CN117031294A - Battery multi-fault detection method, device and storage medium - Google Patents
Battery multi-fault detection method, device and storage medium Download PDFInfo
- Publication number
- CN117031294A CN117031294A CN202310886095.0A CN202310886095A CN117031294A CN 117031294 A CN117031294 A CN 117031294A CN 202310886095 A CN202310886095 A CN 202310886095A CN 117031294 A CN117031294 A CN 117031294A
- Authority
- CN
- China
- Prior art keywords
- battery
- statistics
- correlation coefficient
- statistic
- fault
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 44
- 238000003860 storage Methods 0.000 title claims abstract description 13
- 238000000034 method Methods 0.000 claims abstract description 42
- HBBGRARXTFLTSG-UHFFFAOYSA-N Lithium ion Chemical class [Li+] HBBGRARXTFLTSG-UHFFFAOYSA-N 0.000 claims abstract description 20
- 229910001416 lithium ion Inorganic materials 0.000 claims abstract description 20
- 238000005259 measurement Methods 0.000 claims abstract description 14
- 230000002159 abnormal effect Effects 0.000 claims description 20
- 239000013598 vector Substances 0.000 claims description 19
- 238000004590 computer program Methods 0.000 claims description 17
- 230000001186 cumulative effect Effects 0.000 claims description 15
- 239000011159 matrix material Substances 0.000 claims description 15
- 238000004364 calculation method Methods 0.000 claims description 14
- 239000000178 monomer Substances 0.000 claims description 12
- 238000000926 separation method Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 3
- 238000012880 independent component analysis Methods 0.000 abstract description 13
- 238000000513 principal component analysis Methods 0.000 abstract description 10
- 238000003745 diagnosis Methods 0.000 abstract description 6
- 230000008569 process Effects 0.000 abstract description 6
- 238000000605 extraction Methods 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 9
- 230000006870 function Effects 0.000 description 6
- 230000000694 effects Effects 0.000 description 5
- 238000012545 processing Methods 0.000 description 5
- 230000008859 change Effects 0.000 description 4
- 238000009826 distribution Methods 0.000 description 4
- 238000000354 decomposition reaction Methods 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 230000005856 abnormality Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000001629 suppression Effects 0.000 description 2
- 206010000117 Abnormal behaviour Diseases 0.000 description 1
- 208000032953 Device battery issue Diseases 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
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
- 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]
- G01R31/378—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC] specially adapted for the type of battery or accumulator
-
- 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)
Abstract
The application relates to the technical field of battery fault diagnosis, in particular to a method, a device, equipment and a computer storage medium for detecting multiple faults of a battery. According to the battery multi-fault detection method, a cross voltage measurement topological structure of a lithium ion battery system is firstly established, correlation coefficients of two voltages are calculated based on the topological structure, then component extraction is performed on the high-dimensional non-Gaussian correlation coefficients in parallel based on a diagnosis model (ICA-PCA) combining Independent Component Analysis (ICA) and Principal Component Analysis (PCA), and comprehensive fault indexes are calculated, so that a fault detection process is simplified, short circuits and sensor faults are identified and positioned through a contribution graph method, fault detection efficiency is improved, and fault type identification and fault positioning are realized.
Description
Technical Field
The application relates to the technical field of battery fault diagnosis, in particular to a method, a device, equipment and a computer storage medium for detecting multiple faults of a battery.
Background
Correlation coefficient based methods are commonly used for battery fault detection, but currently in use, a single correlation coefficient is typically analyzed at a time. For lithium ion battery systems having a large number of cells, it is inefficient to monitor each correlation coefficient individually. Therefore, it is necessary to establish a parallel processing method with a high-dimensional non-gaussian correlation coefficient. In addition, in the traditional fault detection method based on the correlation coefficient, the problem of how to identify the fault type is not solved yet.
Disclosure of Invention
Therefore, the application aims to solve the technical problems of low efficiency and failure type identification in the prior art.
