CN115951254A - Lithium ion battery health state and remaining service life estimation method - Google Patents
Lithium ion battery health state and remaining service life estimation method Download PDFInfo
- Publication number
- CN115951254A CN115951254A CN202211617711.4A CN202211617711A CN115951254A CN 115951254 A CN115951254 A CN 115951254A CN 202211617711 A CN202211617711 A CN 202211617711A CN 115951254 A CN115951254 A CN 115951254A
- Authority
- CN
- China
- Prior art keywords
- ant
- lion
- optimal
- fitness
- elite
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 42
- 230000036541 health Effects 0.000 title claims abstract description 38
- 229910001416 lithium ion Inorganic materials 0.000 title claims abstract description 17
- HBBGRARXTFLTSG-UHFFFAOYSA-N Lithium ion Chemical compound [Li+] HBBGRARXTFLTSG-UHFFFAOYSA-N 0.000 title claims abstract description 16
- 241001206881 Myrmeleon inconspicuus Species 0.000 claims abstract description 80
- 238000012549 training Methods 0.000 claims abstract description 35
- 239000013598 vector Substances 0.000 claims abstract description 31
- WHXSMMKQMYFTQS-UHFFFAOYSA-N Lithium Chemical compound [Li] WHXSMMKQMYFTQS-UHFFFAOYSA-N 0.000 claims abstract description 21
- 229910052744 lithium Inorganic materials 0.000 claims abstract description 21
- 238000012360 testing method Methods 0.000 claims abstract description 19
- 238000005457 optimization Methods 0.000 claims abstract description 12
- 230000035772 mutation Effects 0.000 claims abstract description 5
- 238000012545 processing Methods 0.000 claims abstract description 4
- 230000006870 function Effects 0.000 claims description 38
- 241000257303 Hymenoptera Species 0.000 claims description 27
- 241000282320 Panthera leo Species 0.000 claims description 18
- 241000258923 Neuroptera Species 0.000 claims description 16
- 230000008569 process Effects 0.000 claims description 7
- 238000013507 mapping Methods 0.000 claims description 6
- 230000001186 cumulative effect Effects 0.000 claims description 5
- 238000010277 constant-current charging Methods 0.000 claims description 4
- 238000012706 support-vector machine Methods 0.000 claims description 4
- 238000010280 constant potential charging Methods 0.000 claims description 3
- 230000009977 dual effect Effects 0.000 claims description 3
- 238000013508 migration Methods 0.000 claims description 3
- 230000005012 migration Effects 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 2
- 230000003862 health status Effects 0.000 claims 2
- 230000035508 accumulation Effects 0.000 claims 1
- 238000009825 accumulation Methods 0.000 claims 1
- 238000004146 energy storage Methods 0.000 abstract description 3
- 230000002431 foraging effect Effects 0.000 abstract description 2
- 238000007600 charging Methods 0.000 description 4
- 238000011161 development Methods 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 241000497425 Myrmeleon formicarius Species 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000007599 discharging Methods 0.000 description 1
- 239000007772 electrode material Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000003446 memory effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 239000007774 positive electrode material Substances 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000005295 random walk Methods 0.000 description 1
Images
Classifications
-
- 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
- Secondary Cells (AREA)
Abstract
The invention discloses a method for estimating the health state and the remaining service life of a lithium ion battery, belonging to the technical field of energy storage of new energy batteries; the estimation method comprises the following steps: s1, processing lithium battery data, and respectively extracting health factors related to SOH and RUL to construct feature vectors for generating a training sample and a test sample; s2, obtaining the position of the elite ant lion by using an ant lion optimization algorithm based on Cauchy variation as an optimal parameter combination; s3, inputting the optimal parameter combination of the S2 into a support vector regression model, training and testing the training sample and the testing sample generated in the S1 through the support vector regression model, and outputting SOH and RUL pre-estimated values; for the problem that the traditional population algorithm has local optimal stagnation, a Cauchy mutation operator is innovatively introduced, so that the algorithm jumps out of the local optimal foraging position as soon as possible to find a global optimal solution, the prediction precision of the algorithm is further improved, and the lithium battery state prediction precision and accuracy are improved.
Description
Technical Field
The invention belongs to the technical field of new energy battery energy storage, and particularly relates to a method for estimating the health state and the remaining service life of a lithium ion battery.
