CN107741568A - A kind of lithium battery SOC estimation method that optimization RBF neural is shifted based on state - Google Patents
A kind of lithium battery SOC estimation method that optimization RBF neural is shifted based on state Download PDFInfo
- Publication number
- CN107741568A CN107741568A CN201711093872.7A CN201711093872A CN107741568A CN 107741568 A CN107741568 A CN 107741568A CN 201711093872 A CN201711093872 A CN 201711093872A CN 107741568 A CN107741568 A CN 107741568A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- value
- soc
- rbf
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/367—Software therefor, e.g. for battery testing using modelling or look-up tables
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Evolutionary Computation (AREA)
- Pure & Applied Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Operations Research (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Secondary Cells (AREA)
Abstract
A kind of lithium battery SOC estimation method that optimization RBF neural is shifted based on state.The invention discloses a kind of lithium battery SOC estimation method that optimization RBF neural is shifted based on state, it is related to electric vehicle engineering field.Wherein, this method includes:(1) off-line training sample data is gathered, all training samples are normalized;(2) the lithium battery SOC appraising models based on RBF neural are established;(3) RBF neural network model established is optimized using STA optimized algorithms;(4) the RBF network structures trained and each parameter value are preserved, the RBF networks trained are used for ferric phosphate lithium cell SOC estimation;The present invention can accurately estimate lithium battery SOC, have the characteristics that estimation precision is high, highly reliable, appraising model is simple, can be widely applied to electric automobile power battery technical field.
Description
Art
The present invention relates to electric automobile power battery technical field, in particular it relates to a kind of based on state transfer optimization RBF
The lithium battery SOC estimation method of neutral net.
Background technology
With global environment deterioration and shortage of resources problem it is serious, particularly in recent years domestic PM2.5 pollutions
Aggravation, there occurs huge change, Jiao for developing into everybody growing interest of new-energy automobile for domestic and international orthodox car industry
Point;In numerous alternative energy sources of automobile, electric energy is allowed using electrokinetic cell as main power source with the characteristics of its is safe efficient, cleaning
Or the new-energy automobile of auxiliary power source turns into the main object that everybody studies;Battery management system (BMS) is based on optimization
Management and the control system of protection electrokinetic cell, while be also working condition, the pipe of performance for assessing the vehicle-mounted battery of electric automobile
Reason system, support vehicles safety traffic, maintain battery-efficient utilization and stability;The core of battery management system is battery
SOC (State Of Charge, i.e. battery charge state) accurate estimation;And nonlinear system of the battery as complexity, it is difficult to
Find the mathematical relationship of the parameter such as battery SOC and voltage, electric current, temperature, internal resistance;According to electric currents of the SOC of battery with battery, electricity
The parameters such as pressure, temperature, degree of aging are presented nonlinear feature, how using battery can survey supplemental characteristic to realize present battery
Accurately estimation is all the time the key problem of battery management system and is badly in need of the technical barrier solved, Ye Shiyi dump energy
It is important and rich in the task of challenge.
Currently used SOC estimation method has:Open circuit voltage method, Impedance Analysis, Ah counting method, neural network and
Kalman filtering method etc.;The shortcomings that open circuit voltage method is notable needs battery to stand for a long time when being measurement, it usually needs several to arrive
More than ten hour, this method was only applicable to parking electric automobile state verification SOC;Impedance Analysis is by studying cell resistance
Battery SOC is estimated with SOC relation;But SOC is carried out using the battery impedance to estimate difficulty precision that is also very big, and estimating
It can not be guaranteed;Ah counting method is the more SOC estimation method of actual use, and Ah counting method occurs in the application
Some problems;Error in current measurement, SOC can be caused to produce calculation error, error constantly accumulates, increasing, and error exists
Can be bigger in the case of high temperature and current fluctuation;SOC is carried out to battery using Kalman filtering method (KF, Kalman Filter)
During estimation, battery is taken as a dynamical system research, and SOC is treated as an internal state of system;Kalman filtering algorithm
Following two problems are primarily present for SOC estimations:First, stronger to battery model dependence, to obtain accurate SOC, it is necessary to
Accurate battery model is established, and the order of accuarcy of battery model and complexity are directly proportional;Second, Kalman Algorithm
It is larger comprising a large amount of matrix operations, amount of calculation.
The content of the invention
It is an object of the invention to provide a kind of lithium battery SOC estimation method that optimization RBF neural is shifted based on state;
The present invention solves its concrete technical problems and adopted the technical scheme that:
A kind of lithium battery SOC estimation method that optimization RBF neural is shifted based on state, its step are as follows:
(A) off-line training sample data is obtained, sample data includes charge-discharge magnification at intervals of 0.2C, temperature interval 5
Monomer terminal voltage, charging and discharging currents, lug temperature and the cycle life parameter of degree Celsius lower lithium battery and corresponding SOC numbers
According to, the input layer vector using monomer terminal voltage, charging and discharging currents, lug temperature and cycle life parameter as network, SOC conducts
The output layer vector of network;According to following formula, all training samples are normalized:
Wherein, R be actual sample actual value, R* be normalized after data, RmaxFor corresponding types data sample
Maximum, RminFor the minimum value of corresponding types data sample.