In order to solve the technical problems, the application provides a battery multi-fault detection method, which comprises the following steps:
establishing a cross voltage measurement topological structure of the lithium ion battery system, and calculating a recursion correlation coefficient between every two battery voltages based on the topological structure;
inputting the recursive correlation coefficient into an ICA-PCA model, extracting a non-Gaussian part, a Gaussian part and a noise part respectively, and calculating I of the non-Gaussian part 2 T of statistics, gaussian part 2 SPE statistics of statistics and noise parts;
estimating the I by using a nuclear density estimation method 2 Statistics, T 2 A threshold of statistics and SPE statistics, and according to the I 2 Statistics, T 2 Calculating a comprehensive statistic by using the statistic, the SPE statistic and a threshold value corresponding to each statistic;
and estimating the threshold value of the comprehensive statistic by adopting a kernel density estimation method, isolating abnormal correlation coefficient signals through a contribution graph if the comprehensive statistic exceeds the corresponding threshold value, and identifying the fault type and the positioning fault position according to the arrangement of the abnormal correlation coefficient signals.
Preferably, the establishing a lithium ion battery system cross voltage measurement topology includes:
grouping a plurality of monomers in the lithium ion battery system by taking n monomers as a group;
the n monomers in each group are connected in parallel, and each group of monomers are connected in series;
and setting a voltage sensor in parallel for every two adjacent monomer groups to obtain the cross voltage measurement topological structure of the lithium ion battery system.
Preferably, the formula for calculating the recursive correlation coefficient between the battery voltages based on the topology structure is as follows:
P k ＝P k-1 +x i y i -x i-w y i-w
Q k ＝Q k-1 +x i -x i-w
R k ＝R k-1 +y i -y i-w
wherein x and y are measured values of two adjacent voltage sensors, k is sampling time, w is window width, and P k Is x i y i -x i-w y i-w Cumulative sum of Q k Is x i -x i-w Is the cumulative sum of R k Is y i -y i-w Is the cumulative sum of S k Is thatIs the cumulative sum of T k Is->Cumulative sum of P 0 、Q 0 、R 0 、S 0 、T 0 Is constant, x i Is the ith value, y of the voltage sequence of a battery i Is the ith value, x of the voltage sequence of the adjacent cell i-w Is the i-w value, y of the voltage sequence of a battery i-w Is the i-w th value of the voltage sequence of the adjacent cell.
Preferably, said inputting said recursive correlation coefficients into an ICA-PCA model, extracting non-gaussian, gaussian and noise parts respectively comprises:
collecting the recursive correlation coefficient into a vector with x= [ C ] 1,2 ,C 2,3 ,...,C m,1 ] T M is the number of cells in the battery pack, C 1,2 Between the voltages of the first battery and the second batteryRelated recursion coefficient of C 2,3 C is the relative recursion coefficient between the voltages of the second battery and the third battery m,1 A relative recursion coefficient between the voltage of the last battery and the voltage of the first battery;
normalizing the vector and inputting the vector into an ICA model, and extracting a non-Gaussian partInputting the rest into PCA model, extracting Gauss part +.>And noise part->
Wherein W is s Is the non-gaussian part, x, of the separation matrix W new Calculating a new correlation coefficient signal vector after obtaining a new measured value for the voltage sensor on line, wherein P is the load of the residual part, I is an identity matrix, and A= (Λ) -1/2 U T ) - 1 B s U is a eigenvector matrix, Λ is a diagonal matrix composed of eigenvalues, B s Is lambda -1/2 U T The non-gaussian portion of a.
Preferably, said method according to said I 2 Statistics, T 2 The statistic, SPE statistic and threshold value corresponding to each statistic calculate the formula of the comprehensive statistic as follows:
wherein I is 2 Representation I 2 The statistics of the statistics are obtained,representation I 2 Threshold corresponding to statistics, T 2 Representing T 2 Statistics (1)/(>Representing T 2 Threshold corresponding to statistic, SPE represents SPE statistic, J SPE,th Representing the threshold value to which the SPE statistic corresponds.
Preferably, the I is estimated using a nuclear density estimation method 2 Statistics, T 2 The statistics and the threshold value of SPE statistics further comprise:
and searching for the optimal window width based on a trend curve of the threshold value estimated by the kernel density estimation method along with the window width.
