CN111985156A - Method for predicting health state of battery - Google Patents

Method for predicting health state of battery Download PDF

Info

Publication number
CN111985156A
CN111985156A CN202010834612.6A CN202010834612A CN111985156A CN 111985156 A CN111985156 A CN 111985156A CN 202010834612 A CN202010834612 A CN 202010834612A CN 111985156 A CN111985156 A CN 111985156A
Authority
CN
China
Prior art keywords
battery
constant
soh
parameters
discharging
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.)
Granted
Application number
CN202010834612.6A
Other languages
Chinese (zh)
Other versions
CN111985156B (en
Inventor
盛瀚民
刘鑫
程玉华
邵晋梁
陈凯
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.)
University of Electronic Science and Technology of China
Original Assignee
University of Electronic Science and Technology of China
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 University of Electronic Science and Technology of China filed Critical University of Electronic Science and Technology of China
Priority to CN202010834612.6A priority Critical patent/CN111985156B/en
Publication of CN111985156A publication Critical patent/CN111985156A/en
Application granted granted Critical
Publication of CN111985156B publication Critical patent/CN111985156B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • 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]
    • G01R31/396—Acquisition or processing of data for testing or for monitoring individual cells or groups of cells within a battery
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/08—Probabilistic or stochastic CAD
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N7/00—Computing arrangements based on specific mathematical models
    • G06N7/01—Probabilistic graphical models, e.g. probabilistic networks

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Theoretical Computer Science (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computer Hardware Design (AREA)
  • Geometry (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Secondary Cells (AREA)
  • Charge And Discharge Circuits For Batteries Or The Like (AREA)

Abstract

The invention discloses a method for predicting the health state of a battery, which comprises the steps of extracting five characteristics of historical constant-current charging time, constant-voltage charging time, maximum value of a capacity increment curve, equal-pressure difference time interval of discharging, ohmic internal resistance and the like of the battery in each complete charging and discharging process, then processing the extracted characteristics, and establishing a transfer Gaussian process regression model, thereby predicting the SOH value of a new battery in the subsequent use process.