(B) the lithium battery SOC appraising models based on RBF neural are established:
First, basic function of the Gaussian function for RBF neural hidden layer node is selected, its expression is as follows:
Wherein, x is 4 dimension input vectors, correspond to respectively the monomer terminal voltage of lithium battery, charging and discharging currents, lug temperature and
Cycle life parameter, ciIt is the center vector of i-th of neuron node of hidden layer, dimension is identical with input vector x, σi 2For i-th
The center width of individual neuron node.
Secondly according to training sample, the input number using the discharge and recharge data of any time lithium battery as RBF neural
According to the output data using SOC data corresponding to the moment as RBF neural, then the model is corresponding exports expression formula such as
Under:
WhereinExported for network, i.e. SOC estimated values, wiRepresent connection of i-th of neuron node to output node
Weights, k are network node in hidden layer.
(C) RBF neural network model established is optimized using STA optimized algorithms:
First, RBF network inputs sample sets are classified with K-means clustering algorithms, reduce node in hidden layer,
Determination for the number k of classification, determined using apart from cost principle:
lk1For average infima species border distance, be defined as all cluster centres to its nearest cluster centre apart from sum
Average value, ciFor the center of ith cluster, cjFor the center of j-th of cluster;
lk2For average inter- object distance, the average value of all clustering cluster inner distance sums is defined as, n is total sample number, mi
For the total sample number of ith cluster, xijFor j-th of sample of intra-cluster.
lk=lk1-lk2 (6)
lkWhen for classification number being k apart from cost, then optimal classification number k should cause in maximum classification number kmaxIt is interior,
lkFor maximum, i.e.,:
lk=max (li), i=1,2 ... kmax (7)
Meanwhile for the classification number k of determination, when being clustered with K-means algorithms, to overcome K-means to calculate
The shortcomings that method is easily trapped into local optimum when being clustered, using STA algorithm optimize K-means algorithms in cluster process it is every
The selection of the central point of secondary iteration, corresponding optimization problem can be equivalent to below equation:
Wherein Ck(i+1)Represent a state, specially k center position, lk(i+1)It is corresponding Ck(i+1)Under state away from
From cost, specific formula for calculation such as formula (4) to formula (6);AiCan be considered as that the operation of optimized algorithm is calculated for state-transition matrix
Son;After STA algorithm optimizes, you can obtain classification number as the optimal classification scheme in the case of k.
Secondly, the central point c of RBF networks is determined using STA algorithmi, extension width σi 2And wi;RBF nets are optimized using STA
Network parameter problem is represented by following relation:
Wherein, SkRepresent a state, map network center ci, extension width σi 2And connection weight wiOne group of solution, Ak
It is state-transition matrix, for the operation operator of optimized algorithm, SOCiFor the SOC value of i-th of sample,For estimating for i-th sample
Calculation value, n are total number of samples, err (xk+1) object function, the mean square error of SOC actual values and estimated value is defined as, is used
STA training RBF networks comprise the following steps that:
(a) population at individual number SE=30, the random equality initialization c in feasible zone are initializedi、σi 2And wiThree variables,
Initial population is produced, produces SE group initial feasible solutions;
(b) one group of c for making object function f reach minimum value is selected from current populationi、σi 2With w values, best is designated as, it is right
The cost answered is fbest, and best is copied as into the colony that number of individuals is SE, is designated as S (k), and carrying out stretching by formula (10) obtains
To new population:
S (k+1)=S (k)+γ ReS(k) (10)
Wherein, γ is normal number, referred to as contraction-expansion factor, value 1, Re∈Rn×nFor a random diagonal matrix, x (k+1) is
New populations of the S (k) after stretching.
The optimum individual in population after stretching is newbest, and corresponding cost is gbest, if gbest
Less than fbest, then translation transformation is carried out to individual newbest by formula (11), and updates the best after translation transformation and fbest,
Otherwise without translation transformation.
Wherein, β is normal number, referred to as shift factor, value 1, Rm∈Rn×nFor a stochastic variable, its value
In the range of [0,1], S (k-1) is that newbest individuals carry out the value before stretching;
(c) best being copied as into the colony that number of individuals is SE, right back-pushed-type (12) carries out rotation transformation and obtains new population,
Optimum individual newbest after selection rotation transformation in population, corresponding cost is gbest;
Wherein, α is normal number, referred to as twiddle factor, value 1, R ∈ Rn×nRandom matrix, its element value [- 1,
1] scope, | | | |2For 2 norms of vector, In is unit matrix;
If gbest is less than fbest, by formula (11) carry out translation transformation, and update the best after translation transformation and
Fbest, otherwise without translation transformation;
(d) best is copied as into the colony that number of individuals is SE, right back-pushed-type (13) carries out principal axis transformation and judged, selected
Individual optimal in population is designated as newbest after conversion, and corresponding cost is gbest;
Sk+1=Sk+δRaSk (13)
Wherein, δ is the axle factor, value 1, RaFor random diagonal matrix;
If gbest is less than fbest, by formula (11) carry out translation transformation, and update the best after translation transformation and
Fbest, otherwise without translation transformation;
(e) repeat step b) to step d), until fitness meets minimum requirement or reaches iterations;
(f) the kernel function center c that preservation optimizes to obtain through STAi, extension width σi 2And connection weight wi;(D) instruction is preserved
The RBF network structures perfected and each parameter value, the RBF networks trained to be used for ferric phosphate lithium cell SOC estimation.
