CN108804764A - A kind of aging of lithium battery trend forecasting method based on extreme learning machine - Google Patents
A kind of aging of lithium battery trend forecasting method based on extreme learning machine Download PDFInfo
- Publication number
- CN108804764A CN108804764A CN201810427737.XA CN201810427737A CN108804764A CN 108804764 A CN108804764 A CN 108804764A CN 201810427737 A CN201810427737 A CN 201810427737A CN 108804764 A CN108804764 A CN 108804764A
- Authority
- CN
- China
- Prior art keywords
- model
- neuron
- learning machine
- data set
- lithium battery
- 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
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
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- Bioinformatics & Computational Biology (AREA)
- Software Systems (AREA)
- Evolutionary Biology (AREA)
- Mathematical Physics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Geometry (AREA)
- Genetics & Genomics (AREA)
- Physiology (AREA)
- Computer Hardware Design (AREA)
- Secondary Cells (AREA)
Abstract
The invention discloses a kind of aging of lithium battery trend forecasting method based on extreme learning machine, original time series Accurate Model of this method limit of utilization learning machine to the lithium battery charging voltage of acquisition, using Volterra series models as the input layer of extreme learning machine model, simultaneously, to improve the accuracy of cell degradation data model, the hidden layer neuron with higher precision of prediction is generated by genetic algorithm in the structure forecast model stage, aging of lithium battery trend is predicted by the prediction model of lithium battery, the results show, this method has good estimated performance, precision is high.
Description
Technical field
The invention belongs to battery life predicting technical fields, more specifically, are related to a kind of based on extreme learning machine
Aging of lithium battery trend forecasting method.
Background technology
With the gradually popularization of the new energy vehicles such as electric vehicle, lithium ion battery is as main " green energy resource "
It has received widespread attention.Lithium battery due to it is light-weight, adaptable, environmentally protective the advantages that, become the weight in new energy traffic
Want the energy.However lithium battery is in use, with the increase of charge and discharge number, the capacity of battery can constantly decay, and work as drop
When as low as below some threshold value, battery is up to end of life.Therefore, during the use of battery, to cell degradation data
Trend prediction is carried out, its health status is understood in due course, knows life information in advance, to be effectively prevented from since the energy content of battery declines
The safety accident for moving back initiation, is of great significance.
Currently, common battery life predicting method is divided into two classes:The physics of failure (PhysicsofFailure, PoF) mould
Type and data-driven method (DataDriven).Failure physical model needs largely about lithium ion battery material characteristic and failure
The information of mechanism, but sensor technology common at present is difficult to measure the electrochemical reaction of inside battery.Data-driven method is logical
The training that monitoring cell degradation data carry out prediction model is crossed, the implicit information between input and output is excavated, to not
The decline for carrying out energy carries out trend prediction.Compared with physical failure modelling, data-driven method need not specifically study failure
Mechanism is realized simply, therefore has obtained universal application.But current data-driven fado is agreed with now using analytic modell analytical model
There is sample data, since cell performance degradation rule is complicated, non-linear relation is apparent, the analytic modell analytical model found by such methods
And realistic model can have certain deviation, and the precision of life prediction result is caused to reduce.
Invention content
It is an object of the invention to overcome the deficiencies of the prior art and provide a kind of aging of lithium battery based on extreme learning machine
Trend forecasting method, this method limit of utilization learning machine accurately build the original time series of the lithium battery charging voltage of acquisition
Mould, using Volterra series models as the input layer of extreme learning machine model, meanwhile, to improve cell degradation data model
Accuracy generates the hidden layer neuron with higher precision of prediction by genetic algorithm in the structure forecast model stage, passes through
The prediction model of lithium battery predicts that aging of lithium battery trend, the results show, this method have good estimated performance, precision
It is high.