Description

Method for predicting health state of battery
Technical Field
The invention belongs to the technical field of battery health state estimation, and particularly relates to a method for predicting the battery health state.
Background
The state of health of the battery is closely related to the driving range, safety and reliability of the electric vehicle. Since the degradation mechanism of the battery is complex and the influence factors are numerous, accurate and reliable estimation of the state of health SOH of the battery is a difficult problem in the battery management technology.
The data driving method represented by machine learning is flexible, does not need modeling, has good nonlinear mapping capability, and is a research hotspot in the field at present. Researchers have proposed a variety of data-driven SOH estimation methods, however, studies are currently mainly focused on the modeling process of the specific power battery state of health under experimental conditions, and how to predict the performance of a new battery without historical SOH data is still an unsolved problem. Due to different battery types and different use environments, it is difficult to ensure that the training data and the predicted object have the same data distribution, and when the training data and the actually predicted data distribution of the battery are different, the reliability of a general data-driven model is difficult to ensure. Therefore, the practical problem in the research of the battery SOH estimation when the data knowledge obtained by the laboratory is applied to the tested object is solved.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a method for predicting the state of health of a battery, which extracts and processes characteristics from different existing batteries so as to predict the SOH of the battery.
To achieve the above object, the present invention provides a method for predicting a state of health of a battery, comprising the steps of:
(1) extracting the feature
(1.1) extracting five characteristics of constant-current charging time, constant-voltage charging time, maximum value of a capacity increment curve, equal-voltage difference time interval of discharging and ohmic internal resistance of the conventional battery S in each complete charging and discharging process and SOH (state of health) to form a set
Figure BDA0002639215180000011
Wherein x and y represent the eigenvector and SOH, respectively, subscript is the cycle number, and m represents the total cycle number of the battery S in the entire life cycle;
(1.2) extracting five characteristics of constant-current charging time, constant-voltage charging time, maximum value of capacity increment curve, constant-voltage difference time interval of discharging and ohmic internal resistance of the new battery T in the previous n times of complete charging and discharging processes and SOH (state of health) to form a set
Figure BDA0002639215180000021
(2) Data preprocessing
Using mapminmax function in matlab to respectively pair DSAnd DTNormalizing the characteristic x;
(3) determining the parameter to be optimized
(3.1) setting optimization parameters
Figure BDA0002639215180000022
The initialization parameters Θ, rand (5,1), where l, σ denote the parameters to be optimized in the gaussian kernel function,
Figure BDA0002639215180000023
represents the measurement error of the cell S and the cell T, and λ represents the transfer rate;
(3.2) setting DSSatisfying a prior distribution
Figure BDA0002639215180000024
DTSatisfying a prior distribution
Figure BDA0002639215180000025
Wherein the parameters
Figure BDA0002639215180000026
Im,InRespectively representing m × m dimension and n × n dimension unit matrixes;
(3.3) calculating the joint distribution probability p (Y)T,YS|XS,XT)；
Figure BDA0002639215180000027
Wherein K () is a Gaussian kernel function satisfying
Figure BDA0002639215180000028
Superscript T denotes transpose, KSSAnd KTTRespectively represent DSAnd DTRespective kernel matrix, KSTA kernel matrix representing a cross-domain;
(3.4) probability p (Y) according to joint distributionT,YS|XS,XT) Calculating conditional distribution probability p (Y) by using Bayes formulaT|YS,XS,XT)；
Figure BDA0002639215180000029
(3.5) optimization of parameters by pairs
Figure BDA00026392151800000210
Taking appropriate values to make the conditional distribution probability p (Y)T|YS,XS,XT) Max, thus constructing a maximized optimization objective function:
Figure BDA0002639215180000031
wherein the content of the first and second substances,
Figure BDA0002639215180000032
optimizing a target function by utilizing a covariance matrix adaptive evolution strategy so as to determine the value of an optimization parameter theta;
(4) predicting SOH of battery T
(4.1) extracting five characteristics of constant-current charging time, constant-voltage charging time, maximum value of capacity increment curve, constant-voltage difference time interval of discharging and ohmic internal resistance in the complete charging and discharging process of the (n + 1) th time of the battery to form a set
Figure BDA0002639215180000033
x*Representing the feature vector, y, extracted at the (n + 1) th time*Represents the SOH to be predicted;