The present invention is estimated electric automobile lithium battery SOC based on the angle of data, is gone out from the external behavior of lithium battery
Hair, establishes RBF neural network model, accurate to estimate lithium battery SOC;When considering lithium battery external characteristic, electricity is not only allowed for
The influence of cell voltage, electric current, temperature to battery SOC, also introduces battery cycle life parameter, eliminates due to cell degradation pair
The error that battery SOC estimation is brought, at the same it is high representative to have training sample set, and charge-discharge magnification interval is set as
0.2C, temperature interval are set as 5 degrees Celsius;Network hidden layer number is determined with apart from cost principle, introduces average infima species
Border distance and average inter- object distance, for the center point set after cluster, more discrete explanation Clustering Effect is better, i.e. some central point
It is the bigger the better with a distance from the central point nearest from it, average infima species border distance definition is all cluster centres to nearest with it
The average value apart from sum of cluster centre;For some specific class, the sample point in class is from nearlyer theory with a distance from class center
The effect of bright cluster is better, the average value of all clustering cluster inner distance sums of average inter- object distance;The introducing of the two indexs
The degree of accuracy of classification can be improved;Meanwhile when being clustered to sample set, to overcome K-means clustering algorithms to be easily trapped into
The shortcomings that local optimum, introduce state transfer global optimization approach and global optimization is carried out to K-means;Adoption status transfer simultaneously
Algorithm optimization network parameter, obtain optimal central point ci, extension width σi 2And wi。
Brief description of the drawings
Fig. 1 is a kind of flow for the lithium battery SOC estimation method that optimization RBF neural is shifted based on state of the present invention
Figure;
Fig. 2 is a kind of SOC appraising models based on RBF neural of the present invention;
Fig. 3 is a kind of 1C electric discharges for the lithium battery SOC estimation method that optimization RBF neural is shifted based on state of the present invention
At 25 DEG C, SOC estimation result comparison diagrams;
Fig. 4 is a kind of 1C electric discharges for the lithium battery SOC estimation method that optimization RBF neural is shifted based on state of the present invention
At 25 DEG C, SOC estimation error comparison diagrams.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, below in conjunction with accompanying drawing and specific implementation
Mode is described in further detail to the present invention;But it is understood that protection scope of the present invention is not limited by embodiment
System;
The present embodiment uses hundred million latitude lithium energy LF56K-56AH as object, according to embodiments of the present invention, there is provided one kind is based on
Lithium battery SOC real-time estimating methods Fig. 1 of neutral net is the flow chart of this method, and it is comprised the following steps that:
(A) off-line training sample data is gathered, sample data includes charge-discharge magnification at intervals of 0.2C, temperature interval 5
Monomer terminal voltage, charging and discharging currents, lug temperature and the cycle life parameter of degree Celsius lower lithium battery and corresponding SOC numbers
According to;There is high representativeness for training sample selected by guarantee, in 5 kinds of discharge-rates (being set as 0.2C to 1C), 7 kinds of temperature (settings
For 0 DEG C to 30 DEG C) discharge test is carried out totally in the case of 35 kinds to obtain training sample set, the sample size of each case is 100
Individual, then total training sample size is 3500;According to following formula, training sample is normalized:
Wherein, R be actual sample actual value, R* be normalized after data, RmaxFor corresponding types data sample
Maximum, RminFor the minimum value of corresponding types data sample;
(B) reference picture 2, the lithium battery SOC appraising models based on RBF neural are established, respectively with the list of electrokinetic cell
The input vector of body end voltage (V), electric current (I), environment temperature (T) and cycle life parameter (Cn) as RBF networks, with charged
State SOC is as output vector;It is specially:
First, basic function of the Gaussian function for RBF neural hidden layer node is selected, its expression is as follows:
Wherein, x is 4 dimension input vectors, correspond to respectively the monomer terminal voltage of lithium battery, charging and discharging currents, lug temperature and
Cycle life parameter, ciIt is the center vector of i-th of neuron node of hidden layer, dimension is identical with input vector x, σi 2For i-th
The center width of individual neuron node.
Secondly according to training sample, the input number using the discharge and recharge data of any time lithium battery as RBF neural
According to the output data using SOC data corresponding to the moment as RBF neural, then the model is corresponding exports expression formula such as
Under:
WhereinExported for network, i.e. SOC estimated values, k is network node in hidden layer, wiRepresent i-th of neuron
Connection weight of the node to output node.
