD-MRVR, so that the service life of the battery under the multi-stress condition is more completely predicted.
In order to make the technical solutions of the present invention better understood, the present invention will be described in further detail with reference to the accompanying drawings and the detailed description below. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1:
an EMD-MRVR-based method for predicting battery life under multiple stresses, as shown in fig. 1, includes the following steps:
s1, inputting time sequence data which can reflect battery aging and is obtained by performing tests under different stress conditions and unified failure threshold values which are applicable under all stress conditions;
the stress condition is the combination of different discharge depths, temperatures, charge multiplying factors, discharge multiplying factors, circulation intervals and charge-discharge cut-off voltages.
In this embodiment, the time series data adopts battery capacity;
the unified failure threshold comprises a battery capacity failure threshold or a battery impedance failure threshold, and when the battery capacity is smaller than the battery capacity failure threshold or the battery impedance is larger than the battery impedance failure threshold, the battery is judged to be failed; in this embodiment, for the battery capacity, the battery capacity failure threshold is 75% of the initial capacity of the battery;
s2, preprocessing the input time sequence data to obtain preprocessed time sequence data;
for the input time sequence data, the linear interpolation method is used for filling the missing value of the time sequence data and replacing the abnormal value of the time sequence data, and the last valid data y before the missing value or the abnormal value is taken0And corresponding time t0Taking the first valid data y after the missing value or the abnormal value1And corresponding time t1And obtaining a filling value of the missing value or a replacement value of the abnormal value according to the time t of the missing value or the abnormal value, which is specifically as follows:
Figure BDA0003486489240000061
and intercepting the time sequence data before the battery fails according to the unified failure threshold value for the time sequence data x after filling or replacement, and carrying out data normalization to obtain the time sequence data x after final data preprocessing*The method comprises the following steps:
Figure BDA0003486489240000062
wherein max is the maximum value of the time sequence data before the battery failure, and min is the minimum value of the time sequence data before the battery failure.
S3, decomposing the preprocessed time sequence data into a high-frequency connotation modal component and a low-frequency residual component by empirical mode decomposition, and fitting the residual component to obtain a battery aging empirical model;
by adopting Empirical Mode Decomposition (EMD), as shown in fig. 2, the preprocessed time series data is decomposed into three high-frequency content Mode components (IMFs) and one low-frequency residual component, so that high-frequency noise in the preprocessed time series data is removed, a health index sequence is smoother, and data fitting and establishment of a battery aging empirical model are facilitated; the comparison effect of the residual component and the original data point obtained under a certain stress condition is shown in fig. 3, and the comparison effect proves the above statement.
The fitting of the residual component is an empirical fit, in this embodiment, a quadratic polynomial fitting model is used as the battery aging empirical model, and the model expression is as follows:
f(x)＝p1x4+p2x3+p3x2+p4x+p5；
wherein p is1、p2、p3、p4、p5Model parameters of a fourth-order polynomial battery aging empirical model. The effect of a quartic polynomial fit under certain stress conditions is shown in figure 4.
S4, taking the different stress conditions mentioned in the step S1 as input, taking the model parameters of the corresponding battery aging empirical model as output, and training a multi-output relevance vector regression (MRVR) model;
in this embodiment, according to the document Fast multi-output-based aging vector regression, or Template-based aging and tracking of 3D hand motion, or Template-based aging and tracking of multi-output-based aging regression, the multi-output correlation vector regression is used to learn the different stress conditions and the corresponding model parameters of the battery aging experience model in step S1, and train the multi-output correlation vector regression (MRVR) model
S5, inputting a new stress condition different from the step S1 into the trained multi-output relevance vector regression (MRVR) model, and obtaining model parameters of the battery aging empirical model under the stress condition, so that the battery aging empirical model under the stress condition is obtained, the battery service life prediction under the stress condition is further realized, and a battery manufacturer and a maintainer are assisted to carry out production and maintenance work;
inputting a stress condition different from the step S1 by using a trained multi-output correlation vector regression (MRVR) model, predicting model parameters of a battery aging empirical model under the stress condition by using the trained multi-output correlation vector regression (MRVR) model so as to obtain the battery aging empirical model under the stress condition, namely a complete aging curve, and obtaining a battery life value under the stress condition by taking the time (cycle number) of the aging curve from the beginning to the unified failure threshold value so as to realize the prediction of the battery life under each stress condition; the regression model is suitable for small-batch training samples, so that the life prediction is generalized to more stress conditions without tests on the premise of less stress condition combination tests, the battery life prediction is carried out, operation and maintenance personnel are assisted to overhaul and replace, meanwhile, the test cost is saved, and the complete aging curve or the complete aging process under the stress conditions without tests is made up. As shown in fig. 5, the solid line in the graph is the aging curve of the battery under the stress condition in which the test has been performed, and the dotted line is the aging curve of the battery under the stress condition in which the test has not been performed to be predicted.
Example 2:
in this embodiment, the same as in embodiment 1 is input for various conditions, except that the characteristics closely related to the state of health of the battery obtained in each cycle of the battery that has been operated are taken as the ordinate and the cycle number is taken as the abscissa, as shown in fig. 5, the curve is smoothed by Empirical Mode Decomposition (EMD), fig. 6 is the comparison effect before and after smoothing, and then the obtained residual is curve-fitted by polynomial of degree 8 to obtain the fitting coefficient of the curve, as shown in fig. 7. By establishing a mapping relation between the input working condition and the fitting coefficient of the curve, namely establishing a multi-output correlation vector machine, the aging curve coefficient under different working conditions is estimated, and the cycle life under unknown working conditions is predicted.
The characteristics closely related to the state of health of the battery include a constant current discharge time or a maximum temperature time.
Example 3:
in this embodiment, the time series data is battery impedance, and for the battery impedance, the battery impedance failure threshold is 2 times of the initial impedance of the battery.
The above embodiments are only specific embodiments of the present invention, and are not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.