feature vector x is mapped to by mapminmax function in matlab*Carrying out normalization processing;
(4.2) optimizing the parameters
Figure BDA0002639215180000034
Substituting into the joint distribution probability calculation formula in step (3.3), recalculating new joint distribution probability, denoted as p (Y | X),
Figure BDA0002639215180000035
(4.3) calculating Y and Y*Is given by the joint distribution probability p (y)*,Y∣x*,X)；
Figure BDA0002639215180000036
Wherein the content of the first and second substances,
Figure BDA0002639215180000037
K*represents DSAnd DTThe feature vector X in (1)
Figure BDA0002639215180000038
In x*Kernel matrix of, K**To represent
Figure BDA0002639215180000039
In x*From the nucleusA matrix;
(4.4) calculating the conditional distribution probability p (y) by using a Bayesian formula*∣x*,X,Y)；
Figure BDA00026392151800000310
(4.5) probability p (y) according to conditional distribution*∣x*X, Y), let Y*＝K*K-1And Y, and is taken as SOH in the n +1 th complete charging and discharging process of the new battery T.
The invention aims to realize the following steps:
the invention relates to a method for predicting the state of health of a battery, which comprises the steps of extracting five characteristics of historical constant-current charging time, constant-voltage charging time, maximum value of a capacity increment curve, equal-pressure difference time interval of discharging, ohmic internal resistance and the like of the battery in each complete charging and discharging process, then processing the extracted characteristics, and establishing a transition Gaussian process regression model, thereby predicting the SOH value of a new battery in the subsequent use process.
Drawings
FIG. 1 is a flow chart of a method of predicting battery state of health in accordance with the present invention;
FIG. 2 is a SOH curve during the entire cycle of charge and discharge with NASA (national aviation and aerospace administration) data collector cell numbers B0005, B0006, B00018, respectively;
FIG. 3 is a graph of the comparative effect of the present invention on a NASA dataset with the LSSVM (least squares support vector machine) algorithm and the LSTM (long short term memory network) algorithm;
FIG. 4 is a SOH curve during the entire cycle of charge and discharge for CLACE (university of Maryland) data set batteries with CS2_33, CS2_35, CS2_36, CS2_38, and CX2_36, respectively.
FIG. 5 is a graph showing the comparative effect of the present invention on CLACE data set with LSSVM (least squares support vector machine) algorithm and LSTM (long short term memory network) algorithm in CLACE data set.
Detailed Description
The following description of the embodiments of the present invention is provided in order to better understand the present invention for those skilled in the art with reference to the accompanying drawings. It is to be expressly noted that in the following description, a detailed description of known functions and designs will be omitted when it may obscure the subject matter of the present invention.
Examples
For convenience of description, the related terms appearing in the detailed description are first described
Ann (artificial Neural network): an artificial neural network;
LSSVM (least Square Support Vector machine): a least squares support vector machine;
GPR (Gaussian Process regression) Gaussian Process regression;
LSTM (Long Short Term memory): a long-short term memory network;
rmse (root Mean Square error) root Mean Square error;
mae (mean Square error): averaging the absolute errors;
SSE (the Sum of Squares error);
MAPE (mean Absolute percent error);
AIW: (the average width of the 95% confidence error interval) is the average width of the 95% error bound.
FIG. 1 is a flow chart of a method for predicting battery state of health in accordance with the present invention.
In this embodiment, as shown in fig. 1, a method for predicting the state of health of a battery according to the present invention includes the following steps:
s1, extracting characteristics
S1.1, recording the number of times of cyclic charge and discharge of an existing battery and the voltage, current and other data in each charge and discharge process in the use process of the existing battery, and taking the public data set of NASA as an example, firstly extracting the constant-current charging time of the battery S with the serial number B0005 in each complete charge and discharge processFive characteristics of constant voltage charging time, maximum value of capacity increment curve, constant voltage difference time interval of discharge and ohmic internal resistance, and health state SOH
Figure BDA0002639215180000051
Wherein x and y represent the eigenvector and SOH, respectively, subscript is the cycle number, and m represents the total cycle number of the battery S in the entire life cycle;
in the embodiment, the extracted features are some features which are frequently used in the field of SOH prediction at present, and the method is not limited to the features and has good performance under the condition of adding or reducing the features.