(C) RBF neural network model established is optimized using STA optimized algorithms:
First, network hidden layer node is determined using K-means algorithms;It is defeated to RBF networks with K-means clustering algorithms
Enter sample set to be classified, reduce node in hidden layer, the determination for the number k of classification, using apart from Least-cost principle
It is determined that:
lk1For infima species border average distance, be defined as all cluster centres to its nearest cluster centre apart from sum
Average value, ciFor the center of ith cluster, cjFor the center of j-th of cluster;
lk2For average inter- object distance, the average value of all clustering cluster inner distance sums is defined as, n is total sample number, mi
For the total sample number of ith cluster, xijFor j-th of sample of intra-cluster;
lk=lk1-lk2 (6)
lkWhen for classification number being k apart from cost, then optimal classification number k should cause in maximum classification number kmaxIt is interior,
lkFor maximum, i.e.,:
lk=max (li), i=1,2 ... kmax (7)
Meanwhile for the classification number k of determination, when being clustered with K-means algorithms, to overcome K-means to calculate
The shortcomings that method is easily trapped into local optimum when being clustered, using STA algorithm optimize K-means algorithms in cluster process it is every
The selection of the central point of secondary iteration, corresponding optimization problem can be equivalent to below equation:
Wherein Ck(i+1)Represent a state, specially k center position, lk(i+1)It is corresponding Ck(i+1)Under state away from
From cost, specific formula for calculation such as formula (4) to formula (6);AiCan be considered as that the operation of optimized algorithm is calculated for state-transition matrix
Son;After STA algorithm optimizes, you can classification number is obtained as the optimal classification scheme in the case of k, in this example, it is determined that most
Excellent central point number is 31.
Secondly, the central point c of RBF networks is determined using STA algorithmi, extension width σi 2And wi;RBF nets are optimized using STA
Network parameter problem is represented by following relation:
Wherein, Sk represents a state, map network center ci, extension width σi 2And connection weight wiOne group of solution, Ak
It is state-transition matrix, for the operation operator of optimized algorithm, SOCiFor the SOC value of i-th of sample,For estimating for i-th sample
Calculation value, n are total number of samples, err (xk+1) object function, the mean square error of SOC actual values and estimated value is defined as, is used
STA training RBF networks comprise the following steps that:
(a) population at individual number SE=30, the random equality initialization c in feasible zone are initializedi、σi 2And wiThree variables,
Initial population is produced, produces SE group initial feasible solutions;
(b) one group of c for making object function f reach minimum value is selected from current populationi、σi 2With w values, best is designated as, it is right
The cost answered is fbest, and best is copied as into the colony that number of individuals is SE, is designated as S (k), and carrying out stretching by formula (10) obtains
To new population:
S (k+1)=S (k)+γ ReS(k) (10)
Wherein, γ is normal number, referred to as contraction-expansion factor, value 1, Re∈Rn×nFor a random diagonal matrix, x (k+1) is
New populations of the S (k) after stretching;
The optimum individual in population after stretching is newbest, and corresponding cost is gbest, if gbest
Less than fbest, then translation transformation is carried out to individual newbest by formula (11), and updates the best after translation transformation and fbest,
Otherwise without translation transformation;
Wherein, β is normal number, referred to as shift factor, value 1, Rm∈Rn×nFor a stochastic variable, its value is in [0,1]
In the range of, S (k-1) is that newbest individuals carry out the value before stretching;
(c) best being copied as into the colony that number of individuals is SE, right back-pushed-type (12) carries out rotation transformation and obtains new population,
Optimum individual newbest after selection rotation transformation in population, corresponding cost is gbest;
Wherein, α is normal number, referred to as twiddle factor, value 1, R ∈ Rn×nRandom matrix, its element value [- 1,
1] scope, | | | |2For 2 norms of vector, InIt is unit matrix;
If gbest is less than fbest, by formula (11) carry out translation transformation, and update the best after translation transformation and
Fbest, otherwise without translation transformation;
(d) best is copied as into the colony that number of individuals is SE, right back-pushed-type (13) carries out principal axis transformation and judged, selected
Individual optimal in population is designated as newbest after conversion, and corresponding cost is gbest;
Sk+1=Sk+δRaSk (13)
Wherein, δ is the axle factor, value 1, RaFor random diagonal matrix;
If gbest is less than fbest, by formula (11) carry out translation transformation, and update the best after translation transformation and
Fbest, otherwise without translation transformation;
(e) repeat step b) to step d), until fitness meets minimum requirement or reaches iterations;
(f) the kernel function center c that preservation optimizes to obtain through STAi, extension width σi 2And connection weight wi;(D) warp is preserved
The RBF network structures and each parameter value that step (3) trains, at normal temperatures, invention proposition is respectively adopted
Method, BP neural network method and Ah counting method estimated the state-of-charge of electrokinetic cell at 25 DEG C of 1C discharge-rates,
Estimation result is as shown in Figure 3;SOC estimation errors as shown in figure 4,;From figure 3, it can be seen that the estimation result of algorithm is put forward herein
Closer to actual value;As can be drawn from Figure 4, the estimation error of STA-K ' M-RBF neural network algorithms is in 2% or so, BP nerves
The estimation error of network algorithm is 5% or so, the estimation error 7% or so of Ah counting method.
Above is the preferable implementation to the present invention is illustrated, but the invention is not limited in the embodiment,
Those skilled in the art can also make a variety of equivalent variations or replacement on the premise of without prejudice to spirit of the invention, this
A little equivalent or replacement is all contained in the application claim limited range.