Background
The lithium ion battery has become a mainstream technical solution for electric vehicles and energy storage systems by virtue of the advantages of high energy density, low self-discharge rate, no memory effect, long service life and the like. However, the lithium battery has deteriorated performance in cycle use, and the safe and reliable use of new energy equipment is directly affected due to external environment, corrosion of internal electrode materials, aging of a diaphragm and the like. Therefore, accurate and efficient lithium battery health state estimation and residual life prediction are important prerequisites for safe operation of the battery management system.
In recent years, many methods have been proposed to predict the lithium battery state of health (SOH) and remaining life (RUL) accuracy, including a data-driven method, without knowing the internal degradation mechanism of the battery, and have become one of the main techniques for predicting the lithium battery state. However, a large amount of sufficient training sample data is needed for a data-driven modeling mode, and the lithium battery real-time monitoring data distribution has the characteristics of small sample data and long time window, so that available data is more lacking, which can seriously affect the lithium battery state estimation precision.
Disclosure of Invention
Aiming at the defects of the prior art, the invention aims to provide a method for estimating the health state and the remaining service life of a lithium ion battery.
The purpose of the invention can be realized by the following technical scheme:
a method for estimating the health state and the remaining service life of a lithium ion battery comprises the following steps:
s1, processing lithium battery data, and respectively extracting health factors related to SOH and RUL to construct feature vectors for generating a training sample and a test sample;
s2, obtaining the position of the elite ant lion by using an ant lion optimization algorithm based on Cauchy variation as an optimal parameter combination;
and S3, inputting the optimal parameter combination of the S2 into a support vector regression model, training and testing the training sample and the testing sample generated in the S1 through the support vector regression model, and outputting SOH and RUL estimated values.
Further, in S1, the health factors related to SOH include: average discharge voltage, constant current discharge time, constant voltage rise charge time interval and constant current fall charge time interval; health factors associated with RUL include: internal resistance, constant voltage charging time and constant current charging time.
Further, in S2, the step of obtaining the position of the elite lion includes:
s21, randomly initializing the position of each ant and each ant lion to obtain the position of each ant and each ant lion;
s22, substituting the positions of each ant and each ant lion obtained in the step S21 as parameters into a support vector regression model for training, calculating the fitness according to a fitness function established by mean square error, sequencing all ants and ant lions from high fitness to low fitness, and taking the individual with the optimal fitness value in the ant lion population as the current elite ant lion;
s23, performing optimization selection on ant lions through roulette search, capturing ants and judging iteration termination conditions; with the migration of ants, the adaptability of the ants becomes high and falls into a trap, and then elite lion is selected and iterated, and the elite lion is updated; in the searching process, the selection cumulative probability q of each elite ant lion is calculated according to the fitness obtained in the step S22 i Randomly generating r E [0,1 ∈]If q is i >r, selecting the ant lion to be stored as elite ant lion; searching until the maximum iteration number is reached, and taking the saved position of the elite ant lion as an optimal parameter combination; when the two adjacent optimal values in the ant lion colony are recorded as historical optimal values, namely the algorithm falls into local optimal stagnation, the operation enters S24;
s24, updating positions of ants and ant lions based on Cauchy variation;
s25, calculating the updated fitness of the ants, and judging whether the ant lion position is the best ant lion position; if the ant fitness of the position is higher than the fitness of the updated ant lion, the maximum precision is achieved, or the maximum iteration times of the algorithm are achieved, the optimal ant lion position is obtained, otherwise, the algorithm is switched to S22.
Further, in S21, ants are based on: x (t) = [0,cumnum (2 r (t)) 1 )-1),···,cumsum(2r(t n )-1)]Walking randomly in the search space;
in the formula: cumsum is the calculated running sum, r (t) is a random function:
further, in S22, the fitness function is:
where MSE represents the mean square error between the actual and predicted values, n is the number of samples in the training sample set,
and y i Predicted and actual values for the ith SOH and RUL, respectively.
Further, in S23, the cumulative probability of choice of each elite lion is calculated according to the fitness obtained in S22, and the expression is as follows:
further, in S24, a cauchy mutation operation is embedded, the ant lion population position is updated, the optimal value is updated, and the ant lion position with the maximum fitness in the ant lion population is updated to the position of the elite ant lion, which is expressed as:
Ant t ＝Ant 0 +Ant o *Cauchy(0,1)
Antlion t ＝Antlion 0 +Antlion 0 *Cauchy(0,1)
wherein, ant 0 、Antlion 0 Initial positions of the original Ant lion and Ant, ant t 、Antlion t Is the new position updated by ant lion and ant after Cauchy variation, and Cauchy (0, 1) is the standard Cauchy distribution when t = 1.