For achieving the above object, a kind of aging of lithium battery trend forecasting method based on extreme learning machine of the present invention,
Specifically include following steps:
(1) original time series of lithium battery charging voltage are acquired;
(2) phase space reconfiguration is carried out to the original time series of acquisition to correspond to minimum entropy rate using optimal entropy rate method
Delay time and Embedded dimensions as optimal phase space model of delay time and Embedded dimensions, obtain phase space model;
(3) according to phase space Construction of A Model Volterra series models, composition sequence is obtained, and using composition sequence as defeated
Enter sequence, the list entries is divided into learning data set, correction data set and test data set;
(4) it using the learning data set as the input layer of extreme learning machine model, is selected according to maximum relation degree principle
Each optimal neuron, and each optimal neuron is added to the hidden layer of the extreme learning machine model, construct initial hidden layer;
(5) the extreme learning machine model established by initial hidden layer using the correction data set re -training is based on Cp
Criterion determines the neuron number of final hidden layer, to obtain initial predicted model;
(6) learning data set and correction data set are combined as to the input layer of the initial predicted model, passed through
Training determines the hidden layer of the initial predicted model again, obtains final prediction model;
(7) test data set is input to the final prediction model and carries out prediction aging of lithium battery trend.
Advantageous effect of the present invention:Terseness, the height of the accuracy and extreme learning machine of present invention combination Volterra models
The characteristics of robustness, this method limit of utilization learning machine method is to time series Accurate Model, with Volterra series models
Input layer of the composition sequence as extreme learning machine model, while having by genetic algorithm generation in the structure forecast model stage
The hidden layer neuron of higher precision of prediction further increases the precision of sequence prediction.The simulation experiment result show set forth herein
Method on precision of prediction more existing several typical prediction techniques have higher promotion, and tectonic model is more succinct,
Accurately, efficiently.
In order to better illustrate technical solution of the present invention, following improvement can be also done:
Further, further include using the minimum angle Return Law in the Volterra series models of construction in the step (3)
Parameter screened.
Advantageous effect using above-mentioned further scheme is:So that the cell degradation Data Tendency Forecast Based model established is more
Accurately.
Further, the step (4) specifically includes:
(41) weights and deviation for waiting for each neuron in scavenger of hidden layer are generated at random;
(42) using the learning data set as the input layer of extreme learning machine model, calculating waits for each neuron in scavenger
Output valve, and calculate the degree of correlation of current predictive residual error corresponding with real output value;
(43) it selects and is added as optimal neuron, and by optimal neuron with each maximum neuron of the output valve degree of correlation
In current hidden layer, and update the output layer weights and desired output of extreme learning machine model;
(44) according to maximum relation degree principle, the weights and deviation of each neuron in scavenger are waited for using genetic algorithm update,
It selects with the maximum each neuron of the current desired output degree of correlation as optimal neuron, and institute is added in each optimal neuron
The hidden layer for stating extreme learning machine model constructs initial hidden layer.
Advantageous effect using above-mentioned further scheme is:Select the neuron for generating optimal hidden layer.
Further, the C of the step (5)pCriterion is:
Wherein, i is i-th of neuron, RSSPFor assessment prediction model for the precision of prediction of correction data set, definition
For:
Wherein, j is the number of data intensive data, tjFor the voltage actual value at jth moment, yjLearn for the jth moment limit
The output valve of model;
MSEfullIt is precision of prediction value of the learning data set under the initial predicted model comprising whole neurons, l is school
The parameter number of correction data;
By minimum CpThe corresponding extreme learning machine model of value is as initial predicted model.
Advantageous effect using above-mentioned further scheme is:Generate the hidden layer with minimum neuron.
Description of the drawings
Fig. 1 is the prediction technique flow chart of the present invention;
Fig. 2 is the effect that the present invention predicts four groups of lithium battery charging voltage datas of acquisition using prediction technique
Figure.
Specific implementation mode
The specific implementation mode of the present invention is described below in conjunction with the accompanying drawings, preferably so as to those skilled in the art
Understand the present invention.Requiring particular attention is that in the following description, when known function and the detailed description of design perhaps
When can desalinate the main contents of the present invention, these descriptions will be ignored herein.