Preferably, the isolating abnormal correlation coefficient signals through the contribution graph, and identifying the fault type and locating the fault position according to the arrangement of the abnormal correlation coefficient signals includes:
defining the contribution of the jth correlation coefficient signal to phi as:
wherein, xi j Is the column vector with the j element being 1 and the other elements being 0, phi being the comprehensive statistic parameter, x new Calculating a new correlation coefficient signal vector after obtaining a new measured value for the voltage sensor on line;
normalizing contributions of the plurality of correlation coefficient signals to phi:
calculating the contribution rate of each correlation coefficient signal:
if the j-th correlation coefficient signal contribution rate and the j-2-th correlation coefficient signal contribution rate are abnormal two correlation coefficient signal contribution rates, the j-th monomer has an internal short circuit fault;
if the j-th correlation coefficient signal contribution rate and the j-1-th correlation coefficient signal contribution rate are abnormal two correlation coefficient signal contribution rates, the j-th sensor fails.
The application also provides a device comprising:
the recursion correlation coefficient calculation module is used for establishing a cross voltage measurement topological structure of the lithium ion battery system and calculating recursion correlation coefficients between battery voltages based on the topological structure;
a statistic calculation module for inputting the recursive correlation coefficient into ICA-PCA model, extracting non-Gaussian part, gaussian part and noise part respectively, and calculating I of non-Gaussian part 2 T of statistics, gaussian part 2 SPE statistics of statistics and noise parts;
a comprehensive statistic calculation module for estimating the I by using nuclear density estimation method 2 Statistics, T 2 A threshold of statistics and SPE statistics, and according to the I 2 Statistics, T 2 Calculating a comprehensive statistic by using the statistic, the SPE statistic and a threshold value corresponding to each statistic;
and the fault detection module is used for estimating the threshold value of the comprehensive statistic by adopting a nuclear density estimation method, isolating abnormal correlation coefficient signals through a contribution graph if the comprehensive statistic exceeds the corresponding threshold value, and identifying fault types and positioning fault positions according to the arrangement of the abnormal correlation coefficient signals.
The application also provides a battery multi-fault detection device, comprising:
a memory for storing a computer program;
and the processor is used for realizing the steps of the battery multi-fault detection method when executing the computer program.
The present application also provides a computer readable storage medium having stored thereon a computer program which when executed by a processor implements the steps of a battery multi-fault detection method as described above.
Compared with the prior art, the technical scheme of the application has the following advantages:
according to the battery multi-fault detection method, a cross voltage measurement topological structure of a lithium ion battery system is firstly established, correlation coefficients of two voltages are calculated based on the topological structure, then component extraction is performed on the high-dimensional non-Gaussian correlation coefficients in parallel based on a diagnosis model (ICA-PCA) combining Independent Component Analysis (ICA) and Principal Component Analysis (PCA), and comprehensive fault indexes are calculated, so that a fault detection process is simplified, short circuits and sensor faults are identified and positioned through a contribution graph method, fault detection efficiency is improved, and fault type identification and fault positioning are realized.
Drawings
In order that the application may be more readily understood, a more particular description of the application will be rendered by reference to specific embodiments thereof which are illustrated in the appended drawings, in which:
FIG. 1 is a flow chart of an implementation of a method for detecting multiple battery faults;
FIG. 2 is a flowchart illustrating a method for detecting multiple battery failures according to an embodiment of the present application;
fig. 3 is a schematic diagram of the equivalent principle of a lithium ion battery system and a series battery and the topology of a voltage sensor of the battery;
FIG. 4 is a schematic diagram of two maximum contributions of an interval;
fig. 5 is a schematic diagram of two adjacent maximum contributions.
Detailed Description
The core of the application is to provide a method, a device, equipment and a computer storage medium for detecting multiple faults of a battery, so that the fault detection efficiency is improved, and the fault type identification and the fault positioning are realized.