Claims (1)
1. a kind of lithium battery SOC estimation method that optimization RBF neural is shifted based on state, it is characterised in that including following step
Suddenly:
(A) off-line training sample data is obtained, sample data includes charge-discharge magnification at intervals of 0.2C, and temperature interval is 5 Celsius
Monomer terminal voltage, charging and discharging currents, lug temperature and the cycle life parameter and corresponding SOC data of the lower lithium battery of degree, will
The input layer vector of monomer terminal voltage, charging and discharging currents, lug temperature and cycle life parameter as network, SOC is as network
Output layer vector;According to following formula, all training samples are normalized:
<mrow>
<msup>
<mi>R</mi>
<mo>*</mo>
</msup>
<mo>=</mo>
<mfrac>
<mrow>
<mi>R</mi>
<mo>-</mo>
<msub>
<mi>R</mi>
<mrow>
<mi>m</mi>
<mi>i</mi>
<mi>n</mi>
</mrow>
</msub>
</mrow>
<mrow>
<msub>
<mi>R</mi>
<mrow>
<mi>m</mi>
<mi>a</mi>
<mi>x</mi>
</mrow>
</msub>
<mo>-</mo>
<msub>
<mi>R</mi>
<mrow>
<mi>m</mi>
<mi>i</mi>
<mi>n</mi>
</mrow>
</msub>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, R be actual sample actual value, R* be normalized after data, RmaxFor corresponding types data sample most
Big value, RminFor the minimum value of corresponding types data sample;
(B) the lithium battery SOC appraising models based on RBF neural are established:
First, basic function of the Gaussian function for RBF neural hidden layer node is selected, its expression is as follows:
Wherein, x is 4 dimension input vectors, corresponds to monomer terminal voltage, charging and discharging currents, lug temperature and the circulation of lithium battery respectively
Life parameter, ciIt is the center vector of i-th of neuron node of hidden layer, dimension is identical with input vector x, σi 2For i-th of god
Center width through first node;
Secondly according to training sample, the input data using the discharge and recharge data of any time lithium battery as RBF neural will
Output data of the SOC data as RBF neural corresponding to the moment, then output expression formula is as follows corresponding to the model:
WhereinExported for network, i.e. SOC estimated values, wiRepresent i-th of neuron node to the connection weight of output node, k
For network node in hidden layer;
(C) RBF neural network model established is optimized using STA optimized algorithms:
First, RBF network inputs sample sets are classified with K-means clustering algorithms, reduces node in hidden layer, for
The number k of classification determination, determined using apart from cost principle:
<mrow>
<msub>
<mi>l</mi>
<mrow>
<mi>k</mi>
<mn>1</mn>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mi>k</mi>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>k</mi>
</munderover>
<munder>
<mrow>
<mi>m</mi>
<mi>i</mi>
<mi>n</mi>
</mrow>
<mrow>
<mn>1</mn>
<mo>&le;</mo>
<mi>j</mi>
<mo>&le;</mo>
<mi>k</mi>
<mo>,</mo>
<mi>j</mi>
<mo>&NotEqual;</mo>
<mi>i</mi>
</mrow>
</munder>
<mo>|</mo>
<msub>
<mi>c</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>c</mi>
<mi>j</mi>
</msub>
<mo>|</mo>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>4</mn>
<mo>)</mo>
</mrow>
</mrow>
lk1For average infima species border distance, all cluster centres are defined as to being averaged apart from sum with its nearest cluster centre
Value, ciFor the center of ith cluster, cjFor the center of j-th of cluster;
<mrow>
<msub>
<mi>l</mi>
<mrow>
<mi>k</mi>
<mn>2</mn>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mi>n</mi>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>k</mi>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<msub>
<mi>m</mi>
<mi>i</mi>
</msub>
</munderover>
<mo>|</mo>
<msub>
<mi>x</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mo>-</mo>
<msub>
<mi>c</mi>
<mi>i</mi>
</msub>
<mo>|</mo>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>5</mn>
<mo>)</mo>
</mrow>
</mrow>
lk2For average inter- object distance, the average value of all clustering cluster inner distance sums is defined as, n is total sample number, miFor i-th
The total sample number of individual cluster, xijFor j-th of sample of intra-cluster;
lk=lk1-lk2 (6)
lkWhen for classification number being k apart from cost, then optimal classification number k should cause in maximum classification number kmaxIt is interior, lkFor
Maximum, i.e.,:
lk=max (li), i=1,2 ... kmax (7)
Meanwhile for the classification number k of determination, when being clustered with K-means algorithms, to overcome K-means algorithms to exist
The shortcomings that local optimum is easily trapped into when being clustered, changed every time in cluster process using STA algorithm optimization K-means algorithms
The selection of the central point in generation, corresponding optimization problem can be equivalent to below equation:
<mrow>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msub>
<mi>C</mi>
<mrow>
<mi>k</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>+</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
</msub>
<mo>=</mo>
<msub>
<mi>A</mi>
<mi>i</mi>
</msub>
<msub>
<mi>C</mi>
<mrow>
<mi>k</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
</mrow>
</msub>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>l</mi>
<mrow>
<mi>k</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>+</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
</msub>
<mo>=</mo>
<msub>
<mi>l</mi>
<mrow>