Further, in S3, the specific steps of obtaining the estimated SOH and RUL values include:
s31, combining the optimal parameters obtained in the S2 to serve as parameters of the support vector regression model, inputting the training samples in the S1 into the model, and training the optimal support vector regression model;
and S32, verifying the trained model by the S31 through the test sample obtained in the S1, and predicting the SOH and RUL of the battery.
Further, in S31, the specific steps of training the optimal support vector regression model are as follows:
1) Mapping the low-dimensional samples to a higher-dimensional space by a nonlinear mapping function phi (x) with a functional relationship of f (x) = w.phi (x) + b; wherein w is a weight parameter and b is a bias parameter; establishing an objective function:
wherein ξ i * And xi i Is the relaxation variable of the ith training sample, and epsilon is an insensitive loss function; c is a penalty parameter;
2) Solving an objective function through Lagrange, converting the problem of finding an optimal value with conditions into a function with unlimited conditions, solving partial derivatives of each parameter, and adopting a regression function formula of a support vector machine with dual theorems and a kernel function as follows:
wherein x is i For the ith feature input vector, K (x) i X) is a kernel function, α i And alpha i * Is a lagrange multiplier.
A computer storage medium stores a readable program, and when the program runs, executes the above estimation method.
The invention has the beneficial effects that:
1. aiming at the problem that the traditional population algorithm has local optimal stagnation, a Cauchy mutation operator is innovatively introduced to enable the algorithm to jump out of a local optimal foraging position as soon as possible to find a global optimal solution, so that the prediction precision of the algorithm is improved, the convergence time of the algorithm is reduced, the design is reasonable, the lithium battery state prediction precision and accuracy are improved, and the method has great significance in improving the service performance and the safety performance of the battery;
2. the method selects the support vector regression model, has excellent performance on a battery data set with the characteristics of small samples and nonlinearity, solves the problem of parameter selection of the support vector regression model through the ant lion intelligent algorithm, and improves the generalization capability and the fitting capability of model training;
3. compared with the traditional SOH prediction and residual life prediction method, the method can realize high-precision and real-time online multi-state estimation of the lithium ion battery, and has better practicability in practical engineering application;
4. the method can be applied to new energy application scenes such as development of a battery management system BMS of a new energy automobile, and can effectively solve the problem that the conventional battery management system BMS algorithm is lack of adaptability and interpretability.
Drawings
In order to more clearly illustrate the embodiments or prior art solutions of the present invention, the drawings used in the description of the embodiments or prior art will be briefly described below, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
Fig. 1 is a flow chart illustrating estimation of a state of health and a remaining life of a lithium battery according to an embodiment of the present invention;
FIG. 2 is a graph of operational parameters of a battery data set used in an embodiment of the present invention;
fig. 3 is a schematic diagram of curves of four health factors of a B05 battery in the method for estimating the SOH state of a lithium battery according to the embodiment of the present invention;
fig. 4 is a schematic diagram illustrating curves of three health factors of a CS35 battery in an estimation method of a RUL state of a lithium battery according to an embodiment of the present invention;
FIG. 5 is a comparison chart of estimation results of lithium battery state of health estimation using different methods according to an embodiment of the present invention;
fig. 6 is a diagram comparing the Mean Absolute Error (MAE) and the Root Mean Square Error (RMSE) when different methods are used in the method for estimating the state of health and the remaining life of the lithium battery according to the embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. 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.
As shown in fig. 1, a method for estimating the state of health and remaining service life of a lithium ion battery includes the following steps:
s1, preprocessing lithium battery data;
and respectively extracting health factors with high correlation degree with the state of health (SOH) and the remaining service life (RUL) to construct a feature vector so as to generate a training sample and a test sample. Health state estimation the health factors chosen include: average discharge voltage, constant current discharge time, constant voltage rise charge time interval and constant current fall charge time interval; the health factors selected for the estimation of the residual service life are as follows: internal resistance, constant voltage charging time and constant current charging time; the proportion of training samples and test samples was 60% and 40%, respectively.
S2, carrying out CALO algorithm optimization on the parameters to obtain the position of the elite ant lion as an optimal parameter combination;
compared with other data driving methods, the Support Vector Regression (SVR) has perfect theoretical basis and strong anti-generalization capability, and has unique advantages in processing the small sample and nonlinear fitting problems of lithium battery state estimation. In order to effectively solve the problem of parameter selection of the SVR prediction model, the traditional ant lion optimization algorithm ALO has good exploration and development capabilities, however, the algorithm still has the following problems: the random walk of ants is easily limited around elite ant lions, the global exploration and development capability is lost, and the optimization precision is reduced.