S1.2, extracting five characteristics of constant-current charging time, constant-voltage charging time, maximum value of capacity increment curve, equal-voltage difference time interval of discharging and ohmic internal resistance of the battery T with the serial number of B0018 in the previous n times of complete charging and discharging processes and SOH of the health state to form a set
Figure BDA0002639215180000052
n＜＜m。
In this embodiment, in order to compensate for the insufficient data information, we migrate the relevant information in other battery data to the target object, where DTRepresenting data generated by predicting the target object, DSRepresenting an auxiliary data set, called the source domain in migration learning, for providing additional information, we want as little data as possible in this data set, so we want n < m in the experimental setup. In the present embodiment, we can predict the SOH of the eighty percent battery behind the battery T by using only the twenty percent data of the battery T, and through the pre-training of the twenty percent data, the estimation of the subsequent SOH can be obtained by real-time prediction, which is very good for practical application. In the algorithm, firstly, a model and parameters are pre-trained by using source domain data with label values and a small part of target domain data, and then the trained parameters are used for predicting label-free data of a target domain by using Gaussian process regression. Thus, the present invention is also a novel mobilityThe application of the conventional algorithm in the field of battery SOH estimation, and the fact that the transfer learning algorithm is used for SOH estimation means that laboratory data can be used for predicting the SOH of a battery in a real life scene, historical data of an old battery can also be used for predicting the SOH of a new battery, and even when the old battery and the new battery belong to different types of batteries, the transfer learning algorithm has a good effect.
S2, preprocessing data
Using mapminmax function in matlab to respectively pair DSAnd DTNormalizing the characteristic x to 0-1 interval;
s3, determining parameters to be optimized
S3.1, setting optimization parameters
Figure BDA0002639215180000061
The initialization parameters Θ, rand (5,1), where l, σ denote the parameters to be optimized in the gaussian kernel function,
Figure BDA0002639215180000062
represents the measurement error of the cell S and the cell T, and λ represents the transfer rate;
s3.2, setting DSSatisfying a prior distribution
Figure BDA0002639215180000063
DTSatisfying a prior distribution
Figure BDA0002639215180000064
Wherein the parameters
Figure BDA0002639215180000065
Im,InRespectively representing m × m dimension and n × n dimension unit matrixes;
in the present embodiment, for computational complexity and higher accuracy considerations, only D may be selected hereSThe first twenty percent data of (a) or corresponding to DTData of the middle SOH phase. The original symbols are used here for easier understanding of the description of the invention.
S3.3, calculating joint distribution probability p (Y)T,YS|XS,XT)；
Figure BDA0002639215180000066
Wherein K () is a Gaussian kernel function satisfying
Figure BDA0002639215180000071
Superscript T denotes transpose, KSSAnd KTTRespectively represent DSAnd DTRespective kernel matrix, KSTA kernel matrix representing a cross-domain;
s3.4, according to the joint distribution probability p (Y)T,YS|XS,XT) Calculating conditional distribution probability p (Y) by using Bayes formulaT|YS,XS,XT)；
Figure BDA0002639215180000072
S3.5, optimizing parameters by pairs
Figure BDA0002639215180000073
Taking appropriate values to make the conditional distribution probability p (Y)T|YS,XS,XT) Max, thus constructing a maximized optimization objective function:
Figure BDA0002639215180000074
wherein the content of the first and second substances,
Figure BDA0002639215180000075
optimizing a target function by utilizing a covariance matrix adaptive evolution strategy so as to determine the value of an optimization parameter theta;
s4, predicting SOH of new battery T
S4.1, extracting constant-current charging time and constant-voltage charging time in the n +1 th complete charging and discharging process of the new batteryThe five characteristics of the maximum value of the capacity increment curve, the equal differential time interval of the discharge and the ohmic internal resistance form a set
Figure BDA0002639215180000076
x*Representing the feature vector, y, extracted at the (n + 1) th time*Represents the SOH to be predicted;
feature vector x is mapped to by mapminmax function in matlab*Carrying out normalization processing;
s4.2, optimizing the parameters
Figure BDA0002639215180000077
Substituting into the joint distribution probability calculation formula in step (3.3), recalculating new joint distribution probability, denoted as p (Y | X),
Figure BDA0002639215180000078
s4.3, calculating Y and Y*Is given by the joint distribution probability p (y)*,Y∣x*,X)；
Figure BDA0002639215180000079
Wherein the content of the first and second substances,
Figure BDA0002639215180000081
K*represents DSAnd DTThe feature vector X in (1)