In order to better understand the aspects of the present application, the present application will be described in further detail with reference to the accompanying drawings and detailed description. It will be apparent that the described embodiments are only some, but not all, embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
Referring to fig. 1 and fig. 2, fig. 1 is a flowchart illustrating an implementation of a method for detecting multiple faults of a battery according to the present application, and fig. 2 is a flowchart illustrating an implementation of a method for detecting multiple faults of a battery according to an embodiment of the present application; the specific operation steps are as follows:
s101, establishing a cross voltage measurement topological structure of a lithium ion battery system, and calculating a recursion correlation coefficient between battery voltages based on the topological structure;
s102, inputting the recursion correlation coefficient into an ICA-PCA model, respectively extracting a non-Gaussian part, a Gaussian part and a noise part, and calculating I of the non-Gaussian part 2 T of statistics, gaussian part 2 SPE statistics of statistics and noise parts;
s103, estimating the I by adopting a nuclear density estimation method 2 Statistics, T 2 A threshold of statistics and SPE statistics, and according to the I 2 Statistics, T 2 Calculating a comprehensive statistic by using the statistic, the SPE statistic and a threshold value corresponding to each statistic;
and S104, estimating a threshold value of the comprehensive statistic by adopting a kernel density estimation method, isolating abnormal correlation coefficient signals through a contribution graph if the comprehensive statistic exceeds a corresponding threshold value, and identifying fault types and positioning fault positions according to the arrangement of the abnormal correlation coefficient signals.
According to the battery multi-fault detection method, a cross voltage measurement topological structure of a lithium ion battery system is firstly established, correlation coefficients of two voltages are calculated based on the topological structure, then component extraction is performed on the parallel of the high-dimensional non-Gaussian correlation coefficients based on a diagnosis model (ICA-PCA) combining Independent Component Analysis (ICA) and Principal Component Analysis (PCA), and comprehensive fault indexes are calculated, so that a fault detection process is simplified, short circuits and sensor faults are identified and positioned through a contribution graph method, fault detection efficiency is improved, and fault type identification and fault positioning are realized.
Based on the above embodiments, the present embodiment describes step S101 in detail:
lithium ion battery systems may contain hundreds or even thousands of battery cells, and it is therefore difficult to configure a sensor for each cell. One common method of organizing the cells in a lithium ion battery system is to first connect them in parallel and then connect them in series so that there are multiple cells in each group, as shown in fig. 3. Thus, a lithium ion battery system may be equivalent to a series-connected battery. Due to the self-balancing effect, multiple cells in the same group have relatively consistent electrical behavior and aging processes, so parallel cells are typically managed as a whole. Therefore, only one sensor needs to be configured for a group of batteries connected in parallel. However, since the abnormal behavior of multiple faults on the voltage signal is very similar, the faulty battery cannot be effectively isolated and identified by a simple single voltage sensor topology. The present application addresses this problem with a cross-cell sensor topology as shown in fig. 3. Each voltage sensor measures the voltage of two adjacent cells, while adjacent voltage sensors cross-share one cell. Different fault types will trigger individual or simultaneous changes of different sensors, providing a basis for identifying multiple faults.
The use of the voltage correlation coefficient instead of the voltage measurement value for abnormality detection can effectively avoid the problem of erroneous judgment caused by inconsistency between the battery cells. In a cross-cell sensor topology, each voltage sensor measures the voltage of two adjacent equivalent cells, labeled V j (j=1,..m, m is the number of cells), where the last sensor measures the m-th cell and the first cell. The correlation coefficient is calculated as follows:
P k ＝P k-1 +x i y i -x i-w y i-w
Q k ＝Q k-1 +x i -x i-w
R k ＝R k-1 +y i -y i-w
wherein x and y are measured values of two adjacent voltage sensors, k is sampling time, w is window width, and P k Is x i y i -x i-w y i-w Cumulative sum of Q k Is x i -x i-w Is the cumulative sum of R k Is y i -y i-w Is the cumulative sum of S k Is thatIs the cumulative sum of T k Is->Cumulative sum of P 0 、Q 0 、R 0 、S 0 、T 0 Is constant, the specific value is determined according to the calculation requirement, x i Is the ith value, y of the voltage sequence of a battery i Is the ith value, x of the voltage sequence of the adjacent cell i-w Is the i-w value, y of the voltage sequence of a battery i-w Is the i-w th value of the voltage sequence of the adjacent cell.