Aiming at the situations, an improved ant lion optimization algorithm CALO is provided, a Cauchy mutation operator and a weighted elite strategy are added, the proportion of ants randomly walking around elite ant lions and ant lions selected by roulette in different periods can be adjusted, the random walking of the ants is greatly influenced by the ant lions selected by the roulette in the early period of iteration, the random walking of the ants is greatly influenced by the elite ant lions in the later period of iteration, the exploration capability and the development capability of the ALO algorithm are effectively improved, the situation that the ALO algorithm is trapped in local optimization is avoided, and the convergence precision is improved.
The specific CALO algorithm process is as follows:
s21, randomly initializing the population: the position of each ant and each ant lion in the improved ant lion optimization algorithm (CALO) corresponds to a group of parameter combinations (C, sigma) to be optimized in the support vector regression model, and the position of each ant and each ant lion is randomly initialized according to the parameter ranges of the penalty factor C and the kernel function parameter sigma in the support vector regression model; ants are according to the following steps: x (t) = [0,cumnum (2 r (t)) 1 )-1),···,cumsum(2r(t n )-1)]Randomly walking in the search space to obtain the positions of each ant and each ant lion, wherein the position of the ith ant or the ith ant lion is (C) i ，σ i ) (ii) a In the above formula, cumsum is the calculated cumulative sum, and r (t) is a random function:
s22, calculating a fitness function value: and (3) substituting the positions of each ant and each ant lion obtained in the step (S21) as parameters into a support vector regression model for training, and establishing a fitness function according to the Mean Square Error (MSE):
wherein MSE represents the mean square error between the actual value and the predicted value, n is the number of samples in the training sample set,
and y i Predicted values and actual values for the ith SOH and RUL, respectively; the fitness of all ants and ant lions is ranked from high to low, and the individual with the optimal fitness value in the ant lions population is used as the current elite ant lions; traps are established proportional to fitness. />
In addition, after the positions of each ant and each ant lion obtained in the step S21 are substituted as parameters into the support vector regression model for training, an evaluation standard can be established by using the average absolute error and the root mean square error as fitness functions, that is, the formulas are respectively
Wherein MAE is the mean absolute error, RMSE is the root mean square error,
and y i The predicted value and the actual value of the ith SOH and RUL are respectively, and n is the number of samples.
S23, performing optimization selection on ant lions through roulette search, capturing ants and judging iteration termination conditions: with the migration of ants, the adaptability of the ants becomes high and the ants fall into traps, and then elite lion is selected and iterated, and elite lion is updated; in the searching process, the selection cumulative probability of each elite lion is calculated according to the fitness MSE obtained in S22, namely:
wherein q is i Accumulating probabilities for each individual ant lion; at this time, r is randomly generated to be [0,1 ]]If q is i >r, selecting the ant lion to be stored as elite ant lion; always on searchUntil reaching the maximum iteration times, taking the saved position of the elite ant lion as the optimal parameter combination; when the optimal values with almost no difference between two adjacent iterations in the ant lion group are recorded as historical optimal values, namely the algorithm falls into local optimal stagnation, the next step is carried out.
S24, updating positions of ants and ant lions based on Cauchy variation: embedding Kouchi variation operation, further updating the ant lion group positions and updating the optimal values, updating the ant lion position with the maximum fitness in the ant lion group to be the position of the elite ant lion, wherein the expression is as follows:
Ant t ＝Ant 0 +Ant o *Cauchy(0,1)
Antlion t ＝Antlion 0 +Antlion 0 *Cauchy(0,1)
wherein, ant 0 、Antlion 0 Initial positions of the original Ant lion and Ant, ant t 、Antlion t Is the new position updated by ant lion and ant after Cauchy variation, and Cauchy (0, 1) is the standard Cauchy distribution when t = 1.
S25, calculating the updated fitness of the ants, and judging whether the ant lion position is the best ant lion position; if the ant fitness of the position is higher than the fitness of the updated ant lion, the maximum precision is achieved, or the maximum iteration times of the algorithm are achieved, the optimal ant lion position is obtained, otherwise, the algorithm is switched to S22.
S3, training and testing the lithium battery characteristic set (training sample and testing sample) obtained in the S1 through a support vector regression model (SVR model), and outputting the SOH and RUL pre-estimated values;
support Vector Regression (SVR) is an extension of a support vector machine in the regression field, and is often used for solving the problems of nonlinearity and small samples due to good generalization capability and high convergence rate. Because the battery health experiment period is long and the sample size is small, the SVR is very suitable for battery data driven prediction.