<mi>k</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>+</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mn>1</mn>
</mrow>
</msub>
<mo>-</mo>
<msub>
<mi>l</mi>
<mrow>
<mi>k</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>+</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mn>2</mn>
</mrow>
</msub>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>8</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein Ck(i+1)Represent a state, specially k center position, lk(i+1)It is corresponding Ck(i+1)Distance generation under state
Valency, specific formula for calculation such as formula (4) to formula (6);AiCan be considered as the operation operator of optimized algorithm for state-transition matrix;
After STA algorithm optimizes, you can obtain classification number as the optimal classification scheme in the case of k;
Secondly, the central point c of RBF networks is determined using STA algorithmi, extension width σi 2And wi;Using STA optimization RBF network ginsengs
Number problem is represented by following relation:
<mrow>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msub>
<mi>S</mi>
<mrow>
<mi>k</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>=</mo>
<msub>
<mi>A</mi>
<mi>k</mi>
</msub>
<msub>
<mi>S</mi>
<mi>k</mi>
</msub>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>e</mi>
<mi>r</mi>
<mi>r</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mrow>
<mi>k</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mi>n</mi>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</munderover>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>SOC</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mover>
<mrow>
<mi>S</mi>
<mi>O</mi>
<mi>C</mi>
</mrow>
<mo>^</mo>
</mover>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>9</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, SkRepresent a state, map network center ci, extension width σi 2And connection weight wiOne group of solution, AkIt is shape
State transfer matrix, for the operation operator of optimized algorithm, SOCiFor the SOC value of i-th of sample,For the estimation of i-th of sample
Value, n are total number of samples, err (xk+1) object function, the mean square error of SOC actual values and estimated value is defined as, uses STA
Training RBF networks comprise the following steps that:
(a) population at individual number SE=30, the random equality initialization c in feasible zone are initializedi、σi 2And wiThree variables, produce just
Beginning population, produce SE group initial feasible solutions;
(b) one group of c for making object function f reach minimum value is selected from current populationi、σi 2With w values, best is designated as, it is corresponding
Cost is fbest, and best is copied as into the colony that number of individuals is SE, is designated as S (k), and carrying out stretching by formula (10) obtains newly
Population:
S (k+1)=S (k)+γ ReS(k) (10)
Wherein, γ is normal number, referred to as contraction-expansion factor, value 1, Re∈Rn×nFor a random diagonal matrix, x (k+1) is S (k)
New population after stretching;
The optimum individual in population after stretching is newbest, and corresponding cost is gbest, if gbest is less than
Fbest, then translation transformation is carried out to individual newbest by formula (11), and update the best after translation transformation and fbest, otherwise
Without translation transformation;
<mrow>
<mi>S</mi>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>+</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mi>S</mi>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
<mo>+</mo>
<msub>
<mi>&beta;R</mi>
<mi>m</mi>
</msub>
<mfrac>
<mrow>
<mo>&lsqb;</mo>
<mi>S</mi>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mi>S</mi>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mo>&rsqb;</mo>
</mrow>
<mrow>
<mo>|</mo>
<mo>|</mo>
<mi>S</mi>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mi>S</mi>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mo>|</mo>
<msub>
<mo>|</mo>
<mn>2</mn>
</msub>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>11</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, β is normal number, referred to as shift factor, value 1, Rm∈Rn×nFor a stochastic variable, its value is in [0,1] scope
Interior, S (k-1) is that newbest individuals carry out the value before stretching;
(c) best is copied as into the colony that number of individuals is SE, right back-pushed-type (12) carries out rotation transformation and obtains new population, selects
Optimum individual newbest after rotation transformation in population, corresponding cost are gbest;
<mrow>
<mi>S</mi>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>+</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mrow>
<mo>(</mo>
<msub>
<mi>I</mi>
<mi>n</mi>
</msub>
<mo>+</mo>
<mi>&alpha;</mi>
<mfrac>
<mn>1</mn>
<mrow>
<mi>n</mi>
<mo>|</mo>
<mo>|</mo>
<mi>S</mi>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
<mo>|</mo>
<msub>
<mo>|</mo>
<mn>2</mn>
</msub>
</mrow>
</mfrac>
<mi>R</mi>
<mo>)</mo>
</mrow>
<mi>S</mi>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>12</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, α is normal number, referred to as twiddle factor, value 1, R ∈ Rn×nRandom matrix, its element value is in [- 1,1] model
Enclose, | | | |2For 2 norms of vector, InIt is unit matrix;
If gbest is less than fbest, translation transformation is carried out by formula (11), and updates the best after translation transformation and fbest,