Based on the above embodiments, the present embodiment describes step S102 in detail:
all correlation coefficient signals are collected into a vector x= [ C ] 1,2 ,C 2,3 ,...,C m,1 ] T For parallel processing, m is the number of cells in the battery pack, C 1,2 C is the relative recursion coefficient between the voltages of the first battery and the second battery 2,3 To the relative recursion coefficient between the voltages of the second and third batteries, and so on, C m,1 Is the relative recursion coefficient between the last cell and the first cell voltage. x should be normalized before use, i.e. processed as vector samples with unity variance and zero mean. x is a strong non-gaussian distribution, which is first input into the ICA model to extract its non-gaussian portion. By ICA, x can be decomposed into unknown independent components, s 1 ,s 2 ,…,s γ Namely:
x＝As+e
wherein s= [ s ] 1 ,s 2 ,…,s γ ] T ∈R γ .A＝[a 1 ,a 2 ,…,a γ ]∈R m×γ Is a mixed matrix, e.epsilon.R m Is residual error, gamma is the component quantity, and gamma is less than or equal to m. ICA needs to estimate a and s from x, where ICA aims at finding a separation matrix W to reconstruct data whose components are as independent as possible, and the reconstructed data is represented by the following formula, where we need to find the separation matrix W in the following formula:
for the random vector x, the covariance of the random vector x is subjected to eigenvalue decomposition R x ＝E(xx T ) (E represents the desire):
R x ＝UΛU T
wherein U is a eigenvector matrix, Λ is a diagonal matrix composed of eigenvalues, and defined vectors z and B are:
z＝Λ -1/2 U T x≈Λ -1/2 U T As＝Bs
B＝Λ -1/2 U T A
s can be reconstructed as:
calculating W:
W＝B T Λ -1/2 U T
m independent components can be obtained by ICA, from which it is necessary to further select γ dominant components. Using euclidean norms (L 2 ) Standard: the order of the individual components is determined by the Euclidean norm (L 2 ) And (5) determining. After reordering, B and W are adjusted to:
after the non-gaussian portion is extracted, the remaining portion (e) is approximated as a gaussian distribution, which is further input to a PCA model to extract the principal component (gaussian main portion) and the residual (gaussian noise portion). The remainder of ICA extraction is assembled into matrix E, then matrix E of the input PCA model is:
where P and T are the load and score matrices of E and F is the remainder.
After PCA decomposition, e is split into two parts:
wherein,main component space belonging to the Gaussian part main signal, < ->Belonging to a residual subspace reflecting gaussian partial noise.
Once the voltage is transferredThe sensor acquires a new measured value on line, calculates and acquires a new correlation coefficient signal vector x new Normalization is required according to the mean value and standard deviation of training samples, and then decomposition is performed into three parts through an ICA-PCA model: non-gaussian partGauss part->And noise part->
Wherein a= (Λ) -1/2 U T ) -1 B s
The non-Gaussian portion is generally represented by I 2 Statistics, while the remaining two are typically represented by T 2 Statistics and SPE statistics:
based on the above embodiments, the present embodiment describes in detail step S103:
the threshold of statistics is determined by kernel density estimation, at I 2 Statistics are as examples:
I 2 the probability density function of the statistic is f p (I 2 ) The threshold isAmong these statistics. Due toThreshold->For a given confidence level limit α, the following is calculated:
(1) Get I when N battery systems are normal 2 Statistics and kernel density estimation to obtain I 2 Statistic minimum valueI 2 Maximum value of statistics->Step size ΔI 2 Step number k, and probability density function of the distribution interval of these statistics.
(2) According to the calculation principle of the trapezoidal method, the total probability covered by the probability density curve is as follows:
(3) For a given α, at a step ΔI 2 Period searchUntil the following conditions are satisfied:
at this time, theSet to threshold +.>
T 2 Threshold of statistics and SPE statisticsAnd J SPE,th And may be calculated in the same manner. In the practical application process, in order to simplify the fault detection logic, the statistics are integrated into a comprehensive statistic phi and a comprehensive statistic parameter phi:
and (3) with
Wherein:
H 3 ＝(I-AW s ) T (I-PP T )(I-AW s )
threshold J φ,th The corresponding phi is also determined by an algorithm based on the kernel density estimation.