The specific algorithm process is as follows:
and S31, taking the optimal parameter combination (C, sigma) obtained in the S2 as a parameter of the Support Vector Regression (SVR) model, inputting the training samples in the data set preprocessed in the S1 into the model, and training out the optimal SVR model.
The SVR model is defined as follows:
mapping the low-dimensional samples to a higher-dimensional space by a nonlinear mapping function phi (x) with a functional relationship of f (x) = w.phi (x) + b; wherein w is a weight parameter and b is a bias parameter; establishing an objective function:
wherein ξ i * And xi i The relaxation variable of the ith training sample is epsilon, and the epsilon is an insensitive loss function; and C is a penalty parameter which is used for applying penalty to the observed value outside the control boundary so as to prevent overfitting.
Solving an objective function through Lagrange, converting the problem of finding an optimal value with conditions into a function with unlimited conditions, solving partial derivatives of each parameter, utilizing dual theorem and adding a kernel function, and obtaining a regression function formula of the trained optimal support vector machine, wherein the regression function formula is as follows:
wherein x is i For the ith feature input vector, K (x) i X) is a kernel function, α i And alpha i * Is a lagrange multiplier.
S32, verifying the trained model by the S31 through the test samples in the S1 data set, and predicting the state of health (SOH) and the remaining service life (RUL) of the battery;
and (3) taking the test sample as the input of the model trained in S31, and selecting an unoptimized ALO-SVR algorithm for comparison in order to verify the prediction performance of the CALO-SVR algorithm.
In order to compare the prediction performance of the algorithm more intuitively, the average absolute error (MAE) and the root Mean Square Error (MSE) are selected as fitness functions to establish evaluation criteria, namely the formulas are respectively as follows:
wherein: MAE is the mean absolute error, RMSE is the root mean square error,
and yi are the predicted and actual values of the ith SOH and RUL, respectively, and n is the number of samples.
The feasibility and the effectiveness of the invention on the prediction of the state of health and the residual life of the battery are verified through experiments as follows:
the operating parameters of the battery data set used in the present invention are shown in fig. 2. The health state estimation experimental data adopted by the invention is from NASAPCIE research center, the data set comprises the charge and discharge experimental data of four lithium ion 18650 type rechargeable batteries, the rated capacity of the battery is 2 A.h, and the charge cut-off voltage is 4.2V. Experimental data for four li-ion batteries, no. B5, B6, B7 and B18, operating at 24 degrees room temperature in 3 different operating modes, the charging process was carried out in a constant current mode of 1.5A until the battery voltage reached 4.2V, and then continued in a constant voltage mode until the charging current dropped to 20mA. The discharging process is performed with a constant current of 2A until the cell voltages of the cells 5, 6, 7 and 18 are reduced to 2.7V, 2.5V, 2.2V and 2.5V, respectively. The residual life prediction experimental data are from CALCE research center of Maryland university, and the data set comprises 4 batteries with positive electrode materials of lithium cobaltate and nominal capacities of 1 A.h: CS35, CS36, CS37 and CS38. Constant current charging was performed at a rate of 0.5C under the same room temperature environment until the cell voltage reached 4.2V. The charging is then continued in constant voltage mode until the charging current drops to 50mA. The discharge test was performed in 1C constant current output mode until the voltage dropped to 2.7V.
The feasibility and the effectiveness of a prediction model are verified by using a B5 battery with 168 charge-discharge cycle times and a CS35 battery with 882 cycle times, the first 60% of samples are used as a training sample set, the remaining 40% of samples are used as a test sample set, a health factor selected according to a charge-discharge curve is used as a sample input feature, and the SOH and the RUL of the battery are respectively used as sample output features. Fig. 3 shows four health factors selected for state of health prediction in the present invention, and fig. 4 shows three health factors selected for remaining life prediction.
According to the invention, two methods, namely ALO-SVR and CALO-SVR, are selected to compare SOH of the battery, and the SOH corresponds to three lithium batteries, and the labels are B5, B6 and B07 respectively. FIG. 5 shows the SOH estimation results of the battery B05 under the two methods, when 60% of the samples are used as the training set, the comparison can be made, the SOH estimated by CALO-SVR is the most accurate, and the fitness of ALO-SVR is not high enough. The ALO-SVR has better estimation accuracy in the early stage, but gradually deviates from the true SOH in the later stage. In the later stage, the prediction precision of the CALO-SVR is still very high compared with that of the ALO-SVR, and is caused by dynamic adjustment of the inertial weight of the common ant lion.
