CN102569922B - Improved storage battery SOC estimation method based on consistency of unit cell - Google Patents

Improved storage battery SOC estimation method based on consistency of unit cell Download PDF

Info

Publication number
CN102569922B
CN102569922B CN201210055701.6A CN201210055701A CN102569922B CN 102569922 B CN102569922 B CN 102569922B CN 201210055701 A CN201210055701 A CN 201210055701A CN 102569922 B CN102569922 B CN 102569922B
Authority
CN
China
Prior art keywords
soc
rule base
cell
fuzzy rule
estimate
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.)
Expired - Fee Related
Application number
CN201210055701.6A
Other languages
Chinese (zh)
Other versions
CN102569922A (en
Inventor
戴海峰
魏学哲
孙泽昌
王佳元
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.)
Shanghai pine Power Supply Technology Co., Ltd.
Original Assignee
Tongji University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tongji University filed Critical Tongji University
Priority to CN201210055701.6A priority Critical patent/CN102569922B/en
Publication of CN102569922A publication Critical patent/CN102569922A/en
Application granted granted Critical
Publication of CN102569922B publication Critical patent/CN102569922B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

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 relates to an improved storage battery SOC (State Of Charge) estimation method based on the consistency of a unit cell. In the method, the training data set of a storage battery, SOC predicted data corrected on the basis of consistency, and a self-adoption neural network are use for building a fuzzy rule base with an optimized structure and parameters; and then the fuzzy rule base after off-line learning is implanted into the fuzzy inference machine of an embedded controller of BMS (Battery Management System), so as to carry out on-line correction to storage battery SOC estimation. Compared with the prior art, the invention has the advantage that SOC differences in all unit cells inside the battery can be factually reflected on entire SOC estimation.