The window width w has a large effect on the correlation coefficient signal, a smaller w amplifies the effect of noise and interference, and an excessive w reduces the sensitivity of the correlation coefficient signal to anomalies. Currently w is determined mainly by empirical values. To better address this problem, in one embodiment, the present application establishes a semi-quantitative preference criterion, i.e., using a threshold based on kernel density estimation to coarsely quantify the change in correlation coefficients under different window widths in a fault-free condition, and determines an optimal threshold based on a trend curve of the threshold evolution with window width. The specific criteria are as follows:
since the threshold based on the kernel density estimation can be automatically adjusted according to the distribution of the training data, the false alarm rate under the condition of no fault does not exceed the limit (1-alpha) multiplied by 100 percent. Thus, a threshold based on the kernel density estimate can be used to reflect the change in correlation coefficient at different w during normal operation, given a. W corresponding to the minimum threshold means the optimal suppression of noise and interference by the correlation coefficient at the window width. Therefore, the above embodiment can be used to describe the step S104 in detail using the threshold-dependent window based on the kernel density estimation:
the contribution graph method is a relatively common abnormality diagnosis method at present. The contribution graph method is obtained by calculating the contribution rate of each process variable to statistics in principal component space and residual space and drawing a histogram, and is a reflection of the influence degree of the change of each variable on the stability of the system statistical model. The application improves the contribution graph method to inhibit tailing effect so as to accurately isolate problematic correlation coefficient signals after faults occur.
So long as phi exceeds its threshold J φ,th The arrangement of the isolated problematic correlation coefficient signals (spaced or adjacent) is used to identify the fault type and locate the faulty cell/sensor.
Once phi exceeds J φ,th Then it is considered that a fault is detected and in order to identify the type of fault, it is necessary to isolate the problematic correlation coefficient signal by means of a contribution graph:
defining the contribution of the jth correlation coefficient signal to phi as:
wherein, xi j Is the column vector with the j-th element being 1 and the other elements being 0, and phi is the comprehensive statistic parameter.
The larger the more likely the jth correlation coefficient is problematic, in order to enhance the suppression of shift effects, the following is done:
first, based on training data x n For a pair ofNormalized get->So that the contribution of all correlation coefficients to phi is in the same range:
then, the contribution rate of each correlation coefficient signal is calculated:
ranging from 0 to 1, the average contribution rate (1/m) is a suitable threshold, and after isolating the problematic correlation coefficient signal, the type of current fault can be identified according to the following fault identification logic:
short circuit fault: if the j (j=1, 2,., m) th cell fails with an internal short circuit, thenThe j-th and j-1 th voltage sensors will change at the same time, resulting in C j-2,j-1 And C j,j+1 Exception while other values remain unchanged. The contribution graph visually shows the two largest correlation coefficient contribution values of the interval, as in fig. 4.
Sensor failure: if the j (j=1, 2,., m) th sensor fails, the voltage measured by the j-th voltage sensor changes, resulting in C j-1,j And C j,j+1 Exception while other values remain unchanged as in fig. 5.
The embodiment of the application also provides a battery multi-fault detection device; the specific apparatus may include:
the recursion correlation coefficient calculation module is used for establishing a cross voltage measurement topological structure of the lithium ion battery system and calculating recursion correlation coefficients between battery voltages based on the topological structure;
a statistic calculation module for inputting the recursive correlation coefficient into ICA-PCA model, extracting non-Gaussian part, gaussian part and noise part respectively, and calculating I of non-Gaussian part 2 T of statistics, gaussian part 2 SPE statistics of statistics and noise parts;
a comprehensive statistic calculation module for estimating the I by using nuclear density estimation method 2 Statistics, T 2 A threshold of statistics and SPE statistics, and according to the I 2 Statistics, T 2 Calculating a comprehensive statistic by using the statistic, the SPE statistic and a threshold value corresponding to each statistic;
and the fault detection module is used for estimating the threshold value of the comprehensive statistic by adopting a nuclear density estimation method, isolating abnormal correlation coefficient signals through a contribution graph if the comprehensive statistic exceeds the corresponding threshold value, and identifying fault types and positioning fault positions according to the arrangement of the abnormal correlation coefficient signals.