As shown in Figure 1, the present invention is low to service life of lithium battery prediction result precision in the prior art in order to solve the problems, such as, carry
A kind of aging of lithium battery trend forecasting method based on extreme learning machine is gone out, specific steps include:
(1) the original time series X={ x of lithium battery charging voltage are acquired1,x2,...,xN, wherein xiIndicated for the i-th moment
Lithium battery charging voltage numerical value;
(2) to the original time series X={ x of acquisition1,x2,...,xNPhase space reconfiguration is carried out, utilize optimal entropy rate side
Method calculates delay time T and Embedded dimensions m under optimum state, and using delay time T with Embedded dimensions m as optimal mutually empty
Between model delay time ToptWith Embedded dimensions mopt, obtain phase space model Xs;The entropy rate of the phase space tectonic model can determine
Justice is:
Wherein, H (x, m, τ) is micro- entropy of present construction phase space sequence,<H(x,m,τ)>For the average value of micro- entropy, ρjFor
Euclidean distance between j-th of delay vector and nearest neighbor point, n=N- (m-1) τ, and CeFor Euler's constant, value is
0.5772.When entropy rate reaches minimum, corresponding phase space model is optimal solution, and according to its corresponding parameter, construction is mutually empty
Between model be:
Xs={ xt,xt-τ,xt-2τ,...,xt-(m-1)τ}
Wherein, XsFor the phase space model established according to delay time T and sequence dimension m, xt-kτFor original time series
The voltage value at t-k τ moment, k=0,1,2 ... (m-1),.
(3) according to phase space model, Volterra series models are constructed, this method is using Second-Order Volterra series
Model:
Wherein, Xv,iFor Second-Order Volterra series model, xi-kτIt is initiation sequence in the voltage value at i-k τ moment, k=0,
1,2 ... (m-1), xi-pτxi-qτFor the quadratic term of Volterra series, p=0,1,2 ... (m-1), q=, 1,2 ... (m-
1)。
It, can be to the parameter in model into traveling to make the cell degradation Data Tendency Forecast Based model established more accurate
The screening of one step.This method screens the parameter in the Volterra series models of construction using the minimum angle Return Law, will
For composition sequence after screening as list entries, list entries is divided into learning data set, correction data set and test by the present invention
Three parts of data set.
(4) it using the learning data set as the input layer of extreme learning machine model, is selected according to maximum relation degree principle
Each optimal neuron, and each optimal neuron is added to the hidden layer of the extreme learning machine model, initial hidden layer is constructed,
It specifically includes:
(41) generate hidden layer at random waits for the weight w of each neuron and deviation b in scavenger;
(42) using the learning data set as the input layer of extreme learning machine model, calculating waits for each neuron in scavenger
Output valve, and the degree of correlation of current predictive residual error corresponding with real output value is calculated, process is:
For waiting for i-th of neuron in scavenger, according to learning data set, the output valve h of the neuron is calculatedi, and
Calculate the degree of correlation of current predictive residual error corresponding with real output value:
ci=(ttest-y)Thi
Wherein, y is the predicted value of the extreme learning machine model of construction, ttestFor the real output value of data set, ciIt is i-th
The degree of correlation of a hidden layer neuron output valve and real output value.
(43) it selects and is added as optimal neuron, and by optimal neuron with each highest neuron of the output valve degree of correlation
The hidden layer currently established, and update the output layer weights and desired output of extreme learning machine model
Wherein, H is the output valve of current hidden layer, H+For the pseudoinverse of H, hoptFor the output valve of optimal neuron, hT optFor
The transposition of optimal neuron output, I are unit battle array, ttestFor the real output value of data set;
To reduce the greedy degree of algorithm, avoids omitting the higher neuron of performance, the output of existing model be done as follows
Update:
Wherein, y is the output valve of existing model,For desired output, γ is the weighted value for reducing greedy degree,
aiIt is the degree of correlation for waiting for i-th of neuron and current residue in scavenger of hidden layer, γ and aiCalculating it is as follows:
Wherein, cmaxFor the maximum relation degree for waiting for each neuron output value and desired output in scavenger of hidden layer, ciFor
The degree of correlation for waiting for i-th of neuron output value and desired output in scavenger of hidden layer, aiIt is that hidden layer waits for i-th in scavenger
The degree of correlation of a neuron and current residue, hi,normThe normalization output for waiting for scavenger neuron for i-th.