Description

A kind ofly based on the conforming batteries SOC of cell, estimate to improve one's methods
Technical field
The present invention relates to a kind of batteries SOC and estimate to improve one's methods, especially relate to and a kind ofly based on the conforming batteries SOC of cell, estimate to improve one's methods.
Background technology
Traditional battery management system (Battery Management System, BMS) to inside battery state, particularly to state-of-charge (State ofcharge, while SOC) estimating, often whole battery pack is regarded as to an integral body and estimated, in fact, inconsistent owing to existing between inner each monomer of battery pack, therefore, with integrality, replace each monomer to manage and have its irrationality.Such as, when the state-of-charge of inner certain batteries of battery pack is 20%, the state-of-charge of another joint is 80%, and obviously, one of this two batteries approached puts, and one approaches and overcharges, and now, management system may be 50% to the state estimation value of battery pack.Due to car load with time whole battery pack is treated as a battery, therefore, when large intensity electric discharge, 20% battery easily occurred to put, and when large intensity charging, 80% battery easily overcharges.In the battery of in the past more common portable set is applied in groups, because number of batteries is in groups fewer, applying working condition changes inviolent, so the impact that between battery, difference causes is not outstanding.But in the application of vehicle mounted dynamic battery, from tens joints to several ten thousand joints not etc., applying working condition changes very violently the quantity of stack battery in addition, is easy to occur abnormal condition as above.This situation can make shorten the useful life of whole Battery pack, and may cause cell to lose efficacy in advance, even causes potential safety hazard, causes system operation and maintenance cost to increase.This problem has become one of bottleneck problem of the development of restriction electric vehicle industrialization and application.
And be limited by the volume of automobile controller and the restriction on cost, each joint cell is all carried out to parameter identification unrealistic.Therefore, be necessary to provide a kind of SOC based on battery pack internal consistency state to estimate modification method under current BMS system architecture.
Summary of the invention
Object of the present invention is exactly to provide a kind of whole group of method in SOC estimation that the SOC difference of each cell in battery pack can be reflected in truly in order to overcome the defect of above-mentioned prior art existence, the method can judge the degree of consistency of each cell on SOC in battery pack, and the consistency based on cell is carried out batteries SOC estimation correction.
Object of the present invention can be achieved through the following technical solutions: a kind ofly based on the conforming batteries SOC of cell, estimate to improve one's methods, the method comprises that the off-line learning modeling process of the fuzzy rule base based on adaptive neural network and the online SOC based on fuzzy rule base estimate makeover process, and concrete steps are as follows:
First utilize the training dataset of batteries model, the SOC prediction data based on consistency correction and adaptive neural network to build to have and optimize structure and the fuzzy rule base of parameter;
Then the fuzzy rule base after off-line learning is implanted in the indistinct logic computer of embedded controller of BMS, batteries SOC is estimated to revise online.
Described off-line learning modeling process specifically comprises the following steps:
1) by cell test set up can reflect cell in use dynamic characteristic cell model and obtain parameter and the distributions rule of cell;
2) according to cell model, set up battery pack model;
3) adopt exciting current as the input of battery pack model, emulation obtains the input data set of fuzzy rule base, and predeterminated target output parameter is set;
4) input data set is input in fuzzy rule base;
5) calculate the poor of fuzzy rule base output valve and predeterminated target output parameter, and the study foundation using it as neural network BP training algorithm, the weighted value of revising fuzzy rule base regular node, returns to step 4) until the difference of the output of fuzzy rule base and predeterminated target output parameter is less than default threshold value;
6) off-line learning finishes.
Described fuzzy rule base comprises obfuscation layer, regular relevance grade computation layer, regular relevance grade normalization layer, defuzzification layer and output layer.
Described input data set is the characteristic quantity that data that BMS can real-time online collection form, and described data that can real-time online collection comprise the delta data of monomer battery voltage, battery voltage, battery pack current and cell SOC.
What described predeterminated target output parameter was whole group of SOC estimated value with based on the revised SOC predicted value of consistency is poor.
Described neural network BP training algorithm comprises the Hybrid Search algorithm that BP training algorithm and BP and least square method combine.
Described online SOC estimates that makeover process comprises the following steps:
1) fuzzy rule base through off-line learning modeling process is implanted in the indistinct logic computer of embedded controller of BMS;
2) parameter of embedded controller Real-time Collection actual battery group, and extract its characteristic quantity as the input of fuzzy rule base, fuzzy rule base calculates the correction that battery pack SOC estimates;
3) adopt conventional method to estimate the SOC estimated value of current actual battery group;
4) correction and SOC estimated value are added to the SOC correction predicted value that can obtain revised actual battery group.
Described conventional method is Kalman filter method, current integration method, fuzzy reasoning method or neural network.
Compared with prior art, the present invention can implement under the hardware system structure of current BMS, and can to the state estimation of whole group of SOC, revise according to the virtual condition of consistency of battery pack, the degree of consistency of each cell on SOC in judgement battery pack, thus more authentic and valid reference information provided for the degree of consistency and the variation tendency of vehicle energy management strategy, judgement battery pack.
Accompanying drawing explanation
Fig. 1 be the present invention can reflect battery dynamic characteristic cell Mathematical Modeling and consisting of battery pack model;
Fig. 2 is the constructed fuzzy rule base structural representation of the present invention;
Fig. 3 is for the fuzzy rule base modeling procedure of SOC consistency correction and training optimizing process flow chart;
Fig. 4 is that whole group of traditional SOC estimates the battery pack SOC on-line prediction flow chart combining with the correction of consistency fuzzy reasoning.
Embodiment
Below in conjunction with the drawings and specific embodiments, the present invention is described in detail.
Embodiment
A kind ofly based on the conforming batteries SOC of cell, estimate to improve one's methods, the method comprises that the off-line learning modeling process of the fuzzy rule base based on adaptive neural network and the online SOC based on fuzzy rule base estimate makeover process, and concrete steps are as follows:
First utilize the training dataset of batteries model, the SOC prediction data based on consistency correction and adaptive neural network to build to have and optimize structure and the fuzzy rule base of parameter;
Then the fuzzy rule base after off-line learning is implanted in the indistinct logic computer of embedded controller of BMS, batteries SOC is estimated to revise online.
As shown in Figure 1, for for obtaining the batteries model with characteristics of distributed parameters of training dataset.The described battery pack of model is for to be composed in series by cell model.In cell model, Uocv is the open circuit voltage of battery; Resistance R 0be used for describing battery ohmic internal resistance, R 1, C 1and R 2, C 2be used for describing the polarity effect of battery.The physical and chemical process very complex of lithium ion battery work, be described clearly also difficult, but, can generally the motion of lithium ion be divided in interelectrode transmitting procedure and two parts of diffusion process on electrode to the impedance that the obstruction being subject to while describing lithium ion activity by two RC links respectively in circuit model or battery table reveal.With the less R of time constant 1c 1link is described the impedance being subject to when lithium ion transmits between electrode, with the larger R of time constant 2c 2the impedance that link is subject to while describing the diffusion of lithium ion in electrode material, because the latter's process is more complicated, except the diffusion of lithium ion in solid phase electrode material, the crystal structure of electrode material also can change due to the accumulation gradually of lithium ion quantity, therefore the suffered hindering factor of the diffusion of lithium ion in electrode is more, also slow many of diffusion in electrolyte compared with it.The SOC-OCV that considers battery only can bring with an electric capacity replacement nonlinearity erron can not put up with, and therefore, in Fig. 1, OCV replaces with a nonlinear function, and it is the function of SOC parameter.The nonlinear equivalent circuit model of this SOC of comprising is comprised of linear segment and non-linear partial, and has passed through the checking that operating mode discharges and recharges and electrochemical impedance spectroscopy is tested, and shows that it can reflect battery dynamic characteristic in the course of the work preferably.
As shown in Figure 2, be the universal architecture of constructed fuzzy rule base, include five layers: obfuscation layer, regular relevance grade computation layer, regular relevance grade normalization layer, defuzzification layer and output layer.The function of each layer is as follows respectively: ground floor: the fuzzy membership function that calculates input
O 1，i＝g xi(x，a i，b i)；O 1，j＝g y(i-2)(y，c j-2，d j-2)
X in above formula, y is node input, g () is degree of membership member function, can be used to calculate the degree of membership of current each input.A wherein i, b i, c i, d ibe called former piece parameter, can be used to regulate the shape of degree of membership member function.
The second layer: computation rule relevance grade
O 2,1=O 1,1* O 1,3=g x1(x, a 1, b 1) * g y1(y, c 1, d 1) be denoted as w 1