The battery multi-fault detection device of the present embodiment is configured to implement the foregoing battery multi-fault detection method, so that the specific implementation of the battery multi-fault detection device may be the example portions of the foregoing battery multi-fault detection method, for example, the recursive correlation coefficient calculation module, the statistics calculation module, the comprehensive statistics calculation module, and the fault detection module are respectively configured to implement steps S101, S102, S103, and S104 in the foregoing battery multi-fault detection method, so that the specific implementation thereof may refer to the descriptions of the corresponding respective portion examples and will not be repeated herein.
The embodiment of the application also provides a battery multi-fault detection device, which comprises: a memory for storing a computer program; and the processor is used for realizing the steps of the battery multi-fault detection method when executing the computer program.
The specific embodiment of the application also provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program realizes the steps of the battery multi-fault detection method when being executed by a processor.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It is apparent that the above examples are given by way of illustration only and are not limiting of the embodiments. Other variations and modifications of the present application will be apparent to those of ordinary skill in the art in light of the foregoing description. It is not necessary here nor is it exhaustive of all embodiments. While still being apparent from variations or modifications that may be made by those skilled in the art are within the scope of the application.
Claims (10)
1. A battery multi-fault detection method, comprising:
establishing a cross voltage measurement topological structure of the lithium ion battery system, and calculating a recursion correlation coefficient between every two battery voltages based on the topological structure;
inputting the recursive correlation coefficient into an ICA-PCA model, extracting a non-Gaussian part, a Gaussian part and a noise part respectively, and calculating I of the non-Gaussian part 2 T of statistics, gaussian part 2 SPE statistics of statistics and noise parts;
estimating the I by using a nuclear density estimation method 2 Statistics, T 2 A threshold of statistics and SPE statistics, and according to the I 2 StatisticsAmount, the T 2 Calculating a comprehensive statistic by using the statistic, the SPE statistic and a threshold value corresponding to each statistic;
and estimating the threshold value of the comprehensive statistic by adopting a kernel density estimation method, isolating abnormal correlation coefficient signals through a contribution graph if the comprehensive statistic exceeds the corresponding threshold value, and identifying the fault type and the positioning fault position according to the arrangement of the abnormal correlation coefficient signals.
2. The battery multi-fault detection method of claim 1, wherein establishing a lithium ion battery system crossover voltage measurement topology comprises:
grouping a plurality of monomers in the lithium ion battery system by taking n monomers as a group;
the n monomers in each group are connected in parallel, and each group of monomers are connected in series;
and setting a voltage sensor in parallel for every two adjacent monomer groups to obtain the cross voltage measurement topological structure of the lithium ion battery system.
3. The battery multi-fault detection method according to claim 2, wherein the formula for calculating the recursive correlation coefficient between the battery voltages based on the topology is:
P k ＝P k-1 +x i y i -x i-w y i-w
Q k ＝Q k-1 +x i -x i-w
R k ＝R k-1 +y i -y i-w
wherein x and y are measured values of two adjacent voltage sensors, k is sampling time, w is window width, and P k Is x i y i -x i- w y i-w Cumulative sum of Q k Is x i -x i-w Is the cumulative sum of R k Is y i -y i-w Is the cumulative sum of S k Is thatIs the cumulative sum of T k Is->Cumulative sum of P 0 、Q 0 、R 0 、S 0 、T 0 Is constant, x i Is the ith value, y of the voltage sequence of a battery i Is the ith value, x of the voltage sequence of the adjacent cell i-w Is the i-w value, y of the voltage sequence of a battery i-w Is the i-w th value of the voltage sequence of the adjacent cell.