Claims (10)

1. An EMD-MRVR-based multi-stress battery life prediction method is characterized by comprising the following steps of:
s1, inputting time sequence data which can reflect battery aging and is obtained by carrying out tests under different stress conditions and a unified failure threshold value which is applicable under all stress conditions;
s2, preprocessing the input time sequence data to obtain preprocessed time sequence data;
s3, decomposing the preprocessed time sequence data into a high-frequency connotation modal component and a low-frequency residual component by empirical mode decomposition, and fitting the residual component to obtain a battery aging empirical model;
s4, taking the different stress conditions mentioned in the step S1 as input, taking the model parameters of the corresponding battery aging empirical model as output, and training a multi-output relevance vector regression (MRVR) model;
and S5, inputting a new stress condition different from the step S1 into the trained multi-output correlation vector regression model, and obtaining model parameters of the battery aging empirical model under the stress condition, so that the battery aging empirical model under the stress condition is obtained, and the battery service life prediction under the stress condition is further realized.
2. The method of claim 1, wherein the stress conditions are different combinations of depth of discharge, temperature, charge rate, discharge rate, cycle interval and charge-discharge cut-off voltage in step S1.
3. The method for predicting battery life under EMD-MRVR as claimed in claim 1, wherein in step S1, the time series data includes battery capacity or impedance;
the unified failure threshold value comprises a battery capacity failure threshold value or a battery impedance failure threshold value, and when the battery capacity is smaller than the battery capacity failure threshold value or the battery impedance is larger than the battery impedance failure threshold value, the battery is judged to be failed.
4. The EMD-MRVR based multi-stress battery life prediction method of claim 3 wherein for a battery capacity, the battery capacity failure threshold is 70% -80% of the initial capacity of the battery; for impedance, the battery impedance failure threshold is 2 times the initial impedance of the battery.
5. The method of claim 1, wherein in step S2, the linear interpolation is used to fill the missing values of the time series data and replace the abnormal values of the time series data, and the last valid data y before the missing values or the abnormal values is taken as the time series data0And corresponding time t0Taking the first valid data y after the missing value or the abnormal value1And corresponding time t1Obtaining a filling value of the missing value or a replacement value y (t) of the abnormal value according to the time t of the missing value or the abnormal value, which is as follows:
Figure FDA0003486489230000021
6. the EMD-MRVR-based multi-stress battery life prediction method of claim 5, wherein the filled or replaced time series data x is subjected to data normalization by intercepting the time series data before battery failure according to a uniform failure threshold to obtain the time series data x after final data preprocessing*The method comprises the following steps:
Figure FDA0003486489230000022
wherein max is the maximum value of the time sequence data before the battery failure, and min is the minimum value of the time sequence data before the battery failure.
7. The method for predicting battery life under multiple stresses according to claim 1, wherein in step S3, Empirical Mode Decomposition (EMD) is used to decompose the preprocessed time series data into high-frequency content mode components and low-frequency residual components, so as to remove high-frequency noise in the preprocessed time series data;
and fitting the residual error component to obtain a battery aging empirical model.
8. The EMD-MRVR-based multi-stress battery life prediction method of claim 7 wherein the fitting to the residual components is an empirical fit comprising a fourth, fifth or sixth order polynomial.
9. The EMD-MRVR-based multi-stress battery life prediction method of claim 1, wherein in step S4, a multi-output correlation vector regression (MRVR) model is trained by learning different stress conditions and corresponding model parameters of the battery aging empirical model in step S1 using a multi-output correlation vector regression.
10. The EMD-MRVR-based multi-stress battery life prediction method of any one of claims 1-9, wherein in step S5, a trained multi-output correlation vector regression (MRVR) model is used to input a stress condition different from that in step S1, and the trained multi-output correlation vector regression (MRVR) model predicts model parameters of the battery aging empirical model under the stress condition, so as to obtain a battery aging empirical model under the stress condition, i.e. a complete aging curve, and a time from the beginning of the aging curve reaching a unified failure threshold is taken to obtain a battery life value under the stress condition, so as to predict the battery life under each stress condition.
CN202210082592.0A 2022-01-24 2022-01-24 EMD-MRVR-based multi-stress battery life prediction method Pending CN114547969A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210082592.0A CN114547969A (en) 2022-01-24 2022-01-24 EMD-MRVR-based multi-stress battery life prediction method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210082592.0A CN114547969A (en) 2022-01-24 2022-01-24 EMD-MRVR-based multi-stress battery life prediction method