Figure BDA0002639215180000082
In x*Kernel matrix of, K**To represent
Figure BDA0002639215180000083
In x*A kernel matrix of self;
s4.4, calculating the conditional distribution probability p (y) by using a Bayesian formula*∣x*,X,Y)；
Figure BDA0002639215180000084
S4.5, distributing probability p (y) according to conditions*∣x*X, Y), let Y*＝K*K-1And Y, and is taken as SOH in the n +1 th complete charging and discharging process of the new battery T.
Example show
The invention compares the results with ANN, LSSVM, GPR and LSTM algorithms, wherein the performance indexes of the algorithms are calculated according to the following formulas
Figure BDA0002639215180000085
Figure BDA0002639215180000086
Figure BDA0002639215180000087
Figure BDA0002639215180000088
Figure BDA0002639215180000089
Where m denotes the number of experimental repetitions, n denotes the number of samples to be predicted, UjAnd LjThe upper and lower bounds of the 95% confidence interval at the jth sample, respectively, are indicated.
FIG. 2 is a SOH curve during the entire cyclic charge and discharge process for NASA (national aviation and aerospace administration) data collector cell numbers B0005, B0006 and B0018, respectively;
as shown in fig. 2, in combination with table 1, it can be seen that the cycle charge and discharge times (i.e., the service life) and the SOH corresponding to each charge and discharge are different for the three types of batteries due to different cut-off voltages.
Figure BDA00026392151800000810
Figure BDA0002639215180000091
TABLE 1
FIG. 3 is a graph of the effectiveness of the present invention validated with NASA data. Three cell combination experiments shown in fig. 2 were used.
As shown in fig. 3(a), the effect of the present invention is much better than that of LSSVM and LSTM algorithms by predicting eighty percent of battery SOH after battery B0006 with battery B0005 and the first twenty percent of battery B0006, and in table 2-case1, for various performance parameters of the validation algorithm index in common, the RMSE, MAE, SSE, MAPE parameters of the present invention are the smallest in several comparative algorithms, and the AIW parameters are much better than the GPR algorithm. Similarly, in fig. 3(B) (c), tables 2-case2, case3, battery B0005 and B0006 are used to predict battery B0018, respectively, and the visualization effect and various index parameters are superior to those of the other categories.
Figure BDA0002639215180000092
TABLE 2
FIG. 4 is a SOH curve during the entire charge and discharge cycle for CLACE (university of Maryland) data set batteries with CS2_33, CS2_35, CS2_36, CS2_38, CX2_36, respectively;
as shown in fig. 4, in conjunction with table 3, we can see that the SOH curves of the 4 batteries of the CS2 type are different due to the difference of the charging current. Whereas the CX2 battery is a different type of battery from the CS2 battery, it is apparent that its SOH curve is quite different from that of the CS2 battery.
Figure BDA0002639215180000093
Figure BDA0002639215180000101
TABLE 3
Figure 5 is a graph of the effect of validating the invention with CLACE data. Three cell combination experiments shown in fig. 4 were used.
As shown in FIG. 5(a), the CS2_33 battery is used for assisting in predicting the CS2_38 battery, the effect is obviously better than that of the LSSVM and the LSTM, and the 95% confidence interval is controlled in a very small range, which is a very strong proof of reliability. Also shown in table 4-case1 are various performance index parameters for more comparative algorithms and our algorithms, and this set of experiments demonstrates the good effect of our method at different charging currents of the battery.
As shown in fig. 5(b), CS2_36 battery is used to assist in predicting CS2_35 battery, the type and charging conditions of both batteries are the same, and we have designed this experiment to demonstrate that our method can achieve good results without migration. The performance indicators of the various algorithms are shown in Table 4-case2, and our algorithm still maintains the best performance.
As shown in FIG. 5(c), the CX2_33 battery is used for assisting in predicting the CS2_38 battery, the charging current and the battery type of the two batteries are different, and the effect is obviously better than that of the LSSVM and the LSTM. Also shown in table 4-case3 are various types of performance index parameters for more comparative algorithms versus our algorithm, and this set of experiments demonstrates the effectiveness of our method at different charging currents and different types of batteries.
Figure BDA0002639215180000102
TABLE 4
Although illustrative embodiments of the present invention have been described above to facilitate the understanding of the present invention by those skilled in the art, it should be understood that the present invention is not limited to the scope of the embodiments, and various changes may be made apparent to those skilled in the art as long as they are within the spirit and scope of the present invention as defined and defined by the appended claims, and all matters of the invention which utilize the inventive concepts are protected.

