CN110187287A - A kind of retired lithium battery complementary energy rapid detection method - Google Patents
A kind of retired lithium battery complementary energy rapid detection method Download PDFInfo
- Publication number
- CN110187287A CN110187287A CN201910549921.6A CN201910549921A CN110187287A CN 110187287 A CN110187287 A CN 110187287A CN 201910549921 A CN201910549921 A CN 201910549921A CN 110187287 A CN110187287 A CN 110187287A
- Authority
- CN
- China
- Prior art keywords
- lithium battery
- discharge
- neural network
- data
- complementary energy
- 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
- 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]
- G01R31/385—Arrangements for measuring battery or accumulator variables
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Secondary Cells (AREA)
- Charge And Discharge Circuits For Batteries Or The Like (AREA)
Abstract
The invention discloses a kind of retired lithium battery complementary energy rapid detection methods, include the following steps: step 1: constructing the BP neural network model for predicting retired lithium battery complementary energy；Step 2: obtaining the required training sample set of neural network model training；Step 3: BP neural network model being trained using training sample set；Step 4: the capacity of retired lithium battery to be measured being detected using the BP neural network after training.It is quickly tested the present invention has the advantages that carrying out residual capacity to retired lithium battery using BP neural network using matlab software, the capacity prediction for providing accurate retired lithium battery, avoids the cumbersome defect that the prior art needs repeated charge to test.
Description
Technical field
The present invention relates to lithium battery detection field, in particular to a kind of retired lithium battery complementary energy rapid detection method.
Background technique
As new-energy automobile is gradually popularized, largely superseded old and useless battery will can be generated.These eliminate the surplus of battery
Covolume amount can reach 80% or more, can be applied to other ting model equipment completely or for energy storage, avoid environmental pollution
And it is energy saving.Therefore, the echelon research on utilization of retired lithium battery has the meaning of reality.Realize that echelon utilizes, then it is right
Retired lithium battery carries out complementary energy detection.Under existing national standards, the detection of lithium battery complementary energy process is more complex and detection time mistake
It is long, it needs to carry out multiple charge and discharge to battery, loss can also be generated to battery itself.
Summary of the invention
It is an object of the invention to overcome the deficiencies of the prior art and provide a kind of quick sides of detection of retired lithium battery complementary energy
Method, for detecting the remaining battery capacity of retired lithium battery.
To achieve the goals above, a kind of the technical solution adopted by the present invention are as follows: the quick side of detection of retired lithium battery complementary energy
Method includes the following steps:
Step 1: constructing the BP neural network model for predicting retired lithium battery complementary energy；
Step 2: obtaining the required training sample set of neural network model training；
Step 3: BP neural network model being trained using training sample set；
Step 4: the capacity of retired lithium battery to be measured being detected using the BP neural network after training.
Training sample set includes multiple groups sample data, and every group of sample data includes charging curve, discharge curve, charging temperature
Degree, discharge temp, residual capacity.
BP neural network model includes input layer, middle layer and output layer, discharge curve data, electric discharge in training sample
Temperature, charging curve data, charging temperature are input layer input data, and output layer corresponding data is residual capacity, pass through charging
Curve data, charging temperature, discharge curve data, discharge temp, residual capacity are trained BP neural network.
The BP neural network is combined using the adaptive algorithm in Neural Network Toolbox in matlab with LM algorithm
Neural network.
The method for obtaining training sample includes that several retired lithium batteries is taken to carry out charge-discharge test, obtains each group of lithium battery
Charging temperature, charging curve data, discharge temp, discharge curve data and corresponding residual capacity data, select remaining hold
The test data for measuring the lithium battery between the 60%-90% of rated capacity forms training sample set.
When carrying out charge-discharge test to retired lithium battery, multiple charge and discharge are carried out to each lithium battery, are repeatedly being filled
Remaining capacity is acquired under electric discharge；The curve data of single charge each time, the impulse discharge curve data under multiple charge and discharge are recorded,
Multiple impulse discharge curve datas are averaging to obtain discharge curve data, multiple single charge curve datas are averaging to obtain
Charging curve data.
When carrying out complementary energy detection to retired lithium battery, lithium battery to be measured is subjected to a charge and discharge and obtains corresponding fill
Electro-temperature, charging curve data, discharge temp, discharge curve data are input to the BP neural network after training, then by BP mind
Residual electric quantity through network output prediction.
The present invention has the advantages that carrying out residual capacity to retired lithium battery using BP neural network using matlab software
It is quickly tested, provides the capacity prediction of accurate retired lithium battery, avoid the prior art and need repeated charge
The cumbersome defect of test.
Detailed description of the invention
Below to each width attached drawing of description of the invention expression content and figure in label be briefly described:
Fig. 1 is complementary energy testing process schematic diagram of the present invention；
Specific embodiment
A specific embodiment of the invention is made further detailed below against attached drawing by the description to optimum embodiment
Thin explanation.
As shown in Figure 1, a kind of retired lithium battery complementary energy rapid detection method, includes the following steps:
Step 1: constructing the BP neural network model for predicting retired lithium battery complementary energy；
Step 2: obtaining the required training sample set of neural network model training；
Step 3: BP neural network model being trained using training sample set；
Step 4: the capacity of retired lithium battery to be measured being detected using the BP neural network after training.
In step 1, the BP neural network of building generally comprises input layer, middle layer and output layer, it is preferable that uses
The BP neural network tool of matlab software combines the quick complementary energy to retired lithium battery with LM algorithm using adaptive algorithm
Detection model optimizes.It combines and BP neural network is optimized using adaptive algorithm and LM algorithm, by network
Weight, threshold value and convergence rate etc. optimize, can effectively improve the convergence rate of BP neural network and reduce prediction and miss
Difference.
In step 2, in order to form training sample, several retired lithium batteries is selected to carry out charge-discharge test, selection
Quantity confirms that the sample size of retired lithium battery selection is more, and training can be more accurate according to actual needs.
Using new prestige CT-4008-5V6A-S1 tester to lithium battery carry out charge-discharge test, according to national standards under, it is right
Each lithium battery carries out multiple charge and discharge, and impulse discharge curve data, the single each time recorded in charge and discharge process fills
Electric curve data is averaged the corresponding a plurality of single charge curve data of each lithium battery to obtain the lithium battery corresponding
The a plurality of impulse discharge curve data of each lithium battery is averaged to obtain the corresponding electric discharge song of the lithium battery by charging curve
Line.Each charging temperature data, discharge temp data are recorded simultaneously, the corresponding charging of charging curve data is used as after being averaging
Temperature, the corresponding discharge temp of discharge temp curve data.Charging curve data refer to voltage change in charging process and when
Between the curve that is correspondingly formed, discharge curve data refer to the curve that voltage change is correspondingly formed at any time in discharge process, curve
Data refer to that curve is characterized in the form of array or function curve by the curve data in charge and discharge process is input to BP nerve net
Network.Measuring data according to tester obtains remaining capacity value simultaneously, which is acquired down according to national standards
's.Then lithium battery corresponding data of the residual capacity between rated capacity 60%-90% are chosen as training sample set.Instruction
Practicing sample set includes multiple groups sample data, and every group of sample data includes charging temperature, charging curve data, discharge temp, electric discharge
Curve data, residual capacity.
The method for obtaining training sample includes that several retired lithium batteries is taken to carry out charge-discharge test, and every group of lithium battery carries out more
Secondary charge and discharge, obtain multiple groups charging and discharging lithium battery under charging curve data, charging temperature, discharge curve data, discharge temp
And corresponding residual capacity data, in actual selection training process, be eliminated dismantling after choosing the electric vehicle use of certain vehicle enterprise
18650 lithium batteries, random 16 section, 18650 lithium battery that takes out tested, and one group of 4 section is then divided into, in 20 DEG C of ± 5 DEG C of conditions
Under, lithium battery is with 2A electric current constant-current charge, until lithium battery voltage reaches 4.2V, termination electric current is 0.1A, 1h is stood, then 20
1C (i.e. electric current 2A) discharges under the conditions of DEG C ± 5 DEG C, when final voltage is set as 2.75V, then stands 1h, and repetitive cycling 3 times.Then
Temperature when corresponding charge and discharge is acquired in charge and discharge process and the voltage in charge and discharge process, current data form charge and discharge
Electric curve and corresponding actual residual capacity data are obtained by the data of acquisition, is held according to the real surplus that measurement obtains
The data of lithium battery of the data decimation residual capacity of acquisition between 60%~90% are constituted sample set, by lithium battery by amount
Charging curve data, discharge curve data, charging temperature, discharge temp, residual capacity be added complementary energy detection calculating in,
The sample data of acquisition is pre-processed, sample data is divided into training sample and test sample, and be normalized.
To obtain the training sample set and test sample of normalized.
The BP neural network of building is trained, BP neural network is trained using training sample, BP nerve net
Network includes input layer, middle layer, output layer, is arranged training termination condition, the corresponding node of input layer include charging curve data,
Discharge curve data and corresponding charging temperature, discharge temp, the residual capacity of output layer corresponding node pass through charging curve
Data, discharge curve data, charging temperature, discharge temp, residual capacity are trained BP neural network.It is soft by MATLAB
BP neural network tool box in part combines the quick complementary energy detection to retired lithium battery using adaptive algorithm with LM algorithm
BP neural network model optimize, training data is trained, until meet training termination condition after terminate instruction
Practice.
When carrying out complementary energy detection to retired lithium battery, lithium battery to be measured is subjected to a charge and discharge and obtains corresponding fill
Electric curve data, discharge curve data and corresponding charging temperature, discharge temp data are input to the BP nerve net after training
Network, then by the residual electric quantity of BP neural network output prediction.Retired lithium is predicted using the BP neural network after training
Battery remaining power, charge and discharge electric standard according to national standards detect charging curve data, discharge curve number in charge and discharge process
Accordingly and charge and discharge electro-temperature is input in the BP neural network after training, is then predicted according to the input data by BP neural network
The remaining capacity value of corresponding lithium battery out.
In order to verify training after Neural Network model predictive accuracy, training after the completion of, by test sample input optimize
It is tested in BP neural network model afterwards, obtains the remaining capacity value (predicted value) and test sample of neural network prediction
In corresponding remaining capacity value (true value), the difference between predicted value and true value is observed, thus to test using BP nerve
Error between the residual capacity and actual capacity of network test.This programme carries out complementary energy detection for same 18650 battery of section
Error 1.3% or so, carry out the error of complementary energy detection 2.2% or so, lower than country's mark for different 18650 batteries of section
Quasi- 5% requires.That is it is predicted by BP neural network in the present processes, it can be after a charge and discharge just
Remaining capacity value can quickly be obtained.
The method that the retired lithium battery complementary energy in of the invention 18650 quickly detects, has fully considered lithium battery under national standard
The detection of charge and discharge complementary energy, too long and lossy to the battery actual influence of detection time.It constructs more than the progress of BP neural network model
It can detection.Multiple charge-discharge test is carried out to 18650 lithium batteries at different temperatures, according to national standards with normal mode of operation
Operation, and acquire data using upper computer module and carry out BP neural network training with data.In prediction by upper computer module
The data of acquisition, the BP neural network model for being input to optimization are predicted, can be found that this method reality by prediction data result
Show 18650 retired lithium battery complementary energy under different temperatures quickly to detect.
Compared with prior art, the beneficial effects of the invention are as follows consider charging and discharging lithium battery complementary energy under national standard to examine
It surveys, detection time is too long and lossy to battery, is detected using BP neural network model to retired lithium battery complementary energy, is 18650
Complementary energy quickly detects and provides a kind of practicable scheme retired lithium battery at different temperatures.
Obviously present invention specific implementation is not subject to the restrictions described above, as long as using method concept and skill of the invention
The improvement for the various unsubstantialities that art scheme carries out, it is within the scope of the present invention.
Claims (7)
1. a kind of retired lithium battery complementary energy rapid detection method, characterized by the following steps:
Step 1: constructing the BP neural network model for predicting retired lithium battery complementary energy；
Step 2: obtaining the required training sample set of neural network model training；
Step 3: BP neural network model being trained using training sample set；
Step 4: the capacity of retired lithium battery to be measured being detected using the BP neural network after training.
2. a kind of retired lithium battery complementary energy rapid detection method as described in claim 1, it is characterised in that: training sample set packet
Multiple groups sample data is included, every group of sample data includes charging curve, discharge curve, charging temperature, discharge temp, residual capacity.
3. a kind of retired lithium battery complementary energy rapid detection method as claimed in claim 1 or 2, it is characterised in that: BP nerve net
Network model includes input layer, middle layer and output layer, discharge curve data, discharge temp in training sample, charging curve number
It is input layer input data according to, charging temperature, output layer corresponding data is residual capacity, passes through charging curve data, charging temperature
Degree, discharge curve data, discharge temp, residual capacity are trained BP neural network.
4. a kind of retired lithium battery complementary energy rapid detection method as claimed in claim 1 or 2, it is characterised in that: the BP mind
The neural network combined using the adaptive algorithm in Neural Network Toolbox in matlab with LM algorithm through network.
5. a kind of retired lithium battery complementary energy rapid detection method as claimed in claim 1 or 2, it is characterised in that: obtain training
The method of sample includes that several retired lithium batteries is taken to carry out charge-discharge test, obtains charging temperature, the charging of each group of lithium battery
Curve data, discharge temp, discharge curve data and corresponding residual capacity data select residual capacity in rated capacity
The test data of lithium battery between 60%-90% forms training sample set.
6. a kind of retired lithium battery complementary energy rapid detection method as claimed in claim 5, it is characterised in that: to retired lithium electricity
When pond carries out charge-discharge test, multiple charge and discharge are carried out to each lithium battery, acquire remaining capacity under multiple charge and discharge；Note
The curve data of single charge each time, the impulse discharge curve data under multiple charge and discharge are recorded, by multiple impulse discharge curve numbers
Discharge curve data are obtained according to averaging, multiple single charge curve datas are averaging to obtain charging curve data.
7. a kind of retired lithium battery complementary energy rapid detection method as claimed in claim 1 or 2, it is characterised in that: to retired
When lithium battery carries out complementary energy detection, lithium battery to be measured is subjected to a charge and discharge and obtains corresponding charging temperature, charging curve
Data, discharge temp, discharge curve data are input to the BP neural network after training, then by BP neural network output prediction
Residual electric quantity.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910549921.6A CN110187287A (en) | 2019-06-24 | 2019-06-24 | A kind of retired lithium battery complementary energy rapid detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910549921.6A CN110187287A (en) | 2019-06-24 | 2019-06-24 | A kind of retired lithium battery complementary energy rapid detection method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110187287A true CN110187287A (en) | 2019-08-30 |
Family
ID=67722972
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910549921.6A Pending CN110187287A (en) | 2019-06-24 | 2019-06-24 | A kind of retired lithium battery complementary energy rapid detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110187287A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111814826A (en) * | 2020-06-08 | 2020-10-23 | 武汉理工大学 | Rapid detection and rating method for residual energy of retired power battery |
| CN112686380A (en) * | 2020-12-28 | 2021-04-20 | 江苏宝航能源技术有限公司 | Neural network-based echelon power cell consistency evaluation method and system |
| CN113238157A (en) * | 2020-12-09 | 2021-08-10 | 北京大学深圳研究生院 | Method for screening through AI detection on retired batteries of electric vehicles |
| CN113820608A (en) * | 2021-08-20 | 2021-12-21 | 北京邮电大学 | Method for predicting remaining capacity of battery in echelon and electronic equipment |
| CN113933718A (en) * | 2021-11-04 | 2022-01-14 | 格林美股份有限公司 | Retired battery capacity sorting method, device, equipment and storage medium |
| CN114280491A (en) * | 2021-12-23 | 2022-04-05 | 中山大学 | Retired battery residual capacity estimation method based on active learning |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101067644A (en) * | 2007-04-20 | 2007-11-07 | 杭州高特电子设备有限公司 | Storage battery performance analytical expert diagnosing method |
| CN104375091A (en) * | 2014-11-18 | 2015-02-25 | 柳州市金旭节能科技有限公司 | Monitoring method for electric vehicle power storage battery |
| CN104953191A (en) * | 2015-05-07 | 2015-09-30 | 国网河南省电力公司电力科学研究院 | Restoration method for performances of transformer substation returned lead-acid storage battery |
| CN107482266A (en) * | 2017-08-28 | 2017-12-15 | 江西优特汽车技术有限公司 | A kind of power battery pack method for group matching |
-
2019
- 2019-06-24 CN CN201910549921.6A patent/CN110187287A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101067644A (en) * | 2007-04-20 | 2007-11-07 | 杭州高特电子设备有限公司 | Storage battery performance analytical expert diagnosing method |
| CN104375091A (en) * | 2014-11-18 | 2015-02-25 | 柳州市金旭节能科技有限公司 | Monitoring method for electric vehicle power storage battery |
| CN104953191A (en) * | 2015-05-07 | 2015-09-30 | 国网河南省电力公司电力科学研究院 | Restoration method for performances of transformer substation returned lead-acid storage battery |
| CN107482266A (en) * | 2017-08-28 | 2017-12-15 | 江西优特汽车技术有限公司 | A kind of power battery pack method for group matching |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111814826A (en) * | 2020-06-08 | 2020-10-23 | 武汉理工大学 | Rapid detection and rating method for residual energy of retired power battery |
| CN111814826B (en) * | 2020-06-08 | 2022-06-03 | 武汉理工大学 | Rapid detection and rating method for residual energy of retired power battery |
| CN113238157A (en) * | 2020-12-09 | 2021-08-10 | 北京大学深圳研究生院 | Method for screening through AI detection on retired batteries of electric vehicles |
| CN113238157B (en) * | 2020-12-09 | 2022-11-22 | 北京大学深圳研究生院 | Method for screening through AI detection on retired batteries of electric vehicles |
| CN112686380A (en) * | 2020-12-28 | 2021-04-20 | 江苏宝航能源技术有限公司 | Neural network-based echelon power cell consistency evaluation method and system |
| CN113820608A (en) * | 2021-08-20 | 2021-12-21 | 北京邮电大学 | Method for predicting remaining capacity of battery in echelon and electronic equipment |
| CN113820608B (en) * | 2021-08-20 | 2022-08-05 | 北京邮电大学 | Method for predicting remaining capacity of battery in echelon and electronic equipment |
| CN113933718A (en) * | 2021-11-04 | 2022-01-14 | 格林美股份有限公司 | Retired battery capacity sorting method, device, equipment and storage medium |
| CN113933718B (en) * | 2021-11-04 | 2024-02-09 | 格林美股份有限公司 | Retired battery capacity sorting method, retired battery capacity sorting device, retired battery capacity sorting equipment and storage medium |
| CN114280491A (en) * | 2021-12-23 | 2022-04-05 | 中山大学 | Retired battery residual capacity estimation method based on active learning |
| CN114280491B (en) * | 2021-12-23 | 2024-01-05 | 中山大学 | Retired battery residual capacity estimation method based on active learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110187287A (en) | A kind of retired lithium battery complementary energy rapid detection method | |
| CN109655754B (en) | Battery performance evaluation method based on multi-dimensional grading of charging process | |
| CN112436202B (en) | Stepped current charging method for preventing lithium precipitation of lithium ion battery cathode | |
| CN110426639B (en) | Lithium ion battery service life prediction method and system based on dynamic impedance spectrum | |
| Yu et al. | A branch current estimation and correction method for a parallel connected battery system based on dual BP neural networks | |
| CN108983106A (en) | Novel power battery health state evaluation method | |
| CN106443461A (en) | Battery energy storage system state assessment method | |
| CN107085187A (en) | Echelon utilizes the determination method and device of battery energy storage system consistency maintenance index | |
| CN112986830A (en) | Lithium battery capacity estimation method based on convolution time memory neural network | |
| CN108732499B (en) | Method and system for detecting cycle life of lithium ion battery | |
| CN112147530B (en) | Battery state evaluation method and device | |
| CN102520367A (en) | Method for evaluating life of space hydrogen-nickel storage batteries | |
| CN113109729B (en) | Vehicle power battery SOH evaluation method based on accelerated aging test and real vehicle working condition | |
| CN104502844A (en) | Power lithium battery deterioration degree diagnosis method based on AC impedance | |
| CN103728570B (en) | Battery-thermal-characteristic-based health state detection method | |
| CN110764009A (en) | Method for estimating battery pack circulation capacity and health state of electric automobile | |
| CN112363075A (en) | Lithium ion battery aging evaluation method | |
| CN108169680A (en) | Power battery health state evaluation method, system and electric vehicle | |
| CN112327170B (en) | Power battery full-period residual life estimation method based on neural network | |
| CN114578237A (en) | Method, system and equipment for rapidly estimating health state of lithium ion battery based on constant-current charging time | |
| CN109613440A (en) | Stage division, device, equipment and the storage medium of battery | |
| CN114280479A (en) | Electrochemical impedance spectrum-based rapid sorting method for retired batteries | |
| CN107505575B (en) | Rapid assessment method for retired power battery | |
| CN111766530A (en) | Method for detecting service life of lithium ion storage battery monomer | |
| CN112686380A (en) | Neural network-based echelon power cell consistency evaluation method and system |
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 |