Meanwhile by the output valve h of optimal neuronoptIt is added in the output H of hidden layer.
(44) according to maximum relation degree principle, the weight w and deviation of each neuron in scavenger are waited for using genetic algorithm optimization
B is selected with the maximum each neuron of the current desired output degree of correlation as optimal neuron, and each optimal neuron is added
The hidden layer of the extreme learning machine model, constructs initial hidden layer.
(5) the extreme learning machine model established by initial hidden layer using the correction data set re -training is based on Cp
Criterion determines the neuron number of final hidden layer, to obtain initial predicted model;
Detailed process is:
Optimal hidden layer is chosen based on Cp criterion, for including the hidden layer of preceding i neuron, Cp rule definitions are such as
Under:
Wherein, i is i-th of neuron, RSSPFor assessment prediction model for the precision of prediction of correction data set, definition
For:
Wherein, j is the number of data intensive data, tjFor the voltage actual value of moment j, yjFor corresponding moment limit study
The output valve of model.
MSEfullIt is precision of prediction value of the learning data set under the initial predicted model comprising whole neurons, l is school
The parameter number of correction data;
By minimum CpThe corresponding extreme learning machine model of value is as initial predicted model.
(6) learning data set and correction data set are combined as to the input layer of the initial predicted model, passed through
Training determines the hidden layer of the initial predicted model again, obtains final prediction model;
β=(HTH)-1HTttrain
Wherein, β is the weights of output layer, and H is the output valve of hidden layer, ttrainFor the real output value of data set.
(7) test data set is input to final prediction model and carries out prediction aging of lithium battery trend.Embodiment 1
Technique effect to illustrate the invention, using a specific application example (aging data of lithium battery) to this hair
It is bright to carry out implementation verification.Experiment uses the aging data of certain lithium battery, by 4 groups of lithium batteries under 1.5A constant current source power supply environment
It charges, measures the charging voltage of lithium battery, until the voltage of battery reaches 4.2V.Multigroup charging voltage data is acquired to be used for
Prediction data.In test, 2/3 measurement data is used as to the training data of model, and 1/3 data are used as test number
According to.To weigh and comparing test performance, LevenbergMarquart algorithms, Bayesian regression algorithm, ratio conjugation ladder are utilized
Degree, ELM, CSELM and the method for the present invention predict voltage data.
Table 1 is 4 groups of lithium battery charging voltage datas acquiring respectively under existing prediction technique and prediction technique of the present invention
Operation result and model parameter.
Table 1
As can be seen from Table 1, prediction technique of the invention has precision of prediction high compared to existing prediction technique, and model is advised
The small feature of mould.
(a) in Fig. 2, (b), (c), (d) figure are respectively the effect that 4 groups of lithium battery charging voltage datas are predicted in table 1
Figure.As can be seen from Figure, the prediction model that this method is established can accurately predict actual battery aging curve.Emulation is real
Test the result shows that, method proposed in this paper more existing several typical prediction techniques on precision of prediction have higher promotion,
And tectonic model is more succinct, accurately, efficiently.
Although the illustrative specific implementation mode of the present invention is described above, in order to the technology of the art
Personnel understand the present invention, it should be apparent that the present invention is not limited to the range of specific implementation mode, to the common skill of the art
For art personnel, if various change the attached claims limit and determine the spirit and scope of the present invention in, these
Variation is it will be apparent that all utilize the innovation and creation of present inventive concept in the row of protection.