Claims (1)

1. A method of predicting battery state of health, comprising the steps of:
(1) extracting the feature
(1.1) extracting five characteristics of constant-current charging time, constant-voltage charging time, maximum value of a capacity increment curve, equal-voltage difference time interval of discharging and ohmic internal resistance of the conventional battery S in each complete charging and discharging process and SOH (state of health) to form a set
Figure FDA0002639215170000011
Wherein x and y represent the eigenvector and SOH, respectively, subscript is the cycle number, and m represents the total cycle number of the battery S in the entire life cycle;
(1.2) extracting five characteristics of constant-current charging time, constant-voltage charging time, maximum value of capacity increment curve, constant-voltage difference time interval of discharging and ohmic internal resistance of the new battery T in the previous n times of complete charging and discharging processes and SOH (state of health) to form a set
Figure FDA0002639215170000012
(2) Data preprocessing
Using mapminmax function in matlab to respectively pair DSAnd DTNormalizing the characteristic x;
(3) determining the parameter to be optimized
(3.1) setting optimization parameters
Figure FDA0002639215170000013
The initialization parameters Θ, rand (5,1), where l, σ denote the parameters to be optimized in the gaussian kernel function,
Figure FDA0002639215170000014
represents the measurement error of battery S and battery T;
(3.2) setting DSSatisfying a prior distribution
Figure FDA0002639215170000015
DTSatisfying a prior distribution
Figure FDA0002639215170000016
Wherein the parameters
Figure FDA0002639215170000017
Im,InRespectively representing m × m dimension and n × n dimension unit matrixes;
(3.3) calculating the joint distribution probability p (Y)T,YS|XS,XT)；
Figure FDA0002639215170000018
Wherein K () is a Gaussian kernel function satisfying
Figure FDA0002639215170000019
Superscript T denotes transpose, KSSAnd KTTRespectively represent DSAnd DTRespective kernel matrix, KSTA kernel matrix representing a cross-domain;
(3.4) probability p (Y) according to joint distributionT,YS|XS,XT) Calculating conditional distribution probability p (Y) by using Bayes formulaT|YS,XS,XT)；
Figure FDA0002639215170000021
(3.5) optimization of parameters by pairs
Figure FDA0002639215170000022
Taking appropriate values to make the conditional distribution probability p (Y)T|YS,XS,XT) Max, thus constructing a maximized optimization objective function:
Figure FDA0002639215170000023
wherein the content of the first and second substances,
Figure FDA0002639215170000024
and then, optimizing the objective function by utilizing a covariance matrix adaptive evolution strategy so as to determine the value of the optimization parameter theta.
(4) Predicting SOH of battery T
(4.1) extracting five characteristics of constant-current charging time, constant-voltage charging time, maximum value of capacity increment curve, constant-voltage difference time interval of discharging and ohmic internal resistance in the complete charging and discharging process of the (n + 1) th time of the battery to form a set
Figure FDA0002639215170000025
x*Representing the feature vector, y, extracted at the (n + 1) th time*Represents the SOH to be predicted;
feature vector x is mapped to by mapminmax function in matlab*Carrying out normalization processing;
(4.2) optimizing the parameters
Figure FDA0002639215170000026
Substituting into the joint distribution probability calculation formula in step (3.3), recalculating new joint distribution probability, denoted as p (Y | X),
Figure FDA0002639215170000027
(4.3) calculating Y and Y*Is given by the joint distribution probability p (y)*,Y∣x*,X)；
Figure FDA0002639215170000028
Wherein the content of the first and second substances,
Figure FDA0002639215170000029
K*represents DSAnd DTThe feature vector X in (1)
Figure FDA00026392151700000210
In x*Kernel matrix of, K**To represent
Figure FDA00026392151700000211
In x*A kernel matrix of self;
(4.4) calculating the conditional distribution probability p (y) by using a Bayesian formula*∣x*,X,Y)；
Figure FDA0002639215170000031
(4.5) probability p (y) according to conditional distribution*∣x*X, Y), let Y*＝K*K-1And Y, and is taken as SOH in the n +1 th complete charging and discharging process of the new battery T.
CN202010834612.6A 2020-08-19 2020-08-19 Method for predicting health state of battery Active CN111985156B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010834612.6A CN111985156B (en) 2020-08-19 2020-08-19 Method for predicting health state of battery