O 2,2=O 1,2* O isosorbide-5-Nitrae=g x2(x, a 2, b 2) * g y2(y, c 2, d 2) be denoted as w 2
This one deck major function is by the output multiplication of ground floor, and by its product with w 1and w 2output.
The 3rd layer: calculate relevance grade normalization
O 3,1 = w 1 w 1 + w 2 = w 1 ‾
O 3,2 = w 2 w 1 + w 2 = w 2 ‾
This one deck is for calculating the w of i rule iratio with whole regular w value sums, triggers intensity level.
The 4th layer: computation rule output
z i = O 4 , i = w ‾ i f i = w ‾ i ( p i x + q i y + r i ) ; i＝1，2
In formula, p i, q i, r ifor the parameter of this node, be commonly called consequent parameter.In this layer, the triggering intensity level of each node and a single order multinomial multiply each other
Layer 5: computing system output
z = w 1 ‾ z 1 + w 2 ‾ z 2
The 4th layer of all signal z of output layer accumulation calculating itotal output, and finally form single output.
In implementation process, the input variable shown in Fig. 2 is actual is by monomer battery voltage is vectorial, battery voltage is vectorial, battery pack current is vectorial and battery pack SOC estimated value vector extracts the rear a plurality of input parameters that generate via characteristic index.
As shown in Figure 3, be the step of off-line learning modeling process.First, build battery pack model as shown in Figure 1.In battery pack, the parameter of each cell can be by real target battery, a batch test obtains; Also can by test, obtain the parameter distribution of a collection of battery, then according to this distribution probability, produce at random.A kind of rear method is compared with the former can be so that data have generality more.Then, current excitation I is applied to and on battery pack model, carries out emulation.The operating mode of current excitation I can gather from real vehicle, also can obtain by the vehicle dynamical system model emulation in the development phase.Due to, in battery pack, the parameter of all monomers is known in simulation model, therefore, can obtain the terminal voltage of each cell under current excitation and the delta data of SOC.The reference correction SOCt of SOC can obtain from the SOC information of each cell.Below a kind of SOC of obtaining with reference to the method for correction:
SOC t(t)＝W adj(t)×SOC a(t)+(1-W adj(t))×SOC pack(t)-SOC pack(t)
SOC a(t)＝max(SOC cell(t))(W adj＞0)
SOC a(t)＝min(SOC cell(t))(W adj＜0)
W adj = W SOC ( t ) + ( 1 - abs ( W soc ( t ) ) × dQ ( t ) mQ
dQ(t)＝SUM(Curr(t-windows)：Curr(t))
The revised battery pack SOC target of SOCt-
The battery pack SOC estimated value that SOCpack-is traditional
Wadj-adjusts weight
Conventional batteries SOC estimated value after Wsoc-standardization
SOCa-adjusts and uses with reference to battery core SOC
Windows-is electric quantity accumulation time window in short-term
MQ-is electric standard value in short-term
The method has been considered the impact of size of current and current SOC operation interval.Can be with reference to the method during concrete enforcement, but be not limited only to use the method.Again according to the structure of the structure fuzzy rule base of Fig. 2, here can be by monomer battery voltage extreme difference, current battery pack SOC estimated value, the BMS such as current excitation can online acquisition or the parameter normalized of estimation after as the input of rule base.During concrete enforcement, also can be not limited to above-mentioned these collectable physical quantitys.Finally, by the method for iteration, rule base is trained to optimization, the revise goal SOCtp of the output SOCt of rule base and reference is compared, definition error function:
E = 1 2 ( SOC t ( t ) - SOC tp ( t ) ) 2
SOCt-model output SOC correction value
The preset SOC correction value of SOCtp-training data
Foundation using E as BP back propagation learning, the weight of each node in iteration modification rule storehouse, until E is less than a certain default termination condition position.So just completed the training optimizing process under a certain operating mode.
During concrete enforcement, in order making to optimize, to train the generalization ability of rear rule base better, the data of different initial condition and different current excitation operating modes can be combined and to train.Also can use the hybrid learning algorithm in conjunction with least square to obtain convergence rate and better convergence result more fast.
As shown in Figure 4, for online SOC estimates makeover process scheme.The SOC of whole Battery pack estimates to realize by Kalman filter, also can obtain by current integration or additive method.The SOC estimated value of whole Battery pack and the monomer battery voltage of BMS Real-time Collection, battery voltage, battery pack current is all using the input as consistency indistinct logic computer.Fuzzy rule base adopts the training of method shown in Fig. 3 to obtain.By indistinct logic computer, obtain based on current conforming SOC correction, then obtain final correction value with the estimated value addition of whole Battery pack SOC.
As mentioned above, the present invention has formed a complete process steps from modeling to application.Due to, it is final that what generate is the model that can be used for indistinct logic computer, therefore, the present invention can apply with current BMS hardware in, and for solve the inconsistent SOC of the causing estimated bias of cell provide a kind of improve one's methods aforementioned.