Fig. 6 shows performance evaluation indicators of two algorithms, i.e., mean Absolute Error (MAE) and Root Mean Square Error (RMSE), and the smaller the corresponding value, i.e., the smaller the error, the better the performance. The estimated performance of the CALO-SVR is significantly improved compared to the ALO-SVR. The maximum MAE and RMSE of CALO-SVR were 0.0226 and 0.0335, respectively, for different types of batteries. MAE and RMSE were reduced by 0.0320 and 0.0311 on average for CALO-SVR compared to ALO-SVR. Furthermore, it can be seen that in the SOH estimation of B5 and the residual life prediction performance of CS35 and CS37, the difference between ALO-SVR and CALO-SVR is relatively small, while the SOH estimation of B06 and B07 and the RUL estimation performance of CS38 are significantly behind the CALO-SVR because the SOH curves of B06 and B07 and the RUL curve of CS38 fluctuate greatly, so the fitting ability and generalization ability of ALO-SVR are inferior to those of CALO-SVR.
In the description herein, references to the description of "one embodiment," "an example," "a specific example" or the like are intended to mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
The foregoing shows and describes the general principles, principal features, and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are described in the specification and illustrated only to illustrate the principle of the present invention, but that various changes and modifications may be made therein without departing from the spirit and scope of the present invention, which fall within the scope of the invention as claimed.
Claims (10)
1. A lithium ion battery health state and remaining service life estimation method is characterized by comprising the following steps:
s1, processing lithium battery data, and respectively extracting health factors related to SOH and RUL to construct a feature vector to generate a training sample and a test sample;
s2, obtaining the position of the elite ant lion by using an ant lion optimization algorithm based on Cauchy variation as an optimal parameter combination;
and S3, inputting the optimal parameter combination of the S2 into a support vector regression model, training and testing the training sample and the testing sample generated in the S1 through the support vector regression model, and outputting SOH and RUL estimated values.
2. The method of claim 1, wherein in S1, the health factors related to SOH include: average discharge voltage, constant current discharge time, constant voltage rise charge time interval and constant current fall charge time interval; health factors associated with RUL include: internal resistance, constant voltage charging time and constant current charging time.
3. The method for estimating the health status and the remaining service life of the lithium ion battery according to claim 1, wherein in S2, the step of obtaining the position of the elite lion comprises:
s21, randomly initializing the position of each ant and each ant lion to obtain the position of each ant and each ant lion;
s22, substituting the position of each ant and each ant lion obtained in the S21 as a parameter into a support vector regression model for training, calculating the fitness according to a fitness function established by mean square error, sequencing all ants and ant lions from high fitness to low fitness, and taking the individual with the optimal fitness value in the ant lion population as the current elite lion;
s23, performing optimization selection on ant lions through roulette search, capturing ants and judging iteration termination conditions; with the migration of ants, the adaptability of the ants becomes high and the ants fall into traps, and then elite lion is selected and iterated, and elite lion is updated; in the searching process, the selection accumulated probability q of each elite lion is calculated according to the fitness obtained in the step S22 i Randomly generating r E [0,1 ∈]If q is i >r, selecting the ant lion to be stored as elite ant lion; searching until the maximum iteration number is reached, and taking the saved position of the elite lion as an optimal parameter combination; when the two adjacent optimal values in the ant lion colony are recorded as historical optimal values, namely the algorithm falls into local optimal stagnation, the operation enters S24;
s24, updating positions of ants and ant lions based on Cauchy variation;
s25, calculating the updated fitness of the ants, and judging whether the ant lion position is the best ant lion position; if the ant fitness of the position is higher than the fitness of the updated ant lion, the maximum precision is achieved, or the maximum iteration times of the algorithm are achieved, the optimal ant lion position is obtained, otherwise, the algorithm is switched to S22.