Otherwise without translation transformation;
(d) best is copied as into the colony that number of individuals is SE, right back-pushed-type (13) carries out principal axis transformation and judged, selection conversion
Individual optimal in population is designated as newbest afterwards, and corresponding cost is gbest;
Sk+1=Sk+δRaSk (13)
Wherein, δ is the axle factor, value 1, RaFor random diagonal matrix;
If gbest is less than fbest, translation transformation is carried out by formula (11), and updates the best after translation transformation and fbest,
Otherwise without translation transformation;
(e) repeat step b) to step d), until fitness meets minimum requirement or reaches iterations;
(f) kernel function center ci, the extension width σ that preservation optimizes to obtain through STAi 2And connection weight wi;(D) preserve and train
RBF network structures and each parameter value, by the RBF networks trained be used for ferric phosphate lithium cell SOC estimation.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711093872.7A CN107741568B (en) | 2017-11-08 | 2017-11-08 | Lithium battery SOC estimation method based on state transition optimization RBF neural network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711093872.7A CN107741568B (en) | 2017-11-08 | 2017-11-08 | Lithium battery SOC estimation method based on state transition optimization RBF neural network |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107741568A true CN107741568A (en) | 2018-02-27 |
CN107741568B CN107741568B (en) | 2019-12-31 |
Family
ID=61234245
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711093872.7A Active CN107741568B (en) | 2017-11-08 | 2017-11-08 | Lithium battery SOC estimation method based on state transition optimization RBF neural network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107741568B (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108519556A (en) * | 2018-04-13 | 2018-09-11 | 重庆邮电大学 | A kind of lithium ion battery SOC prediction techniques based on Recognition with Recurrent Neural Network |
CN108647423A (en) * | 2018-05-04 | 2018-10-12 | 中南大学 | A kind of optimization method of alumina leaching process flash system port size |
CN109031152A (en) * | 2018-10-15 | 2018-12-18 | 重庆科技学院 | Lead-acid accumulator residual capacity flexible measurement method based on neural network transparence |
CN110336336A (en) * | 2018-03-29 | 2019-10-15 | 中国科学院微电子研究所 | A kind of remaining battery capacity calculation method and device |
CN112363058A (en) * | 2020-10-30 | 2021-02-12 | 哈尔滨理工大学 | Lithium ion battery safety degree estimation method and device based on impedance spectrum and Markov characteristic |
CN112462267A (en) * | 2020-11-25 | 2021-03-09 | 珠海格力电器股份有限公司 | Lithium battery SOC estimation method, lithium battery module and power utilization equipment |
CN114740376A (en) * | 2022-05-05 | 2022-07-12 | 杭州科工电子科技有限公司 | On-line diagnosis method for battery state of charge |
CN114895206A (en) * | 2022-04-26 | 2022-08-12 | 合肥工业大学 | Lithium ion battery SOH estimation method based on RBF neural network of improved wolf optimization algorithm |
WO2022213789A1 (en) * | 2021-10-19 | 2022-10-13 | 深圳市德兰明海科技有限公司 | Lithium battery soc estimation method and apparatus, and computer-readable storage medium |
CN116627708A (en) * | 2023-07-24 | 2023-08-22 | 湖南惟储信息技术有限公司 | Storage fault analysis system and method thereof |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103822960A (en) * | 2014-02-17 | 2014-05-28 | 中南大学 | Online multi-metal ion concentration detection method utilizing polarography |
CN104834975A (en) * | 2015-05-13 | 2015-08-12 | 国家电网公司 | Power network load factor prediction method based on intelligent algorithm optimization combination |
CN106443453A (en) * | 2016-07-04 | 2017-02-22 | 陈逸涵 | Lithium battery SOC estimation method based on BP neural network |
US9633148B2 (en) * | 2014-11-28 | 2017-04-25 | International Business Machines Corporation | Static timing analysis in circuit design |
-
2017
- 2017-11-08 CN CN201711093872.7A patent/CN107741568B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103822960A (en) * | 2014-02-17 | 2014-05-28 | 中南大学 | Online multi-metal ion concentration detection method utilizing polarography |
US9633148B2 (en) * | 2014-11-28 | 2017-04-25 | International Business Machines Corporation | Static timing analysis in circuit design |
CN104834975A (en) * | 2015-05-13 | 2015-08-12 | 国家电网公司 | Power network load factor prediction method based on intelligent algorithm optimization combination |
CN106443453A (en) * | 2016-07-04 | 2017-02-22 | 陈逸涵 | Lithium battery SOC estimation method based on BP neural network |
Non-Patent Citations (2)
Title |
---|
BI JUN等: "State of charge estimation of Li-ion batteries in an electric vehicle based on a radial-basis-function neural network", 《CHIN.PHYS.B》 * |
阳春华等: "基于双重状态转移优化RBF神经网络的锂离子电池SOC估算方法", 《第28届中国过程控制会议(CPCC2017)暨纪念中国过程控制会议30周年摘要集》 * |
Cited By (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110336336A (en) * | 2018-03-29 | 2019-10-15 | 中国科学院微电子研究所 | A kind of remaining battery capacity calculation method and device |
CN108519556A (en) * | 2018-04-13 | 2018-09-11 | 重庆邮电大学 | A kind of lithium ion battery SOC prediction techniques based on Recognition with Recurrent Neural Network |
CN108647423A (en) * | 2018-05-04 | 2018-10-12 | 中南大学 | A kind of optimization method of alumina leaching process flash system port size |