Claims (6)

1. based on the conforming batteries SOC of cell, estimate to improve one's methods for one kind, it is characterized in that, the method comprises that the off-line learning modeling process of the fuzzy rule base based on adaptive neural network and the online SOC based on fuzzy rule base estimate makeover process, and concrete steps are as follows:
First utilize the training dataset of batteries model, the SOC prediction data based on consistency correction and adaptive neural network to build to have and optimize structure and the fuzzy rule base of parameter;
Then the fuzzy rule base after off-line learning is implanted in the indistinct logic computer of embedded controller of BMS, batteries SOC is estimated to revise online;
Described off-line learning modeling process specifically comprises the following steps:
1) by cell test set up can reflect cell in use dynamic characteristic cell model and obtain parameter and the distributions rule of cell;
2) according to cell model, set up battery pack model;
3) adopt exciting current as the input of battery pack model, emulation obtains the input data set of fuzzy rule base, and predeterminated target output parameter is set;
4) input data set is input in fuzzy rule base;
5) calculate the poor of fuzzy rule base output valve and predeterminated target output parameter, and the study foundation using it as neural network BP training algorithm, the weighted value of revising fuzzy rule base regular node, returns to step 4) until the difference of the output of fuzzy rule base and predeterminated target output parameter is less than default threshold value;
6) off-line learning finishes;
Described online SOC estimates that makeover process comprises the following steps:
1) fuzzy rule base through off-line learning modeling process is implanted in the indistinct logic computer of embedded controller of BMS;
2) parameter of embedded controller Real-time Collection actual battery group, and extract its characteristic quantity as the input of fuzzy rule base, fuzzy rule base calculates the correction that battery pack SOC estimates;
3) adopt conventional method to estimate the SOC estimated value of current actual battery group;
4) correction and SOC estimated value are added to the SOC correction predicted value that can obtain revised actual battery group.
2. according to claim 1ly a kind ofly based on the conforming batteries SOC of cell, estimate to improve one's methods, it is characterized in that, described fuzzy rule base comprises obfuscation layer, regular relevance grade computation layer, regular relevance grade normalization layer, defuzzification layer and output layer.
3. according to claim 1ly a kind ofly based on the conforming batteries SOC of cell, estimate to improve one's methods, it is characterized in that, described input data set is the characteristic quantity that data that BMS can real-time online collection form, and described data that can real-time online collection comprise the delta data of monomer battery voltage, battery voltage, battery pack current and cell SOC.
4. according to claim 1ly a kind ofly based on the conforming batteries SOC of cell, estimate to improve one's methods, it is characterized in that, what described predeterminated target output parameter was whole group of SOC estimated value with based on the revised SOC predicted value of consistency is poor.
5. according to claim 1ly a kind ofly based on the conforming batteries SOC of cell, estimate to improve one's methods, it is characterized in that, described neural network BP training algorithm comprises the Hybrid Search algorithm that BP training algorithm and BP and least square method combine.
6. according to claim 1ly a kind ofly based on the conforming batteries SOC of cell, estimate to improve one's methods, it is characterized in that, described conventional method is Kalman filter method, current integration method, fuzzy reasoning method or neural network.
CN201210055701.6A 2012-03-05 2012-03-05 Improved storage battery SOC estimation method based on consistency of unit cell Expired - Fee Related CN102569922B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201210055701.6A CN102569922B (en) 2012-03-05 2012-03-05 Improved storage battery SOC estimation method based on consistency of unit cell

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201210055701.6A CN102569922B (en) 2012-03-05 2012-03-05 Improved storage battery SOC estimation method based on consistency of unit cell

Publications (2)

Publication Number Publication Date
CN102569922A CN102569922A (en) 2012-07-11
CN102569922B true CN102569922B (en) 2014-03-05

Family

ID=46414776

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201210055701.6A Expired - Fee Related CN102569922B (en) 2012-03-05 2012-03-05 Improved storage battery SOC estimation method based on consistency of unit cell

Country Status (1)

Country Link
CN (1) CN102569922B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI563276B (en) * 2014-07-21 2016-12-21 Richtek Technology Corp Method of estimating the state of charge of a battery and system thereof

Families Citing this family (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103544330B (en) * 2012-07-13 2016-08-24 北汽福田汽车股份有限公司 The equivalent-circuit model construction method of lithium ion battery
CN103323781B (en) * 2013-05-29 2015-08-05 西安交通大学 Power battery pack on-line parameter detection system and SOC method of estimation
CN103344922B (en) * 2013-07-12 2015-10-28 清华大学 Mixed electric car battery monomer state-of-charge difference detecting method
CN103413981B (en) * 2013-07-24 2015-05-20 清华大学 method and apparatus for battery pack capacity
CN104835988B (en) * 2014-12-16 2017-06-06 北京车和家信息技术有限责任公司 The modification method of battery system and battery system state-of-charge SOC
DE102016119253A1 (en) * 2015-10-20 2017-04-20 GM Global Technology Operations LLC BATTERY STATE ESTIMATING SYSTEMS AND METHODS USING A NONLINEAR RESISTANCE ELEMENT
CN105807231B (en) * 2016-03-14 2018-10-19 深圳供电局有限公司 Method and system for detecting residual capacity of storage battery
CN105912799B (en) * 2016-04-27 2018-12-14 华中科技大学 A kind of modeling method of liquid or semi-liquid metal battery
CN106021923A (en) * 2016-05-19 2016-10-12 江苏理工学院 Method and system for predicting state of charge of power battery of pure electric vehicle
CN106125007A (en) * 2016-08-31 2016-11-16 北京新能源汽车股份有限公司 Method and device for determining residual electric quantity of battery and automobile
US11084387B2 (en) 2019-02-25 2021-08-10 Toyota Research Institute, Inc. Systems, methods, and storage media for arranging a plurality of cells in a vehicle battery pack
CN110118936A (en) * 2019-05-06 2019-08-13 杭州电子科技大学 A kind of estimation method of battery dump energy based on EHF
CN110146823A (en) * 2019-06-06 2019-08-20 重庆大学 A kind of series-connected cell group SOC estimation method based on information fusion
CN110601174B (en) * 2019-07-06 2023-04-18 天津大学 Load modeling and online correction method based on deep learning
CN111274713B (en) * 2020-03-09 2022-08-09 西南交通大学 Method for controlling consistency of remaining service life of multi-pile fuel cell system of motor train unit
CN113554200B (en) * 2020-04-23 2024-02-27 广州汽车集团股份有限公司 Method, system and equipment for predicting voltage inconsistency of power battery
CN112561298B (en) * 2020-12-09 2023-10-13 深圳市优必选科技股份有限公司 Robot battery management method, apparatus, control device and storage medium
CN113156316B (en) * 2021-04-26 2023-11-07 太仓中科赛诺新能源科技有限公司 Brine battery SOC estimation algorithm