4. The method of claim 3, wherein in S21, ants estimate the health status and remaining service life of the lithium ion battery according to: x (t) = [0,cumnum (2 r (t)) 1 )-1),···,cumsum(2r(t n )-1)]In searchWalking randomly in space;
in the formula, cumsum is the sum of the calculated accumulations, and r (t) is a random function:
5. the method according to claim 3, wherein in S22, the fitness function is as follows:
where MSE represents the mean square error between the actual and predicted values, n is the number of samples in the training sample set,
and y i Predicted and actual values for the ith SOH and RUL, respectively. />
6. The method for estimating the health state and remaining service life of a lithium ion battery as claimed in claim 3, wherein in S23, the cumulative probability of choice of each elite lion is calculated according to the fitness obtained in S22, and the expression is as follows:
7. the method for estimating the health state and the remaining service life of the lithium ion battery according to claim 3, wherein a Cauchy variation operation is embedded in S24, the positions of the lion colony are updated, the optimal value is updated, the position of the lion with the highest fitness in the lion colony is updated to the position of the lion, and the expression is as follows:
Ant t ＝Ant 0 +Ant o *Cauchy(0,1)
Antlion t ＝Antlion 0 +Antlion 0 *Cauchy(0,1)
wherein, ant 0 、Antlion 0 Initial positions of the original Ant lion and Ant, ant t 、Antlion t The new location of ant lion and ant renewal after Cauchy mutation, cauchy (0, 1), is the standard Cauchy distribution at t = 1.
8. The method of claim 1, wherein in step S3, the steps of obtaining the estimated SOH and RUL values comprise:
s31, using the optimal parameter combination obtained in the S2 as a parameter of the support vector regression model, inputting the training sample in the S1 into the model, and training out the optimal support vector regression model;
and S32, verifying the trained model by the S31 through the test sample obtained in the S1, and predicting the SOH and RUL of the battery.
9. The method for estimating the state of health and remaining service life of a lithium ion battery according to claim 8, wherein in S31, the specific steps of training the optimal support vector regression model are as follows:
1) Mapping the low-dimensional samples to a higher-dimensional space by a nonlinear mapping function phi (x), wherein the function relation is f (x) = w.phi (x) + b; wherein w is a weight parameter and b is a bias parameter; establishing an objective function:
wherein ξ i * And xi i The relaxation variable of the ith training sample is epsilon, and the epsilon is an insensitive loss function; c is a penalty parameter;
2) Solving an objective function through Lagrange, converting the problem of finding an optimal value with conditions into a function with unlimited conditions, solving partial derivatives of each parameter, and adopting a regression function formula of a support vector machine with dual theorems and a kernel function as follows:
in the formula, x i For the ith feature input vector, K (x) i X) is a kernel function, α i And alpha i * Is a lagrange multiplier.
10. A computer storage medium, in which a readable program is stored which, when executed, performs the method of any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211617711.4A CN115951254A (en) | 2022-12-15 | 2022-12-15 | Lithium ion battery health state and remaining service life estimation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211617711.4A CN115951254A (en) | 2022-12-15 | 2022-12-15 | Lithium ion battery health state and remaining service life estimation method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115951254A true CN115951254A (en) | 2023-04-11 |
Family
ID=87281793
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211617711.4A Pending CN115951254A (en) | 2022-12-15 | 2022-12-15 | Lithium ion battery health state and remaining service life estimation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115951254A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116593903A (en) * | 2023-07-17 | 2023-08-15 | 中国华能集团清洁能源技术研究院有限公司 | Battery remaining life prediction method and device |
| CN116611004A (en) * | 2023-07-20 | 2023-08-18 | 江苏中矿安华科技发展有限公司 | Coal mine safety disaster prediction model training method, early warning method and early warning system |
| CN116609668A (en) * | 2023-04-26 | 2023-08-18 | 淮阴工学院 | Lithium ion battery health state and residual life prediction method |
| CN117129899A (en) * | 2023-08-31 | 2023-11-28 | 重庆跃达新能源有限公司 | Battery health state prediction management system and method |
| CN117233540A (en) * | 2023-11-15 | 2023-12-15 | 广东电网有限责任公司江门供电局 | Metering pipeline fault detection method and system based on deep learning |
| CN117825975A (en) * | 2024-03-05 | 2024-04-05 | 烟台海博电气设备有限公司 | Data-driven lithium ion battery SOH evaluation method and system |
| CN118169582A (en) * | 2024-05-15 | 2024-06-11 | 东方旭能(山东)科技发展有限公司 | Lithium ion battery health state and residual life prediction method |