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010834612.6A CN111985156B (en) 2020-08-19 2020-08-19 Method for predicting health state of battery

Publications (2)

Publication Number Publication Date
CN111985156A true CN111985156A (en) 2020-11-24
CN111985156B CN111985156B (en) 2022-06-14

Family

ID=73435576

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010834612.6A Active CN111985156B (en) 2020-08-19 2020-08-19 Method for predicting health state of battery

Country Status (1)

Country Link
CN (1) CN111985156B (en)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112417767A (en) * 2020-12-09 2021-02-26 东软睿驰汽车技术（沈阳）有限公司 Attenuation trend determination model construction method and attenuation trend determination method
CN112540298A (en) * 2020-12-26 2021-03-23 福州大学 Lithium ion battery SOC estimation method based on deep-migration learning
CN112834945A (en) * 2020-12-31 2021-05-25 东软睿驰汽车技术（沈阳）有限公司 Evaluation model establishing method, battery health state evaluation method and related product
CN113094985A (en) * 2021-03-31 2021-07-09 电子科技大学 Battery SOH prediction method based on cross manifold embedded transfer learning
CN113128672A (en) * 2021-04-16 2021-07-16 昆明理工大学 Lithium ion battery pack SOH estimation method based on transfer learning algorithm
CN113740736A (en) * 2021-08-31 2021-12-03 哈尔滨工业大学 Electric vehicle lithium battery SOH estimation method based on deep network self-adaptation
CN114580262A (en) * 2021-11-18 2022-06-03 吉林大学 Lithium ion battery health state estimation method
CN114791993A (en) * 2022-05-16 2022-07-26 江苏大学 SOH prediction method and system for power battery pack

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102798823A (en) * 2012-06-15 2012-11-28 哈尔滨工业大学 Gaussian process regression-based method for predicting state of health (SOH) of lithium batteries
EP2669696A1 (en) * 2012-02-02 2013-12-04 LG Chem, Ltd. Method and device for predicting state-of-health of battery, and battery management system using same
US20140350877A1 (en) * 2013-05-25 2014-11-27 North Carolina State University Battery parameters, state of charge (soc), and state of health (soh) co-estimation
US20180143257A1 (en) * 2016-11-21 2018-05-24 Battelle Energy Alliance, Llc Systems and methods for estimation and prediction of battery health and performance
CN109031153A (en) * 2018-10-16 2018-12-18 北京交通大学 A kind of health status On-line Estimation method of lithium ion battery
CN109633477A (en) * 2019-01-04 2019-04-16 深圳市计量质量检测研究院 The method for real-time monitoring of battery state-of-health based on EKF-GPR and daily fragment data
CN110398697A (en) * 2019-07-23 2019-11-01 北京工业大学 A kind of lithium ion health status estimation method based on charging process
CN110568361A (en) * 2019-09-12 2019-12-13 华中科技大学 Method for predicting health state of power battery
CN111308379A (en) * 2020-03-13 2020-06-19 北京理工大学 Battery health state estimation method based on local constant voltage charging data
CN111443293A (en) * 2020-03-30 2020-07-24 青岛大学 Lithium battery state of health (SOH) estimation method based on data driving

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2669696A1 (en) * 2012-02-02 2013-12-04 LG Chem, Ltd. Method and device for predicting state-of-health of battery, and battery management system using same
CN102798823A (en) * 2012-06-15 2012-11-28 哈尔滨工业大学 Gaussian process regression-based method for predicting state of health (SOH) of lithium batteries
US20140350877A1 (en) * 2013-05-25 2014-11-27 North Carolina State University Battery parameters, state of charge (soc), and state of health (soh) co-estimation
US20180143257A1 (en) * 2016-11-21 2018-05-24 Battelle Energy Alliance, Llc Systems and methods for estimation and prediction of battery health and performance
CN109031153A (en) * 2018-10-16 2018-12-18 北京交通大学 A kind of health status On-line Estimation method of lithium ion battery
CN109633477A (en) * 2019-01-04 2019-04-16 深圳市计量质量检测研究院 The method for real-time monitoring of battery state-of-health based on EKF-GPR and daily fragment data
CN110398697A (en) * 2019-07-23 2019-11-01 北京工业大学 A kind of lithium ion health status estimation method based on charging process
CN110568361A (en) * 2019-09-12 2019-12-13 华中科技大学 Method for predicting health state of power battery
CN111308379A (en) * 2020-03-13 2020-06-19 北京理工大学 Battery health state estimation method based on local constant voltage charging data
CN111443293A (en) * 2020-03-30 2020-07-24 青岛大学 Lithium battery state of health (SOH) estimation method based on data driving

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
刘皓等: "基于遗传算法和支持向量回归的锂电池健康状态预测", 《南京理工大学学报》 *
郑雪莹等: "基于能量加权高斯过程回归的锂离子电池健康状态预测", 《电子测量与仪器学报》 *
陈凯等: "基于LIBSVM的铅酸蓄电池荷电状态估计", 《电源技术》 *
陈则王等: "基于GA-ELM的锂离子电池RUL间接预测方法", 《计量学报》 *
陈雄姿等: "基于贝叶斯LS-SVR的锂电池剩余寿命概率性预测", 《航空学报》 *