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100859688B1 (en) * 2006-10-12 2008-09-23 삼성에스디아이 주식회사 Battery management system and driving method thereof

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI563276B (en) * 2014-07-21 2016-12-21 Richtek Technology Corp Method of estimating the state of charge of a battery and system thereof

Also Published As

Publication number Publication date
CN102569922A (en) 2012-07-11

Similar Documents

Publication Publication Date Title
CN102569922B (en) Improved storage battery SOC estimation method based on consistency of unit cell
Wang et al. Improved anti-noise adaptive long short-term memory neural network modeling for the robust remaining useful life prediction of lithium-ion batteries
Hannan et al. Neural network approach for estimating state of charge of lithium-ion battery using backtracking search algorithm
Liu et al. An extended Kalman filter based data-driven method for state of charge estimation of Li-ion batteries
Guo et al. A comparative study of different deep learning algorithms for lithium-ion batteries on state-of-charge estimation
CN103926538B (en) Change exponent number RC equivalent-circuit model based on AIC criterion and implementation method
CN105974327A (en) Lithium battery pack SOC prediction method based on neural network and UKF
Liu et al. A novel remaining useful life prediction method for lithium-ion battery based on long short-term memory network optimized by improved sparrow search algorithm
CN106909716A (en) The ferric phosphate lithium cell modeling of meter and capacity loss and SOC methods of estimation
CN105911476B (en) A kind of battery energy storage system SOC prediction techniques based on data mining
Mamo et al. Long short-term memory with attention mechanism for state of charge estimation of lithium-ion batteries
CN105183994A (en) Method and device for predicting powder battery SOC on basis of improved I-ELM
CN106777786A (en) A kind of lithium ion battery SOC estimation method
Boujoudar et al. Lithium-ion batteries modeling and state of charge estimation using artificial neural network
Zhang et al. Intelligent state of charge estimation of battery pack based on particle swarm optimization algorithm improved radical basis function neural network
CN106383315A (en) New energy automobile battery state of charge (SOC) prediction method
CN103954914A (en) Lithium ion battery remaining life direct prediction method based on probability integration
Liu et al. A method for state of charge and state of health estimation of lithium-ion battery based on adaptive unscented Kalman filter
Li et al. A novel state estimation approach based on adaptive unscented Kalman filter for electric vehicles
Xu et al. State of charge estimation for liquid metal battery based on an improved sliding mode observer
AU2021101964A4 (en) Artificial intelligence based smart electric vehicle battery management system
Huang et al. State of charge estimation of li-ion batteries based on the noise-adaptive interacting multiple model
Yang et al. Modified extended Kalman filtering algorithm for precise voltage and state-of-charge estimations of rechargeable batteries
CN114217234B (en) IDE-ASRCKF-based lithium ion battery parameter identification and SOC estimation method
Timilsina et al. Degradation abatement in hybrid electric vehicles using data-driven technique

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
TR01 Transfer of patent right
TR01 Transfer of patent right

Effective date of registration: 20170525

Address after: 201805, Shanghai, Anting Town Park Road, No. 3, 268, 1, A, 2, Jiading District

Patentee after: Shanghai pine Power Supply Technology Co., Ltd.

Address before: 200092 Shanghai City, Yangpu District Siping Road No. 1239

Patentee before: Tongji University

CF01 Termination of patent right due to non-payment of annual fee
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20140305

Termination date: 20210305