| CN118191623A (en) * | 2024-05-17 | 2024-06-14 | 西北工业大学宁波研究院 | Prediction method of residual service life of fuel cell based on AGA-SVR |
-
2022
- 2022-12-15 CN CN202211617711.4A patent/CN115951254A/en active Pending
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116609668B (en) * | 2023-04-26 | 2024-03-26 | 淮阴工学院 | Lithium ion battery health state and residual life prediction method |
| CN116609668A (en) * | 2023-04-26 | 2023-08-18 | 淮阴工学院 | Lithium ion battery health state and residual life prediction method |
| CN116593903B (en) * | 2023-07-17 | 2023-10-20 | 中国华能集团清洁能源技术研究院有限公司 | Battery remaining life prediction method and device |
| CN116593903A (en) * | 2023-07-17 | 2023-08-15 | 中国华能集团清洁能源技术研究院有限公司 | Battery remaining life prediction method and device |
| CN116611004A (en) * | 2023-07-20 | 2023-08-18 | 江苏中矿安华科技发展有限公司 | Coal mine safety disaster prediction model training method, early warning method and early warning system |
| CN116611004B (en) * | 2023-07-20 | 2023-11-10 | 江苏中矿安华科技发展有限公司 | Coal mine safety disaster prediction model training method, early warning method and early warning system |
| CN117129899A (en) * | 2023-08-31 | 2023-11-28 | 重庆跃达新能源有限公司 | Battery health state prediction management system and method |
| CN117129899B (en) * | 2023-08-31 | 2024-05-10 | 重庆跃达新能源有限公司 | Battery health state prediction management system and method |
| CN117233540A (en) * | 2023-11-15 | 2023-12-15 | 广东电网有限责任公司江门供电局 | Metering pipeline fault detection method and system based on deep learning |
| CN117233540B (en) * | 2023-11-15 | 2024-02-20 | 广东电网有限责任公司江门供电局 | Metering pipeline fault detection method and system based on deep learning |
| CN117825975A (en) * | 2024-03-05 | 2024-04-05 | 烟台海博电气设备有限公司 | Data-driven lithium ion battery SOH evaluation method and system |
| CN118169582A (en) * | 2024-05-15 | 2024-06-11 | 东方旭能(山东)科技发展有限公司 | Lithium ion battery health state and residual life prediction method |
| CN118191623A (en) * | 2024-05-17 | 2024-06-14 | 西北工业大学宁波研究院 | Prediction method of residual service life of fuel cell based on AGA-SVR |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115951254A (en) | Lithium ion battery health state and remaining service life estimation method | |
| Chaoui et al. | State of charge and state of health estimation for lithium batteries using recurrent neural networks | |
| CN107368619B (en) | Extended Kalman filtering SOC estimation method | |
| CN113253116A (en) | Lithium ion battery state of charge estimation method and storage medium | |
| CN107957562A (en) | A kind of lithium ion battery residual life on-line prediction method | |
| CN114636932A (en) | Method and system for predicting remaining service life of battery | |
| CN106885994B (en) | Lithium battery remaining life fast detection method based on constant-voltage charge segment | |
| CN114035098A (en) | Lithium battery health state prediction method integrating future working condition information and historical state information | |
| Li et al. | Lithium-ion battery remaining useful life prognostics using data-driven deep learning algorithm | |
| CN114089193A (en) | Method and device for estimating temperature and negative pole potential of battery on line and computer equipment | |
| CN113884936B (en) | ISSA coupling DELM-based lithium ion battery health state prediction method | |
| GB2600129A (en) | Pro-active battery management system (BMS) with lossless active buck balancing and method thereof | |
| CN114814591A (en) | Lithium battery SOE estimation method, device and system | |
| Sun et al. | Battery degradation evaluation based on impedance spectra using a limited number of voltage-capacity curves | |
| CN117420442A (en) | MNQ-LSTM-based battery remaining life prediction method | |
| CN116774045A (en) | Lithium battery health state prediction method based on HHO-SVR | |
| CN111762059A (en) | Multivariable fusion battery pack balancing method considering battery charging and discharging working conditions | |
| CN114896865B (en) | Digital twinning-oriented self-adaptive evolutionary neural network health state online prediction method | |
| Wang et al. | Online battery pack state of charge estimation via ekf-fuzzy logic joint method | |
| CN116258059A (en) | Lithium battery temperature real-time prediction model based on multi-time space width learning model and construction method | |
| Zheng et al. | Remaining useful life indirect prediction of lithium-ion batteries using CNN-BiGRU fusion model and TPE optimization. | |
| CN115308623A (en) | Battery state of charge estimation method based on particle resampling and searcher optimization algorithm | |
| Benrabia et al. | Modeling and Evaluation of Dynamical Properties of Different Energy Storage Systems Using Machine Learning Methods | |
| Sui et al. | Early Prediction of Lithium-Ion Batteries Lifetime via Few-Shot Learning | |
| Fan et al. | LSTM-Attention Mechanism based Remaining Useful Life Prediction of Lithium Batteries |
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 |