Publications (1)

Publication Number Publication Date
CN114547969A true CN114547969A (en) 2022-05-27

Family

ID=81670926

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210082592.0A Pending CN114547969A (en) 2022-01-24 2022-01-24 EMD-MRVR-based multi-stress battery life prediction method

Country Status (1)

Country Link
CN (1) CN114547969A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116840700A (en) * 2023-08-31 2023-10-03 深圳市安德普电源科技有限公司 Method, device, equipment and storage medium for monitoring battery state in real time
CN117454186A (en) * 2023-12-22 2024-01-26 宁德时代新能源科技股份有限公司 Model training method, battery performance prediction method, device, equipment and storage medium

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116840700A (en) * 2023-08-31 2023-10-03 深圳市安德普电源科技有限公司 Method, device, equipment and storage medium for monitoring battery state in real time
CN116840700B (en) * 2023-08-31 2023-10-31 深圳市安德普电源科技有限公司 Method, device, equipment and storage medium for monitoring battery state in real time
CN117454186A (en) * 2023-12-22 2024-01-26 宁德时代新能源科技股份有限公司 Model training method, battery performance prediction method, device, equipment and storage medium
CN117454186B (en) * 2023-12-22 2024-05-14 宁德时代新能源科技股份有限公司 Model training method, battery performance prediction method, device, equipment and storage medium

Similar Documents

Publication Publication Date Title
Li et al. Big data driven lithium-ion battery modeling method based on SDAE-ELM algorithm and data pre-processing technology
CN107957562B (en) Online prediction method for residual life of lithium ion battery
CN107918103B (en) Lithium ion battery residual life prediction method based on gray particle filtering
CN114547969A (en) EMD-MRVR-based multi-stress battery life prediction method
CN109001640B (en) Data processing method and device for power battery
CN107870306A (en) A kind of lithium battery charge state prediction algorithm based under deep neural network
EP3966579A1 (en) Server-side characterisation of rechargeable batteries
CN114372417A (en) Electric vehicle battery health state and remaining life evaluation method based on charging network
CN112068003B (en) Method and device for predicting service life of cadmium-nickel storage battery based on linear wiener process
JP2023501995A (en) Battery performance prediction
EP4031886A1 (en) Condition value for rechargeable batteries
CN115114878A (en) Method and device for online prediction of battery life of energy storage power station and storage medium
CN110766236A (en) Power equipment state trend prediction method based on statistical analysis and deep learning
CN115587489A (en) Method and device for determining hydrogen production model, nonvolatile storage medium and processor
Sydor et al. Recurent Expression for Reliability Indicators of Compound Electropower Systems
CN102095953B (en) A kind of performance of accumulator charger online test method
CN116738865B (en) Energy storage power supply evaluation method and system based on Internet of things
CN116502544B (en) Electric screw press life prediction method and system based on data fusion
CN117445755A (en) Cloud computing-based remote monitoring system for batteries of electric vehicle
CN117520986A (en) Distributed photovoltaic power generation anomaly monitoring method, system, equipment and storage medium
CN116901707A (en) Power battery pack fault early warning method, system and vehicle
CN116794545A (en) Yun Bian cooperation-based battery health prediction method and system
CN116609686A (en) Battery cell consistency assessment method based on cloud platform big data
CN115963401A (en) Battery cyclicity prediction method based on microcontroller equipment and embedded dense neural network
CN111983458A (en) Parallel battery optimal model judging and optimizing method

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