Claims (4)
1. a kind of aging of lithium battery trend forecasting method based on extreme learning machine, which is characterized in that include the following steps:
(1) original time series of lithium battery charging voltage are acquired;
(2) phase space reconfiguration is carried out to the original time series of acquisition, using optimal entropy rate method, prolonged minimum entropy rate is corresponding
The slow delay time and Embedded dimensions of time and Embedded dimensions as optimal phase space model, obtain phase space model;
(3) according to phase space Construction of A Model Volterra series models, composition sequence is obtained, and using composition sequence as input sequence
Row, the list entries are divided into learning data set, correction data set and test data set;
(4) it using the learning data set as the input layer of extreme learning machine model, is selected respectively most according to maximum relation degree principle
Excellent neuron, and each optimal neuron is added to the hidden layer of the extreme learning machine model, construct initial hidden layer;
(5) the extreme learning machine model established by initial hidden layer using the correction data set re -training is based on CpCriterion,
The neuron number for determining final hidden layer, to obtain initial predicted model;
(6) learning data set and correction data set are combined as to the input layer of the initial predicted model, by again
Training determines the hidden layer of the initial predicted model, obtains final prediction model;
(7) test data set is input to the final prediction model and carries out prediction aging of lithium battery trend.
2. the aging of lithium battery trend forecasting method according to claim 1 based on extreme learning machine, which is characterized in that institute
State in step (3) further includes being screened to the parameter in the Volterra series models of construction using the minimum angle Return Law.
3. the aging of lithium battery trend forecasting method according to claim 1 based on extreme learning machine, which is characterized in that institute
Step (4) is stated to specifically include:
(41) weights and deviation for waiting for each neuron in scavenger of hidden layer are generated at random;
(42) using the learning data set as the input layer of extreme learning machine model, the output for waiting for each neuron in scavenger is calculated
Value, and calculate the degree of correlation of current predictive residual error corresponding with real output value;
(43) it selects with each maximum neuron of the output valve degree of correlation as optimal neuron, and optimal neuron is added currently
Hidden layer in, and update the output layer weights and desired output of extreme learning machine model;
(44) according to maximum relation degree principle, the weights and deviation of each neuron in scavenger are waited for using genetic algorithm update, are selected
With the maximum each neuron of the current desired output degree of correlation as optimal neuron, and the pole is added in each optimal neuron
The hidden layer for limiting learning machine model, constructs initial hidden layer.
4. the aging of lithium battery trend forecasting method according to claim 1 based on extreme learning machine, which is characterized in that institute
State the C of step (5)pCriterion is:
Wherein, i is i-th of neuron, RSSPFor assessment prediction model for the precision of prediction of correction data set, it is defined as:
Wherein, j is the number of data intensive data, tjThe voltage actual value for being j for number, yjFor the output of limit learning model
Value;
MSEfullIt is precision of prediction value of the learning data set under the initial predicted model comprising whole neurons, l is correction data
Parameter number;
By minimum CpThe corresponding extreme learning machine model of value is as initial predicted model.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810427737.XA CN108804764A (en) | 2018-05-07 | 2018-05-07 | A kind of aging of lithium battery trend forecasting method based on extreme learning machine |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810427737.XA CN108804764A (en) | 2018-05-07 | 2018-05-07 | A kind of aging of lithium battery trend forecasting method based on extreme learning machine |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108804764A true CN108804764A (en) | 2018-11-13 |