CN109031152A (en) * | 2018-10-15 | 2018-12-18 | 重庆科技学院 | Lead-acid accumulator residual capacity flexible measurement method based on neural network transparence |
CN112363058A (en) * | 2020-10-30 | 2021-02-12 | 哈尔滨理工大学 | Lithium ion battery safety degree estimation method and device based on impedance spectrum and Markov characteristic |
CN112363058B (en) * | 2020-10-30 | 2022-06-24 | 哈尔滨理工大学 | Lithium ion battery safety degree estimation method and device based on impedance spectrum and Markov characteristic |
CN112462267A (en) * | 2020-11-25 | 2021-03-09 | 珠海格力电器股份有限公司 | Lithium battery SOC estimation method, lithium battery module and power utilization equipment |
JP2023520970A (en) * | 2021-10-19 | 2023-05-23 | 深▲せん▼市徳蘭明海科技有限公司 | Lithium battery SOC estimation method, apparatus, and computer-readable storage medium |
WO2022213789A1 (en) * | 2021-10-19 | 2022-10-13 | 深圳市德兰明海科技有限公司 | Lithium battery soc estimation method and apparatus, and computer-readable storage medium |
EP4113139A4 (en) * | 2021-10-19 | 2023-11-08 | Shenzhen Poweroak Newener Co., Ltd | Lithium battery soc estimation method and apparatus, and computer-readable storage medium |
JP7450741B2 (en) | 2021-10-19 | 2024-03-15 | 深▲せん▼市徳蘭明海科技有限公司 | Lithium battery SOC estimation method, device and computer readable storage medium |
CN114895206A (en) * | 2022-04-26 | 2022-08-12 | 合肥工业大学 | Lithium ion battery SOH estimation method based on RBF neural network of improved wolf optimization algorithm |
CN114895206B (en) * | 2022-04-26 | 2023-04-28 | 合肥工业大学 | Lithium ion battery SOH estimation method based on RBF neural network of improved gray wolf optimization algorithm |
CN114740376B (en) * | 2022-05-05 | 2023-01-13 | 杭州科工电子科技有限公司 | On-line diagnosis method for battery state of charge |
CN114740376A (en) * | 2022-05-05 | 2022-07-12 | 杭州科工电子科技有限公司 | On-line diagnosis method for battery state of charge |
CN116627708A (en) * | 2023-07-24 | 2023-08-22 | 湖南惟储信息技术有限公司 | Storage fault analysis system and method thereof |
CN116627708B (en) * | 2023-07-24 | 2023-09-19 | 湖南惟储信息技术有限公司 | Storage fault analysis system and method thereof |
Also Published As
Publication number | Publication date |
---|---|
CN107741568B (en) | 2019-12-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107741568A (en) | A kind of lithium battery SOC estimation method that optimization RBF neural is shifted based on state | |
CN110286332B (en) | Electric vehicle power battery SOC estimation method based on multiple innovation theory | |
Liu et al. | An advanced Lithium-ion battery optimal charging strategy based on a coupled thermoelectric model | |
Ouyang et al. | Improved parameters identification and state of charge estimation for lithium-ion battery with real-time optimal forgetting factor | |
CN111007399B (en) | Lithium battery state of charge prediction method based on improved generation countermeasure network | |
CN102981125B (en) | A kind of electrokinetic cell SOC method of estimation based on RC equivalent model | |
CN104899439B (en) | A kind of lithium ion battery modelling by mechanism method | |
CN107576919A (en) | Power battery charged state estimating system and method based on ARMAX models | |
Ma et al. | Co-estimation of state of charge and state of health for lithium-ion batteries based on fractional-order model with multi-innovations unscented Kalman filter method | |
CN110688808B (en) | Particle swarm and LM optimization hybrid iterative identification method of power battery model | |
CN110525269B (en) | SOC battery pack balance control method | |
CN110346734A (en) | A kind of lithium-ion-power cell health status evaluation method based on machine learning | |
Wu et al. | The SOC estimation of power Li-Ion battery based on ANFIS model | |
CN110850298B (en) | Lithium battery SOH estimation method and system based on data driving | |
CN111458646A (en) | Lithium battery SOC estimation method based on PSO-RBF neural network | |
CN111581904A (en) | Lithium battery SOC and SOH collaborative estimation method considering influence of cycle number | |
CN109839599B (en) | Lithium ion battery SOC estimation method based on second-order EKF algorithm | |
CN103983920B (en) | A kind of method of the model of the electrokinetic cell setting up electric vehicle | |
CN111428433B (en) | Lithium ion battery state calculation method based on hybrid filtering | |
Li et al. | A novel state estimation approach based on adaptive unscented Kalman filter for electric vehicles | |
CN114839536B (en) | Lithium ion battery health state estimation method based on multiple health factors | |
CN114660464A (en) | Lithium ion battery state of charge estimation method | |
CN114217234B (en) | IDE-ASRCKF-based lithium ion battery parameter identification and SOC estimation method | |
CN114545279A (en) | Lithium battery health state estimation method based on neural network ordinary differential equation | |
CN111948539A (en) | Kalman filtering lithium ion battery SOC estimation method based on deep reinforcement learning |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CP01 | Change in the name or title of a patent holder |
Address after: Yuelu District City, Hunan province 410083 Changsha Lushan Road No. 932 Co-patentee after: Guangdong Boliwei Technology Co., Ltd. Patentee after: CENTRAL SOUTH University Address before: Yuelu District City, Hunan province 410083 Changsha Lushan Road No. 932 Co-patentee before: DONGGUAN GREENWAY BATTERY Co.,Ltd. Patentee before: CENTRAL SOUTH University |
|
CP01 | Change in the name or title of a patent holder |