CN108693478A - A kind of method for detecting leakage of lithium-ion-power cell - Google Patents
A kind of method for detecting leakage of lithium-ion-power cell Download PDFInfo
- Publication number
- CN108693478A CN108693478A CN201810341466.6A CN201810341466A CN108693478A CN 108693478 A CN108693478 A CN 108693478A CN 201810341466 A CN201810341466 A CN 201810341466A CN 108693478 A CN108693478 A CN 108693478A
- Authority
- CN
- China
- Prior art keywords
- leakage
- battery
- occurs
- ion
- lithium
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000007637 random forest analysis Methods 0.000 claims abstract description 21
- 238000003745 diagnosis Methods 0.000 claims abstract description 8
- 238000004422 calculation algorithm Methods 0.000 claims description 22
- 238000012549 training Methods 0.000 claims description 12
- 238000003066 decision tree Methods 0.000 claims description 9
- 238000012544 monitoring process Methods 0.000 claims description 6
- 238000012360 testing method Methods 0.000 claims description 5
- 238000002474 experimental method Methods 0.000 claims description 4
- 238000012952 Resampling Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 230000002068 genetic effect Effects 0.000 claims description 3
- 235000013399 edible fruits Nutrition 0.000 claims 1
- 208000032953 Device battery issue Diseases 0.000 abstract description 3
- 230000000694 effects Effects 0.000 abstract description 2
- 239000007788 liquid Substances 0.000 description 4
- HBBGRARXTFLTSG-UHFFFAOYSA-N Lithium ion Chemical compound [Li+] HBBGRARXTFLTSG-UHFFFAOYSA-N 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000005611 electricity Effects 0.000 description 3
- 229910001416 lithium ion Inorganic materials 0.000 description 3
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 239000003792 electrolyte Substances 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 239000000178 monomer Substances 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Landscapes
- Secondary Cells (AREA)
Abstract
The present invention provides a kind of method for detecting leakage of lithium-ion-power cell, can accurately judge whether short trouble causes battery leakage in the case where not opening battery case observation.By establishing the battery model of external short circuit failure, and grader of the operation based on random forest method realizes joint leakage identification.This method is suitable for battery failures diagnostic system, can predict to provide foundation with diagnosis for the fault degree after battery short circuit, have many advantageous effects such as operation is simple, is easily achieved.
Description
Technical field
The present invention relates to cell safety technical field more particularly to a kind of method for detecting leakage of lithium-ion-power cell.
Background technology
External short circuit belongs to a kind of catastrophic discontinuityfailure most commonly seen in vehicle lithium-ion power battery failure, outside occurring
In tens seconds after portion's short circuit, high temperature and high current caused by power battery, it is most likely that cause serious leakage, leak
Electrolyte then can further cause the even more serious consequence including on fire, therefore power battery leakage is differentiated
It is highly important.However, since vehicle lithium-ion power battery is usually disposed with battery case, by unpacking, observation battery is
No leakage was both inconvenient or dangerous, and automobile-used lithium ion battery is often hundreds of batteries series, parallel composition battery packs,
Dense arrangement between battery, even if observing that leakage is also difficult to realize positioning to the battery location of specific leakage.It can be seen that existing
The leakage detection mode of some lithium-ion-power cells still has certain limitation, is not beaten even if this field urgent need is a kind of
The method that leakage detects can be realized by opening battery case also.
Invention content
For technical problem present in above-mentioned this field, the present invention provides a kind of leakage of lithium-ion-power cell inspections
Survey method, specifically includes following steps:
Step 1: carrying out multigroup external short circuit experiment, record current, voltage and temperature to different lithium-ion-power cells
Data;
Step 2: battery model is established to lithium-ion-power cell, and based on the data recorded in the step 1
Parameter identification is carried out to the battery model, and determines and short trouble state threshold and generation leakage state threshold occurs;
Step 3: inline diagnosis is carried out to failure using the battery model established in step 2, and according to power battery
Data judge it and short trouble state occur and the goodness of fit of leakage state occurs;
Step 4: the grader based on random forests algorithm is established, and according to the data recorded in the step 1
The grader is trained, is realized to short trouble state occurs and the monitoring of leakage state occurs.
Step 5: combine the battery model and it is described based on the grader of random forests algorithm to whether leakage occurring
It is detected.
Further, the step 2 specifically includes:
The battery model established uses fractional order impedance model, utilizes electric current, voltage and the temperature when not occurring leakage
Degrees of data is trained the model, to establish out the electricity that can simulate non-leakage state lower outer portion short circuit electrical property change
Pool model;Parameter identification is carried out to the fractional order impedance model based on genetic algorithm.
Further, failure progress inline diagnosis is specifically included described in the step 3:
The voltage that practical power battery voltage and the fractional order impedance model export under short trouble state is carried out
Compare, calculates since short circuit to the goodness of fit χ in the period at current time:
VmFor model prediction terminal voltage, V is the practical terminal voltage of battery, and n is data length.
The generation short trouble state threshold δ that will be determined in the goodness of fit χ and the step 21And it leaks
Liquid status threshold value δ2It is compared.
Further, the grader based on random forests algorithm and its training process established in the step 4 are specific
Including:
4.1, after acquisition short circuit occurs in M lithium-ion-power cell sample leakage battery and non-leakage battery it is maximum warm
Liter and discharge capacity statistical data, are input to as total training dataset in random forests algorithm;
4.2, resampling is carried out using Bootstrap methods, randomly generates N number of training dataset S1,S2,…,SN.Wherein,
Si(1≤i≤N) is the data set that ith has the extraction put back to be formed from M battery sample data;
4.3, it is based on generated N number of training dataset and generates N decision tree at random respectively, after every decision tree is trained to
Grasp the classification capacity for judging whether that leakage occurs by maximum temperature rise and discharge capacity data.
Further, the joint fractional order impedance model described in the step 5 and described it is based on random forest
The grader of algorithm is detected, and is specifically included:
5.1, according to goodness of fit χ and the generation short trouble state threshold δ1And leakage state threshold δ occurs2Ratio
Relatively result is judged, if χ > δ2, it is believed that leakage, assignment K1=0 does not occur;If δ1<χ<δ2, then it is assumed that leakage occurs, assigns
Value K=1;
5.2, run the grader based on random forests algorithm, by collected lithium-ion-power cell maximum temperature rise with put
Electricity is input to as test data set X in trained each decision tree, exports leakage monitoring state K2, when generation leakage
When assignment K2=1;
5.3, when K1 and K2 are simultaneously 1, assignment K=1 indicates that leakage occurs for battery, otherwise K=0, and expression is not leaked
Liquid.
It, being capable of the accurately judgement short circuit in the case where not opening battery case observation according to the above method provided by the present invention
Whether failure causes battery leakage.By establishing the battery model of external short circuit failure, and operation based on random forest side
The grader of method realizes joint leakage identification.This method is suitable for battery failures diagnostic system, can be for after battery short circuit
Fault degree predicts to provide foundation with diagnosis have many advantageous effects such as operation is simple, is easily achieved.
Description of the drawings
Fig. 1 is the flow diagram according to method provided by the present invention
Fig. 2 is the schematic diagram of lithium-ion-power cell fractional order impedance model
Fig. 3 is that power battery terminal voltage predicted value changes over time pass with measured value in a specific example according to the present invention
It is (SoC=20%)
Fig. 4 is that power battery terminal voltage predicted value changes over time pass with measured value in a specific example according to the present invention
It is (SoC=60%)
Fig. 5 is that power battery terminal voltage predicted value changes over time pass with measured value in a specific example according to the present invention
It is (SoC=100%)
Fig. 6 is the grader schematic diagram based on random forests algorithm
Fig. 7 is that leakage detection differentiates result
Specific implementation mode
Below in conjunction with the accompanying drawings to the technical solution of the method for detecting leakage of lithium-ion-power cell provided by the present invention, do
Go out and further illustrates in detail.
As shown in Figure 1, method provided by the present invention specifically includes following steps:
Step 1: carrying out multigroup external short circuit experiment, record current, voltage and temperature to different lithium-ion-power cells
Data;
Step 2: fractional order impedance model is established to lithium-ion-power cell, and based on the institute recorded in the step 1
It states data and parameter identification is carried out to the fractional order impedance model, and determine and short trouble state threshold and generation leakage occurs
State threshold;
Step 3: inline diagnosis is carried out to failure using the fractional order impedance model established in step 2, and according to dynamic
Power battery data judges it and short trouble state occurs and the goodness of fit of leakage state occurs;
Step 4: the grader based on random forests algorithm is established, and according to the data recorded in the step 1
The grader is trained, is realized to short trouble state occurs and the monitoring of leakage state occurs.
Step 5: combine the fractional order impedance model and it is described based on the grader of random forests algorithm to whether sending out
Raw leakage is detected.
In the preferred embodiment of the application, the step 2 specifically includes:
The fractional order impedance model established uses fractional order impedance model, as shown in Fig. 2, V in figureocvFor open-circuit voltage,
RoFor ohmic internal resistance, RctFor electric charge transfer internal resistance, CPE is normal phase angle original paper, and W is weber impedance.When using leakage not occurring
Electric current, voltage and temperature data the model is trained, it is initial for basic, normal, high three kinds under room temperature (20 DEG C)
SoC (20%, 60% and 100%) is carried out, and tests 1 battery cell under identical conditions respectively, and 3 battery cells of record were tested
Cell voltage in journey and curent change, this 3 monomers are without occurring the phenomenon of leakage.Non- leakage can be simulated to establish out
The battery model of state lower outer portion short circuit electrical property change;Parameter is carried out based on genetic algorithm to the fractional order impedance model to distinguish
Know.Power battery terminal voltage prediction result and actual measurement terminal voltage result are as in Figure 3-5.Error mean square root is as listed in table 1.
Table 1
In the preferred embodiment of the application, failure progress inline diagnosis is specifically wrapped described in the step 3
It includes:
The voltage that practical power battery voltage and the fractional order impedance model export under short trouble state is carried out
Compare, calculates since short circuit to the goodness of fit χ in the period at current time:
VmFor model prediction terminal voltage, V is the practical terminal voltage of battery, and n is data length.
The generation short trouble state threshold δ that will be determined in the goodness of fit χ and the step 21And it leaks
Liquid status threshold value δ2It is compared.Optionally 3 initial SoC are respectively that 20%, 60% and 100% battery carries out external short circuit reality
It tests, note electric current, voltage and temperature data.Using the first step trained model prediction external short circuit failure behavior, obtain
Preceding 2s inner terminal voltages predicted value and measured value root-mean-square error, as shown in table 2.
Table 2
In the preferred embodiment of the application, as shown in fig. 6, established in the step 4 based on random forest
The grader and its training process of algorithm specifically include:
4.1, after acquisition short circuit occurs in M lithium-ion-power cell sample leakage battery and non-leakage battery it is maximum warm
Liter and discharge capacity statistical data, are input to as total training dataset in random forests algorithm;
4.2, resampling is carried out using Bootstrap methods, randomly generates N number of training dataset S1,S2,…,SN.Wherein,
Si(1≤i≤N) is the data set that ith has the extraction put back to be formed from M battery sample data;
4.3, it is based on generated N number of training dataset and generates N decision tree at random respectively, after every decision tree is trained to
Grasp the classification capacity for judging whether that leakage occurs by maximum temperature rise and discharge capacity data.
In the preferred embodiment of the application, the joint fractional order impedance model described in the step 5 with
And the grader based on random forests algorithm is detected, and is specifically included:
5.1, according to goodness of fit χ and the generation short trouble state threshold δ1And leakage state threshold δ occurs2Ratio
Relatively result is judged, if χ > δ2, it is believed that leakage, assignment K1=0 does not occur;If δ1<χ<δ2, then it is assumed that leakage occurs, assigns
Value K=1;
5.2, run the grader based on random forests algorithm, by collected lithium-ion-power cell maximum temperature rise with put
Electricity is input to as test data set X in trained each decision tree, exports leakage monitoring state K2, when generation leakage
When assignment K2=1;
5.3, when K1 and K2 are simultaneously 1, assignment K=1 indicates that leakage occurs for battery, otherwise K=0, and expression is not leaked
Liquid.
In the case that known SoC is 100% from the experiment of this case verification, leakage occurs for battery, using patent of the present invention
Last differentiates that result is also coincide with practical to leakage, and classification tree number is as shown in Fig. 7.
It although an embodiment of the present invention has been shown and described, for the ordinary skill in the art, can be with
Understanding without departing from the principles and spirit of the present invention can carry out these embodiments a variety of variations, modification, replace
And modification, the scope of the present invention is defined by the appended.
Claims (5)
1. a kind of method for detecting leakage of lithium-ion-power cell, it is characterised in that:Specifically include following steps:
Step 1: carrying out multigroup external short circuit experiment, record current, voltage and temperature number to different lithium-ion-power cells
According to;
Step 2: battery model is established to lithium-ion-power cell, and based on the data recorded in the step 1 to institute
It states battery model and carries out parameter identification, and determine and short trouble state threshold and generation leakage state threshold occurs;
Step 3: inline diagnosis is carried out to failure using the battery model established in step 2, and according to power battery data
Judge it and short trouble state occurs and the goodness of fit of leakage state occurs;
Step 4: the grader based on random forests algorithm is established, and according to the data recorded in the step 1 to institute
It states grader to be trained, realize to short trouble state occurs and the monitoring of leakage state occurs.
Step 5: combine the battery model and it is described based on the grader of random forests algorithm to whether occur leakage carry out
Detection.
2. the method as described in claim 1, it is characterised in that:The step 2 specifically includes:
The battery model established uses fractional order impedance model, utilizes electric current, voltage and the temperature number when not occurring leakage
It is trained according to the model, to establish out the battery mould that can simulate non-leakage state lower outer portion short circuit electrical property change
Type;Parameter identification is carried out to the fractional order impedance model based on genetic algorithm.
3. method as claimed in claim 2, it is characterised in that:It is specific to failure progress inline diagnosis described in the step 3
Including:
Practical power battery voltage is compared with the voltage that the fractional order impedance model exports under short trouble state,
It calculates since short circuit to the goodness of fit χ in the period at current time:
VmFor model prediction terminal voltage, V is the practical terminal voltage of battery, and n is data length.
The generation short trouble state threshold δ that will be determined in the goodness of fit χ and the step 21And leakage shape occurs
State threshold value δ2It is compared.
4. the method as described in claim 1, it is characterised in that:Established in the step 4 based on random forests algorithm
Grader and its training process specifically include:
4.1, after acquisition short circuit occurs in M lithium-ion-power cell sample the maximum temperature rise of leakage battery and non-leakage battery and
Discharge capacity statistical data is input to as total training dataset in random forests algorithm;
4.2, resampling is carried out using Bootstrap methods, randomly generates N number of training dataset S1,S2,…,SN.Wherein, Si(1
≤ i≤N) it is the data set that ith has the extraction put back to be formed from M battery sample data;
4.3, it is based on generated N number of training dataset and generates N decision tree at random respectively, every decision tree is grasped after being trained to
Judge whether the classification capacity of generation leakage by maximum temperature rise and discharge capacity data.
5. method as claimed in claim 3, it is characterised in that:The joint fractional order modulus of impedance described in the step 5
Type and the grader based on random forests algorithm are detected, and are specifically included:
5.1, according to goodness of fit χ and the generation short trouble state threshold δ1And leakage state threshold δ occurs2Comparison knot
Fruit is judged, if χ > δ2, it is believed that leakage, assignment K1=0 does not occur;If δ1<χ<δ2, then it is assumed that leakage, assignment K=occurs
1;
5.2, the grader based on random forests algorithm is run, by collected lithium-ion-power cell maximum temperature rise and discharge capacity
It is input in trained each decision tree as test data set X, exports leakage monitoring state K2, assigned when leakage occurs
Value K2=1;
5.3, when K1 and K2 are simultaneously 1, assignment K=1 indicates that leakage occurs for battery, otherwise K=0, and leakage does not occur for expression.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810341466.6A CN108693478A (en) | 2018-04-17 | 2018-04-17 | A kind of method for detecting leakage of lithium-ion-power cell |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810341466.6A CN108693478A (en) | 2018-04-17 | 2018-04-17 | A kind of method for detecting leakage of lithium-ion-power cell |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108693478A true CN108693478A (en) | 2018-10-23 |
Family
ID=63845556
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810341466.6A Pending CN108693478A (en) | 2018-04-17 | 2018-04-17 | A kind of method for detecting leakage of lithium-ion-power cell |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108693478A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109657720A (en) * | 2018-12-20 | 2019-04-19 | 浙江大学 | A kind of inline diagnosis method of power transformer shorted-turn fault |
| WO2020155233A1 (en) * | 2019-01-28 | 2020-08-06 | 东北大学 | Method for diagnosing external short circuit fault of lithium-ion battery pack on basis of two-stage model prediction |
| CN111537913A (en) * | 2020-05-14 | 2020-08-14 | 广东汉力威技术有限公司 | Method for diagnosing short circuit by using pre-discharge loop |
| CN116429339A (en) * | 2023-06-07 | 2023-07-14 | 海斯坦普汽车组件(北京)有限公司 | Leakage detection method and system for new energy battery box |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106168799A (en) * | 2016-06-30 | 2016-11-30 | 常伟 | A kind of method carrying out batteries of electric automobile predictive maintenance based on big data machine learning |
| CN106526493A (en) * | 2016-11-01 | 2017-03-22 | 北京理工大学 | Power battery external short circuit fault diagnosing and temperature rise prediction method and system based on BP neural networks |
| CN106980091A (en) * | 2017-03-29 | 2017-07-25 | 北京理工大学 | A kind of electrokinetic cell system health status method of estimation based on fractional model |
| CN107192956A (en) * | 2017-05-19 | 2017-09-22 | 北京理工大学 | A kind of battery short circuit leakage on-line monitoring method and device |
-
2018
- 2018-04-17 CN CN201810341466.6A patent/CN108693478A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106168799A (en) * | 2016-06-30 | 2016-11-30 | 常伟 | A kind of method carrying out batteries of electric automobile predictive maintenance based on big data machine learning |
| CN106526493A (en) * | 2016-11-01 | 2017-03-22 | 北京理工大学 | Power battery external short circuit fault diagnosing and temperature rise prediction method and system based on BP neural networks |
| CN106980091A (en) * | 2017-03-29 | 2017-07-25 | 北京理工大学 | A kind of electrokinetic cell system health status method of estimation based on fractional model |
| CN107192956A (en) * | 2017-05-19 | 2017-09-22 | 北京理工大学 | A kind of battery short circuit leakage on-line monitoring method and device |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109657720A (en) * | 2018-12-20 | 2019-04-19 | 浙江大学 | A kind of inline diagnosis method of power transformer shorted-turn fault |
| CN109657720B (en) * | 2018-12-20 | 2021-05-11 | 浙江大学 | On-line diagnosis method for turn-to-turn short circuit fault of power transformer |
| WO2020155233A1 (en) * | 2019-01-28 | 2020-08-06 | 东北大学 | Method for diagnosing external short circuit fault of lithium-ion battery pack on basis of two-stage model prediction |
| CN111537913A (en) * | 2020-05-14 | 2020-08-14 | 广东汉力威技术有限公司 | Method for diagnosing short circuit by using pre-discharge loop |
| CN116429339A (en) * | 2023-06-07 | 2023-07-14 | 海斯坦普汽车组件(北京)有限公司 | Leakage detection method and system for new energy battery box |
| CN116429339B (en) * | 2023-06-07 | 2023-09-01 | 海斯坦普汽车组件(北京)有限公司 | Leakage detection method and system for new energy battery box |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107153162B (en) | A kind of power battery pack multiple faults online test method | |
| CN110133536B (en) | System, method and apparatus for determining an indicator of leakage current within a battery entity | |
| CN108693478A (en) | A kind of method for detecting leakage of lithium-ion-power cell | |
| CN106353690B (en) | Utilize the method for Petri network diagnosis lithium battery failure | |
| CN105304954A (en) | Battery grouping method and battery grouping system | |
| CN108306065A (en) | Lithium ion battery grouping method and lithium ion battery combo system | |
| Zhang et al. | A multi-fault diagnosis method for lithium-ion battery pack using curvilinear Manhattan distance evaluation and voltage difference analysis | |
| CN104813182B (en) | The steady state detection of abnormal charge event in the cell device being connected in series | |
| CN112098850B (en) | Lithium ion battery voltage fault diagnosis method and system based on SDO algorithm | |
| CN110376530A (en) | Battery internal short-circuit detection device and method | |
| CN109061495A (en) | A kind of hybrid energy-storing battery failure diagnostic method | |
| CN101907688A (en) | Method for detecting electrical property consistency of lithium ion battery | |
| CN108333548A (en) | Insulation resistance measuring apparatus and fault self-diagnosis method | |
| CN113687251B (en) | Double-model-based lithium ion battery pack voltage abnormality fault diagnosis method | |
| CN116027199B (en) | Method for detecting short circuit in whole service life of battery cell based on electrochemical model parameter identification | |
| CN113300437A (en) | Dynamic management and control system of lithium battery energy storage system | |
| CN109507600A (en) | A kind of retired ferric phosphate lithium cell echelon utilizes appraisal procedure | |
| CN108957349A (en) | A kind of lithium ion battery failure detection method and system | |
| CN111580003A (en) | Impedance spectrum-based secondary battery inconsistency identification method and apparatus | |
| US11977122B2 (en) | Battery inspection apparatus | |
| CN104198943B (en) | A kind of lithium battery pack method for testing performance and system | |
| Wu et al. | Research on short-circuit fault-diagnosis strategy of lithium-ion battery in an energy-storage system based on voltage cosine similarity | |
| Wang et al. | Battery pack topology structure on state-of-charge estimation accuracy in electric vehicles | |
| CN112379271B (en) | Passivation-considered capacity detection method for carbon-coated lithium thionyl chloride battery | |
| CN117517969A (en) | Lithium battery aging and internal short circuit fault evaluation method based on relaxation voltage curve |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181023 |
|
| RJ01 | Rejection of invention patent application after publication |