Family
ID=64091989
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810427737.XA Pending CN108804764A (en) | 2018-05-07 | 2018-05-07 | A kind of aging of lithium battery trend forecasting method based on extreme learning machine |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108804764A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110222606A (en) * | 2019-05-24 | 2019-09-10 | 电子科技大学 | Electronic system fault forecast method based on tree search extreme learning machine |
CN110333462A (en) * | 2019-08-08 | 2019-10-15 | 首都师范大学 | A kind of lithium ion battery life-span prediction method under random discharge environment based on DGWO-ELM |
CN111950205A (en) * | 2020-08-25 | 2020-11-17 | 哈尔滨理工大学 | Lithium battery SOH prediction method based on FWA optimization extreme learning machine |
US11165270B2 (en) | 2019-03-21 | 2021-11-02 | Microsoft Technology Licensing, Llc | Predictive management of battery operation |
CN114528772A (en) * | 2022-04-20 | 2022-05-24 | 深圳市森树强电子科技有限公司 | Charger charging prediction method in electromechanical converter control system |
CN114859231A (en) * | 2022-04-27 | 2022-08-05 | 电子科技大学 | Method for predicting remaining life of battery based on wiener process and extreme learning machine |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5714866A (en) * | 1994-09-08 | 1998-02-03 | National Semiconductor Corporation | Method and apparatus for fast battery charging using neural network fuzzy logic based control |
CN103336877A (en) * | 2013-07-25 | 2013-10-02 | 哈尔滨工业大学 | Satellite lithium ion battery residual life prediction system and method based on RVM (relevance vector machine) dynamic reconfiguration |
CN103400190A (en) * | 2013-08-13 | 2013-11-20 | 浙江大学 | Integrated framework method for optimizing extremity learning machine by using genetic algorithm |
CN104978459A (en) * | 2015-07-08 | 2015-10-14 | 电子科技大学 | IGBT remaining useful life prediction method |
CN105183994A (en) * | 2015-09-10 | 2015-12-23 | 广西大学 | Method and device for predicting powder battery SOC on basis of improved I-ELM |
CN105277896A (en) * | 2015-10-26 | 2016-01-27 | 安徽理工大学 | ELM-UKF-based lithium battery remaining service life prediction method |
CN107843843A (en) * | 2017-09-30 | 2018-03-27 | 江苏理工学院 | A kind of on-vehicle battery SOC on-line prediction methods based on big data and extreme learning machine |
-
2018
- 2018-05-07 CN CN201810427737.XA patent/CN108804764A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5714866A (en) * | 1994-09-08 | 1998-02-03 | National Semiconductor Corporation | Method and apparatus for fast battery charging using neural network fuzzy logic based control |
CN103336877A (en) * | 2013-07-25 | 2013-10-02 | 哈尔滨工业大学 | Satellite lithium ion battery residual life prediction system and method based on RVM (relevance vector machine) dynamic reconfiguration |
CN103400190A (en) * | 2013-08-13 | 2013-11-20 | 浙江大学 | Integrated framework method for optimizing extremity learning machine by using genetic algorithm |
CN104978459A (en) * | 2015-07-08 | 2015-10-14 | 电子科技大学 | IGBT remaining useful life prediction method |
CN105183994A (en) * | 2015-09-10 | 2015-12-23 | 广西大学 | Method and device for predicting powder battery SOC on basis of improved I-ELM |
CN105277896A (en) * | 2015-10-26 | 2016-01-27 | 安徽理工大学 | ELM-UKF-based lithium battery remaining service life prediction method |
CN107843843A (en) * | 2017-09-30 | 2018-03-27 | 江苏理工学院 | A kind of on-vehicle battery SOC on-line prediction methods based on big data and extreme learning machine |
Non-Patent Citations (9)
Title |
---|
R. RAZAVI-FAR, S. CHAKRABARTI AND M. SAIF: "Multi-step parallel-strategy for estimating the remaining useful life of batteries", 《2017 IEEE 30TH CANADIAN CONFERENCE ON ELECTRICAL AND COMPUTER ENGINEERING (CCECE)》 * |
刘柱; 姜媛媛; 罗慧; 周利华: "基于最优权阈值ELM算法的锂离子电池RUL预测", 《电源学报》 * |
周维和: "浅析锂电池剩余寿命的ELM间接预测方法", 《中国新技术新产品》 * |
张雨浓等: "基于权值与结构确定法的单极Sigmoid神经网络分类器", 《计算机应用》 * |
张雨浓等: "权值与结构双确定法的RBF神经网络分类器", 《计算技术与自动化》 * |
李国进,董第永,陈双: "磷酸铁锂电池的SOC预测", 《计算机仿真》 * |
王枫; 史永超; 崔纳新; 李超群: "基于灰色模型的锂电池充电技术研究", 《电源学报》 * |
邱翊峰等: "短期电力负荷在小波包分解下的径向基神经网络预测方法", 《数字通信》 * |
陈婕; 金馨: "基于极限学习机的锂电池健康状态预测", 《软件》 * |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11165270B2 (en) | 2019-03-21 | 2021-11-02 | Microsoft Technology Licensing, Llc | Predictive management of battery operation |
CN110222606A (en) * | 2019-05-24 | 2019-09-10 | 电子科技大学 | Electronic system fault forecast method based on tree search extreme learning machine |
CN110222606B (en) * | 2019-05-24 | 2022-09-06 | 电子科技大学 | Early failure prediction method of electronic system based on tree search extreme learning machine |
CN110333462A (en) * | 2019-08-08 | 2019-10-15 | 首都师范大学 | A kind of lithium ion battery life-span prediction method under random discharge environment based on DGWO-ELM |
CN110333462B (en) * | 2019-08-08 | 2021-04-30 | 首都师范大学 | DGWO-ELM-based lithium ion battery life prediction method in random discharge environment |
CN111950205A (en) * | 2020-08-25 | 2020-11-17 | 哈尔滨理工大学 | Lithium battery SOH prediction method based on FWA optimization extreme learning machine |
CN114528772A (en) * | 2022-04-20 | 2022-05-24 | 深圳市森树强电子科技有限公司 | Charger charging prediction method in electromechanical converter control system |
CN114528772B (en) * | 2022-04-20 | 2022-07-01 | 深圳市森树强电子科技有限公司 | Charger charging prediction method in electromechanical converter control system |
CN114859231A (en) * | 2022-04-27 | 2022-08-05 | 电子科技大学 | Method for predicting remaining life of battery based on wiener process and extreme learning machine |
CN114859231B (en) * | 2022-04-27 | 2023-06-09 | 电子科技大学 | Battery remaining life prediction method based on wiener process and extreme learning machine |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108804764A (en) | A kind of aging of lithium battery trend forecasting method based on extreme learning machine | |
CN109991542A (en) | Lithium ion battery residual life prediction technique based on WDE optimization LSTM network | |
Schwenk et al. | Integrating battery aging in the optimization for bidirectional charging of electric vehicles | |
CN111523226B (en) | Storage battery life prediction method based on optimized multilayer residual BP (back propagation) depth network | |
Pan et al. | Transfer learning-based hybrid remaining useful life prediction for lithium-ion batteries under different stresses | |
CN103389471A (en) | Cycle life indirect prediction method for lithium ion battery provided with uncertain intervals on basis of GPR (general purpose register) | |
CN106383315A (en) | New energy automobile battery state of charge (SOC) prediction method | |
Urgun et al. | Importance sampling using multilabel radial basis classification for composite power system reliability evaluation | |
CN104182630A (en) | Residual battery capacity detection method based on simplified least square support vector machine | |
CN105183994A (en) | Method and device for predicting powder battery SOC on basis of improved I-ELM | |
KR20220073829A (en) | Battery Performance Prediction | |
CN111367349A (en) | Photovoltaic MPPT control method and system based on prediction model | |
CN105911476A (en) | Battery energy storage system SOC predication method based on data mining | |
KR20210003596A (en) | Apparatus and method for estimating aging condition of battery | |
Gong et al. | Voltage-stress-based state of charge estimation of pouch lithium-ion batteries using a long short-term memory network | |
CN106372272B (en) | Lithium battery capacity and service life prediction method based on generalized degradation model and multi-scale analysis | |
Gautam et al. | A novel algorithm for scheduling of electric vehicle using adaptive load forecasting with vehicle-to-grid integration | |
CN113917334A (en) | Battery health state estimation method based on evolution LSTM self-encoder | |
KR20210089021A (en) | Simulation system and data distribution method | |
WO2022237661A1 (en) | Lithium battery online life span prediction method based on two-point life span characteristics | |
WO2022126906A1 (en) | Method and system for optimizing battery management system model, and storage medium and electric vehicle | |
CN115621573A (en) | System and method for predicting health state of lithium battery of energy storage power station | |
CN113721151B (en) | Battery capacity estimation model and method based on double-tower deep learning network | |
Wang et al. | Neural Network-Based State of Charge Estimation Method for Lithium-ion Batteries Based on Temperature. | |
Dakshinamoorthy et al. | Estimating battery temperature in dynamic driving conditions using physics informed neural networks |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181113 |