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112417767A (en) * 2020-12-09 2021-02-26 东软睿驰汽车技术（沈阳）有限公司 Attenuation trend determination model construction method and attenuation trend determination method
CN112540298A (en) * 2020-12-26 2021-03-23 福州大学 Lithium ion battery SOC estimation method based on deep-migration learning
CN112834945A (en) * 2020-12-31 2021-05-25 东软睿驰汽车技术（沈阳）有限公司 Evaluation model establishing method, battery health state evaluation method and related product
CN113094985A (en) * 2021-03-31 2021-07-09 电子科技大学 Battery SOH prediction method based on cross manifold embedded transfer learning
CN113094985B (en) * 2021-03-31 2022-06-14 电子科技大学 Battery SOH prediction method based on cross manifold embedded transfer learning
CN113128672A (en) * 2021-04-16 2021-07-16 昆明理工大学 Lithium ion battery pack SOH estimation method based on transfer learning algorithm
CN113740736A (en) * 2021-08-31 2021-12-03 哈尔滨工业大学 Electric vehicle lithium battery SOH estimation method based on deep network self-adaptation
CN113740736B (en) * 2021-08-31 2024-04-02 哈尔滨工业大学 Electric vehicle lithium battery SOH estimation method based on deep network self-adaption
CN114580262A (en) * 2021-11-18 2022-06-03 吉林大学 Lithium ion battery health state estimation method
CN114791993A (en) * 2022-05-16 2022-07-26 江苏大学 SOH prediction method and system for power battery pack
CN114791993B (en) * 2022-05-16 2022-11-11 江苏大学 Power battery pack SOH prediction method and system

Also Published As

Publication number Publication date
CN111985156B (en) 2022-06-14

Similar Documents

Publication Publication Date Title
CN111985156B (en) Method for predicting health state of battery
Sui et al. A review of non-probabilistic machine learning-based state of health estimation techniques for Lithium-ion battery
Meng et al. Long short-term memory network with Bayesian optimization for health prognostics of lithium-ion batteries based on partial incremental capacity analysis
Xu et al. Life prediction of lithium-ion batteries based on stacked denoising autoencoders
Yang et al. State-of-charge estimation of lithium-ion batteries using LSTM and UKF
Xu et al. Improving the state of charge estimation of reused lithium-ion batteries by abating hysteresis using machine learning technique
Tian et al. Deep learning framework for lithium-ion battery state of charge estimation: Recent advances and future perspectives
Chen et al. An overview of data-driven battery health estimation technology for battery management system
Sheng et al. Small sample state of health estimation based on weighted Gaussian process regression
CN107367693A (en) SOC detection system for power battery of electric vehicle
Couture et al. Image-and health indicator-based transfer learning hybridization for battery RUL prediction
Takyi-Aninakwa et al. A hybrid probabilistic correction model for the state of charge estimation of lithium-ion batteries considering dynamic currents and temperatures
Zhang et al. State of health estimation for lithium-ion batteries on few-shot learning
Che et al. Boosting battery state of health estimation based on self-supervised learning
Xue et al. Estimating state of health of lithium-ion batteries based on generalized regression neural network and quantum genetic algorithm
CN116186633A (en) Power consumption abnormality diagnosis method and system based on small sample learning
Shen et al. Temperature adaptive transfer network for cross-domain state-of-charge estimation of li-ion batteries
Li et al. A hybrid framework for predicting the remaining useful life of battery using Gaussian process regression
CN116401770A (en) Quick charge strategy design method based on battery digital twin model and machine learning
Lin et al. Algorithm of BPNN‐UKF based on a fusion model for SOC estimation in lithium‐ion batteries
Dawei et al. State of charge estimation for the Vanadium Redox Flow Battery based on Extended Kalman filter using modified parameter identification
CN114580262A (en) Lithium ion battery health state estimation method
CN113094985B (en) Battery SOH prediction method based on cross manifold embedded transfer learning
CN112327165A (en) Battery SOH prediction method based on unsupervised transfer learning
Zhang et al. Remaining useful life prediction of lithium-ion batteries based on TCN-DCN fusion model combined with IRRS filtering

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant