CN103413033A - Method for predicting storage battery faults - Google Patents
Method for predicting storage battery faults Download PDFInfo
- Publication number
- CN103413033A CN103413033A CN2013103228674A CN201310322867A CN103413033A CN 103413033 A CN103413033 A CN 103413033A CN 2013103228674 A CN2013103228674 A CN 2013103228674A CN 201310322867 A CN201310322867 A CN 201310322867A CN 103413033 A CN103413033 A CN 103413033A
- Authority
- CN
- China
- Prior art keywords
- battery
- fault
- symptom
- membership
- degree
- 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
Images
Landscapes
- Secondary Cells (AREA)
Abstract
The invention discloses a method for predicting storage battery faults, and belongs to the field of storage battery fault prediction and diagnosis. The method is characterized in that whether a battery is broken down is predicted according to related data during the battery use process; according to the knowledge-based method for predicting the health status of the storage battery, namely, a fuzzy prediction method, degrees of membership of various faults are solved according to degrees of membership of certain symptoms and used for representing tendency of various faults, and scientific basis is provided for predicting the health status of the storage battery; and the method is proposed in order to effectively predict the faults of the battery and a battery pack, experimental data is obtained with a nickel-cadmium battery serving as an experimental subject, the DOH (degree of health) of the battery or the battery pack in the whole operation process can be obtained through analysis and processing for real-time data of the nickel-cadmium battery, and simultaneously maintenance information of the battery is given. The method is simple and effective, the disassembly and the discharge testing of the battery are avoided, and the accurate battery health level is obtained.
Description
One, technical field
The present invention relates to failure prediction and the diagnostic field of accumulator.
Two, background technology
Along with the fast development of China's economy, to the requirement of the numerous areas such as the energy, traffic, environment also in continuous raising.Battery pack does not allow the electric power system of cutting off the power supply all, is all an indispensable power-supply system.And the application of battery system in all trades and professions is also more and more extensive.Whether the operation of accumulator is normal, will directly affect normal, the reliable and safe operation of various device in this field.
The battery failures diagnostic techniques is the new technology just grown up in recent years, and the whole world not yet accomplishes tangible results to the research of battery failure diagnostic system at present.In the applicating maintenance of battery, with the most use is artificial testing and diagnosing method, the discharge test namely battery removed by certain multiplying power carries out cubic content measurement, to determine the active volume of accumulator, becomes behindhand battery thereby judge which batteries capacity loss.The discharge capacity test that this diagnostic mode reality is only a kind of off-line, do not carry out comprehensive diagnostic to battery.The defect of its method be the operation element amount large, take time manyly, thereby be not easy to repetitive operation, and repeat cubic content measurement, be disadvantageous to battery, accumulator can not Emergency use.In addition, capacity test needs special test equipment for the high voltage accumulator, may be dangerous.So should avoid as much as possible, directly by electric discharge, carry out measuring capacity and battery is made to diagnosis.
The Forecasting Methodology that we propose is change and consider that other factors diagnose battery by the parameter between the different monomers battery of Real-Time Monitoring, more same electric battery, its thought is: each cell of same electric battery is under identical charging and discharging currents, the performance great majority of each cell are close, but also have the performance of some batteries to exist inconsistent.By considering each battery, in the period, in the change in voltage of this section in the time, can estimate the quality of battery performance to the skew of average voltage and each battery at this section.Be a kind of online real-time predicting method by this method, can effectively avoid many drawbacks of manual testing's diagnosis.
Three, summary of the invention
The objective of the invention is to propose a kind of online, real-time nickel-cadmium battery failure prediction system.Utilize this method can obtain the probability of various faults appearance and health status DOH and the health level of this accumulator (or electric battery), can provide thus the maintenance information of battery.
Whether a kind of method of predicting accumulator failure, is characterized in that the method is for the method for cadmium-nickel storage cell based on expertise, utilize the related data prediction battery in the battery use procedure to break down.Based on the accumulator health status Forecasting Methodology of knowledge, be also the fuzzy prediction method, according to the degree of membership of some symptom, obtain the degree of membership of various faults, the tendentiousness existed in order to characterize various faults, for the prediction that judges cell health state provides scientific basis.Its reasoning process as shown in Figure 1
1) at first utilize battery historical archives data and health data and calculate battery symptom degree of membership in conjunction with the various symptom degree of membership computing formula of battery.Sympotomatic set U={ symptom 1, symptom 2, symptom 3 ... symptom n}, wherein the degree of membership of each symptom forms fuzzy vector α=[μ
1, μ
2, μ
3... μ
n]
T.
The μ in α wherein
nBy following formula, calculate: μ
n=f
sta(I) f
cha(X);
F
sta(I) be function of state,
1. static:
2. charging:
3. electric discharge:
Annotate: I is current value.
F
cha(X) be the external characteristic data function,
F
cha(X) be the degree of membership value of certain state of nickel-cadmium battery performance data;
X is the measured value of certain variation of the external characteristic data of battery;
X
aMean value for certain variation of the external characteristic data of all batteries;
C
rFor constant, C
r=1.5, the relative coefficient of expression external characteristic data.
So just obtained the vectorial α that each symptom degree of membership obtains=[μ
1, μ
2, μ
3... μ
n]
T
2) utilize each symptom degree of membership of battery and in conjunction with weighting Judgement Matrix R, calculate the various fault degrees of membership of battery.
The fault collection V={ fault 1 of battery, fault 2, fault 3 ... fault n}, n herein and the n in the first step are same n.
A) weighting Judgement Matrix R's determines
The weighting Judgement Matrix define two kinds of modes, they are all to take this domain expert's experience as basis.A kind of is when the expert has larger assurance to definite fuzzy matrix, can provide the flexible strategy of each symptom, then carries out normalization or regular, forms the weighting Judgement Matrix; Another kind is when expert's oneself assurance is not very big, can provide the importance ranking that each symptom produces certain fault in conjunction with other a plurality of experts, then this significance sequence qualitatively is mapped as to a kind of flexible strategy and distributes, as the weighting Judgement Matrix.Considering complicacy, the ambiguity of Battery Diagnostic, be namely that veteran expert also is difficult to provide definite weights, so we adopts the second way.
At first two set U and V are carried out to importance by from big to small order sequence, then renumber, now the U={ symptom 1, symptom 2, symptom 3, symptom n} and V={ fault 1, fault 2, fault 3 ... element in these two set of fault n} is arranged by importance, the set of definition when these two set have not been initial, order wherein changes.In addition, the vectorial α of each symptom degree of membership composition=[μ
1, μ
2, μ
3... μ
n]
TAlso to resequence according to U.
Then utilize the binomial coefficient weighted method to carry out determining of matrix R.
Known
N is consistent with above-mentioned n herein.Binomial coefficient obtains one group and be 1 set W thus, is
According to the character of binomial coefficient, this set is the symmetry set that an intermediate value is maximum and progressively reduce to both sides.
Order
W={W
1, W
2..., W
i..., W
n.
Therefore, weighting Judgement Matrix R is:
R on diagonal line in each element in R
11=r
22=r
33=...=r
NnBe equal to the maximal value of set in W, the remaining position of every delegation is by from big to small order assignment successively.
B) calculating of fault degree of membership
Failure collection V={ fault 1, fault 2, fault 3 ... fault n}, wherein the set β of the degree of membership of each element composition=[f
1, f
2, f
3..., f
n]
T, and β=R α
3) utilize the fault degree of membership, provide the health status of battery after comprehensive.
The calculating of DOF and DOH:
Failure degree DOF, to fault degree of membership set β=[f
1, f
2, f
3..., f
n]
TIf there is f=[f
1, f
2, f
3F
n] ∈ F and f
i>=0.5,1≤i≤n, DOF=f
1⊕ f
2⊕ f
3⊕ ... ⊕ f
nOtherwise DOF=f
1∨ f
2∨ f
3∨ ... ∨ f
n.
Wherein operator ∨ is defined as and gets greatly, and operator ⊕ is defined as and gets union, namely meets:
x⊕y=x+y-x×y
Health status DOH, DOH=C
1* DOF+C
2* DOH '
C wherein
1, C
2For constant, C
1=0.7, C
2=0.3; DOF is the failure degree of this calculating; DOH' for last operation calculate DOH, if there is no logout last time, DOH'=0.
According to the value of the DOH calculated, according to form 1, provide the health status of battery.
Form 1 battery health rank and corresponding measure
Health level DOH | Span | Counter-measure |
One-level | <0.4 | Health, without processing |
Secondary | 0.4-0.7 | Strengthen safeguarding |
Three grades | >0.7 | Change |
Four, accompanying drawing explanation
Fig. 1: based on the cell health state prediction flow process of expertise
Fig. 2: based on the cell health state algorithm flow of fuzzy logic
Five, embodiment
For nickel-cadmium battery, its fault collection and sympotomatic set are arranged according to importance:
Fault V={ volume lowering, internal resistance increases, aging };
Symptom U={ is B group voltage/14-monomer voltage in the same time, Δ U/ Δ t, Δ U is very big }; Due to n=3;
Weighting Judgement Matrix like this
Its specific algorithm is as follows, and process flow diagram is illustrated in fig. 2 shown below.
Primary data comprises: the data such as the voltage of monomer and electric current in A+B assembled battery total voltage, A battery voltage, B battery group.
The battery initial data segment that 1) at first will collect.Battery data to annual March is analyzed, and by the data sectional of every day, we only consider the charge and discharge data in period here, and each sampled point is exactly a time period.
2) then calculate the function of state of battery charge and discharge each symptom in period in every day.Because obtained data here are in large electric current, put, fill period, that is to say during this period, its function of state value is constantly equal to 1.
3), after obtaining the function of state value of each symptom of battery, continue to calculate the external characteristic function of this each symptom of battery.The X that each symptom is got
aShown in following form 2.Such as, for symptom 1, the measured value X=0.0475 of electric discharge time period in period, so now the external characteristic functional value f of symptom 1
cha(X)=X/ (X
a* C
r)=0.0475/ (0.05028 * 1.5)=0.6298
Form 2 each symptom X
aValue
4) calculate the membership function of each symptom.By step 2 and 3, calculated function of state f
sta(I) and external characteristic data function f
chaSo the membership function F (t) of each symptom (X),
n=f
sta(I) f
cha(X).
5) complete if the membership function of each symptom all calculates, so just can carry out the calculating of next time period.The namely calculating of point of next sampling time.
6) after the calculating of the symptom degree of membership of all time periods is complete, carry out average value processing, try to achieve this symptom degree of membership on the same day, then calculate the fault degree of membership on the same day.Such as, obtained symptom degree of membership set α=[0.3082,0.4370,0.3339] of certain day
T, known R, the set of fault degree of membership β=R α=[0.3469,0.3791,0.3533] so
T.
7) according to the fault degree of membership obtained, judge the health status of battery, provide simultaneously maintenance information.If β=[0.3469,0.3791,0.3533]
T, DOF=0.3791 so, DOH=0.7 * 0.3791+0.3 * 0=0.2653.Due to now DOH<0.4, therefore deducibility goes out, battery now is healthy, can continue to use.
Claims (1)
1. method of predicting accumulator failure is characterized in that:
1) at first utilize battery historical archives data and health data and calculate battery symptom degree of membership in conjunction with the various symptom degree of membership computing formula of battery; Sympotomatic set U={ symptom 1, symptom 2, symptom 3 ... symptom n}, wherein the degree of membership of each symptom forms fuzzy vector α=[μ
1, μ
2, μ
3... μ
n]
T
The μ in α wherein
nBy following formula, calculate: μ
n=f
sta(I) f
cha(X);
F
sta(I) be function of state,
1. static:
2. charging:
3. electric discharge:
Annotate: I is current value;
F
cha(X) be the external characteristic data function,
F
cha(X) be the degree of membership value of certain state of nickel-cadmium battery performance data;
X is the measured value of certain variation of the external characteristic data of battery;
X
aMean value for certain variation of the external characteristic data of all batteries;
C
rFor constant, C
r=1.5, the relative coefficient of expression external characteristic data;
So just obtained the vectorial α that each symptom degree of membership obtains=[μ
1, μ
2, μ
3... μ
n]
T
2) utilize each symptom degree of membership of battery and in conjunction with weighting Judgement Matrix R, calculate the various fault degrees of membership of battery;
The fault collection V={ fault 1 of battery, fault 2, fault 3 ... fault n}, n herein and the n in the first step are same n;
A) weighting Judgement Matrix R's determines
Definite a plurality of experts of weighting Judgement Matrix provide the importance ranking that each symptom produces certain fault, then this significance sequence qualitatively are mapped as to a kind of flexible strategy and distribute, as the weighting Judgement Matrix;
At first two set U and V are carried out to importance by from big to small order sequence, then renumber, now the U={ symptom 1, symptom 2, symptom 3, symptom n} and V={ fault 1, fault 2, fault 3 ... element in these two set of fault n} is arranged by importance, the set of definition when these two set have not been initial, order wherein changes; In addition, the vectorial α of each symptom degree of membership composition=[μ
1, μ
2, μ
3... μ
n]
TAlso to resequence according to U;
Then utilize the binomial coefficient weighted method to carry out determining of matrix R;
Known
N is consistent with above-mentioned n herein; Binomial coefficient obtains one group and be 1 set W thus, is
According to the character of binomial coefficient, this set is the symmetry set that an intermediate value is maximum and progressively reduce to both sides;
Order
W={W
1, W
2..., W
i..., W
n;
Therefore, weighting Judgement Matrix R is:
R on diagonal line in each element in R
11=r
22=r
33=...=r
NnBe equal to the maximal value of set in W, the remaining position of every delegation is by from big to small order assignment successively;
B) calculating of fault degree of membership
Failure collection V={ fault 1, fault 2, fault 3 ... fault n}, wherein the set β of the degree of membership of each element composition=[f
1, f
2, f
3..., f
n]
T, and β=R α
3) utilize the fault degree of membership, provide the health status of battery after comprehensive;
The calculating of DOF and DOH:
Failure degree DOF, to fault degree of membership set β=[f
1, f
2, f
3..., f
n]
TIf there is f=[f
1, f
2, f
3..., f
n] ∈ F and f
i>=0.5,1≤i≤n, DOD=f
1⊕ f
2⊕ f
3⊕ ... ⊕ f
nOtherwise DOF=f
1∨ f
2∨ f
3∨ ... ∨ f
n
Wherein operator ∨ is defined as and gets greatly, and operator ⊕ is defined as and gets union, namely meets:
x⊕y=x+y-x×y
Health status DOH, DOH=C
1* DOF+C
2* COH '
C wherein
1, C
2For constant, C
1=0.7, C
2=0.3; DOF is the failure degree of this calculating; DOH' for last operation calculate DOH, if there is no logout last time, DOH'=0;
According to the value of the DOH calculated, according to following mode, provide the health status of battery:
Health status DOH span counter-measure
One-level<0.4 health, without outer reason
Secondary 0.4-0.7 strengthens safeguarding
Three grades > 0.7 replacing.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2013103228674A CN103413033A (en) | 2013-07-29 | 2013-07-29 | Method for predicting storage battery faults |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2013103228674A CN103413033A (en) | 2013-07-29 | 2013-07-29 | Method for predicting storage battery faults |
Publications (1)
Publication Number | Publication Date |
---|---|
CN103413033A true CN103413033A (en) | 2013-11-27 |
Family
ID=49606044
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN2013103228674A Pending CN103413033A (en) | 2013-07-29 | 2013-07-29 | Method for predicting storage battery faults |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103413033A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106227942A (en) * | 2016-07-25 | 2016-12-14 | 温州大学 | Electrokinetic cell system method for designing based on Brittleness Theory of Complex System |
CN107144722A (en) * | 2017-04-16 | 2017-09-08 | 知豆电动汽车有限公司 | Battery pressure drop rate alarm method |
CN107340475A (en) * | 2016-04-29 | 2017-11-10 | 株式会社日立制作所 | Battery fault detection method and battery fault detection device |
CN110069810A (en) * | 2019-03-11 | 2019-07-30 | 北京百度网讯科技有限公司 | Battery failures prediction technique, device, equipment and readable storage medium storing program for executing |
CN110116625A (en) * | 2019-05-16 | 2019-08-13 | 辽宁工业大学 | A kind of automobile storage battery fault monitoring method for electric-controlled vehicle |
CN111967191A (en) * | 2020-08-24 | 2020-11-20 | 哈尔滨理工大学 | Fuzzy Bayesian network-based lithium ion power battery safety degree evaluation method and device |
CN114895196A (en) * | 2022-07-13 | 2022-08-12 | 深圳市威特利电源有限公司 | New energy battery fault diagnosis method based on artificial intelligence |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20050119534A1 (en) * | 2003-10-23 | 2005-06-02 | Pfizer, Inc. | Method for predicting the onset or change of a medical condition |
CN101174715A (en) * | 2007-09-28 | 2008-05-07 | 深圳先进技术研究院 | Power battery management system with control and protection function and method thereof |
-
2013
- 2013-07-29 CN CN2013103228674A patent/CN103413033A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20050119534A1 (en) * | 2003-10-23 | 2005-06-02 | Pfizer, Inc. | Method for predicting the onset or change of a medical condition |
CN101174715A (en) * | 2007-09-28 | 2008-05-07 | 深圳先进技术研究院 | Power battery management system with control and protection function and method thereof |
Non-Patent Citations (3)
Title |
---|
刘文杰: "基于模糊理论的电池故障诊断专家系统", 《吉林大学学报》, vol. 21, no. 6, 30 November 2005 (2005-11-30), pages 1 * |
刘文杰: "电池组故障诊断专家系统的研究与实现", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》, no. 7, 15 July 2006 (2006-07-15) * |
梅胜敏: "故障诊断专家系统中的模糊推理方法", 《南京航空航天大学学报》, vol. 27, no. 4, 30 August 1995 (1995-08-30) * |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107340475A (en) * | 2016-04-29 | 2017-11-10 | 株式会社日立制作所 | Battery fault detection method and battery fault detection device |
CN106227942A (en) * | 2016-07-25 | 2016-12-14 | 温州大学 | Electrokinetic cell system method for designing based on Brittleness Theory of Complex System |
CN106227942B (en) * | 2016-07-25 | 2020-05-19 | 温州大学 | Power battery system design method based on complex system brittleness theory |
CN107144722A (en) * | 2017-04-16 | 2017-09-08 | 知豆电动汽车有限公司 | Battery pressure drop rate alarm method |
CN110069810A (en) * | 2019-03-11 | 2019-07-30 | 北京百度网讯科技有限公司 | Battery failures prediction technique, device, equipment and readable storage medium storing program for executing |
CN110069810B (en) * | 2019-03-11 | 2023-04-07 | 北京百度网讯科技有限公司 | Battery failure prediction method, device, equipment and readable storage medium |
CN110116625A (en) * | 2019-05-16 | 2019-08-13 | 辽宁工业大学 | A kind of automobile storage battery fault monitoring method for electric-controlled vehicle |
CN110116625B (en) * | 2019-05-16 | 2020-07-28 | 辽宁工业大学 | Automobile storage battery fault monitoring method for electric control vehicle |
CN111967191A (en) * | 2020-08-24 | 2020-11-20 | 哈尔滨理工大学 | Fuzzy Bayesian network-based lithium ion power battery safety degree evaluation method and device |
CN111967191B (en) * | 2020-08-24 | 2024-03-19 | 哈尔滨理工大学 | Lithium ion power battery safety evaluation method and device based on fuzzy Bayesian network |
CN114895196A (en) * | 2022-07-13 | 2022-08-12 | 深圳市威特利电源有限公司 | New energy battery fault diagnosis method based on artificial intelligence |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103413033A (en) | Method for predicting storage battery faults | |
CN101067644B (en) | Storage battery performance analytical expert diagnosing method | |
CN106033113A (en) | Health state evaluation method for energy-storage battery pack | |
CN102520361B (en) | State of health (SOH) value assessment method of battery pack | |
CN105277896B (en) | Lithium battery method for predicting residual useful life based on ELM MUKF | |
CN108254696A (en) | The health state evaluation method and system of battery | |
CN103412264B (en) | The conforming evaluation method of cell in battery pack | |
CN106443461A (en) | Battery energy storage system state assessment method | |
CN105652106A (en) | Comprehensive performance testing method and apparatus of energy storage system | |
CN103033762B (en) | Difference value differential based lithium iron phosphate power battery power loading capacity dynamic prediction method | |
JPH08233918A (en) | Method and apparatus for estimation of residual unused capacity and residual unused time of battery during discharge | |
CN107367698A (en) | The health status Forecasting Methodology of electric automobile lithium battery group | |
CN103941191A (en) | Energy storing device integrated management method and energy storing device | |
CN104316879A (en) | Method for forecasting service life of lead-acid storage battery | |
CN103886518A (en) | Early warning method for voltage sag based on electric energy quality data mining at monitoring point | |
CN104680024A (en) | Method for predicting remaining useful life of lithium ion battery based on GA (Genetic Algorithms) and ARMA (Auto Regressive and Moving Average) models | |
CN103176137B (en) | Battery pack health condition evaluation method based on battery SOC (state of charge) inhomogeneity | |
CN114444370A (en) | Method and device for predicting accumulated loss life of rechargeable battery by considering operation conditions, electronic equipment and readable storage medium | |
CN113687251B (en) | Double-model-based lithium ion battery pack voltage abnormality fault diagnosis method | |
CN105354768A (en) | Assessment method and system for electric transmission and transformation equipment states | |
CN106033114A (en) | Health state evaluation method of energy storage battery | |
CN113112114A (en) | Energy storage power station online evaluation method and device | |
CN115980612B (en) | Satellite battery pack health state evaluation method, system and equipment | |
CN111509319A (en) | PHM management system for energy storage power supply | |
CN112883634A (en) | DC measurement system state prediction method and system based on multi-dimensional analysis |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20131127 |