4. The battery multi-fault detection method according to claim 1, wherein said inputting the recursive correlation coefficients into an ICA-PCA model, extracting non-gaussian, gaussian and noise parts, respectively, comprises:
collecting the recursive correlation coefficient into a vector with x= [ C ] 1,2 ,C 2,3 ,...,C m,1 ] T M is the number of cells in the battery pack, C 1,2 C is the relative recursion coefficient between the voltages of the first battery and the second battery 2,3 C is the relative recursion coefficient between the voltages of the second battery and the third battery m,1 A relative recursion coefficient between the voltage of the last battery and the voltage of the first battery;
normalizing the vector, inputting the vector into an ICA model, and extracting non-componentsGaussian partInputting the rest into PCA model, extracting Gauss part +.>And noise part->
Wherein W is s Is the non-gaussian part, x, of the separation matrix W new Calculating a new correlation coefficient signal vector after obtaining a new measured value for the voltage sensor on line, wherein P is the load of the residual part, I is an identity matrix, and A= (Λ) -1/2 U T ) -1 B s U is a eigenvector matrix, Λ is a diagonal matrix composed of eigenvalues, B s Is lambda -1/2 U T The non-gaussian portion of a.
5. The battery multi-fault detection method according to claim 1, wherein the step of detecting the I is performed by 2 Statistics, T 2 The statistic, SPE statistic and threshold value corresponding to each statistic calculate the formula of the comprehensive statistic as follows:
wherein I is 2 Representation I 2 The statistics of the statistics are obtained,representation I 2 Threshold corresponding to statistics, T 2 Representing T 2 Statistics (1)/(>Representing T 2 Threshold corresponding to statistic, SPE represents SPE statistic, J SPE,th Representing the threshold value to which the SPE statistic corresponds.
6. The battery multi-fault detection method of claim 1, wherein the I is estimated using a nuclear density estimation method 2 Statistics, T 2 The statistics and the threshold value of SPE statistics further comprise:
and searching for the optimal window width based on a trend curve of the threshold value estimated by the kernel density estimation method along with the window width.
7. The battery multi-fault detection method according to claim 1, wherein the isolating abnormal correlation coefficient signals by the contribution graph and identifying the fault type and locating the fault location according to the arrangement of the abnormal correlation coefficient signals comprises:
defining the contribution of the jth correlation coefficient signal to phi as:
wherein, xi j Is the column vector with the j element being 1 and the other elements being 0, phi being the comprehensive statistic parameter, x new Calculating a new correlation coefficient signal vector after obtaining a new measured value for the voltage sensor on line;
normalizing contributions of the plurality of correlation coefficient signals to phi:
calculating the contribution rate of each correlation coefficient signal:
if the j-th correlation coefficient signal contribution rate and the j-2-th correlation coefficient signal contribution rate are abnormal two correlation coefficient signal contribution rates, the j-th monomer has an internal short circuit fault;
if the j-th correlation coefficient signal contribution rate and the j-1-th correlation coefficient signal contribution rate are abnormal two correlation coefficient signal contribution rates, the j-th sensor fails.
8. A battery multi-fault detection device, comprising:
the recursion correlation coefficient calculation module is used for establishing a cross voltage measurement topological structure of the lithium ion battery system and calculating recursion correlation coefficients between battery voltages based on the topological structure;
a statistic calculation module for inputting the recursive correlation coefficient into ICA-PCA model, extracting non-Gaussian part, gaussian part and noise part respectively, and calculating I of non-Gaussian part 2 T of statistics, gaussian part 2 SPE statistics of statistics and noise parts;
a comprehensive statistic calculation module for estimating the I by using nuclear density estimation method 2 Statistics, T 2 A threshold of statistics and SPE statistics, and according to the I 2 Statistics, T 2 Calculating a comprehensive statistic by using the statistic, the SPE statistic and a threshold value corresponding to each statistic;
and the fault detection module is used for estimating the threshold value of the comprehensive statistic by adopting a nuclear density estimation method, isolating abnormal correlation coefficient signals through a contribution graph if the comprehensive statistic exceeds the corresponding threshold value, and identifying fault types and positioning fault positions according to the arrangement of the abnormal correlation coefficient signals.
9. A battery multi-fault detection apparatus, characterized by comprising:
a memory for storing a computer program;
a processor for implementing the steps of a battery multi-fault detection method as claimed in any one of claims 1 to 7 when executing the computer program.
10. A computer readable storage medium, characterized in that the computer readable storage medium has stored thereon a computer program which, when executed by a processor, implements the steps of a battery multi-fault detection method according to any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310886095.0A CN117031294A (en) | 2023-07-18 | 2023-07-18 | Battery multi-fault detection method, device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310886095.0A CN117031294A (en) | 2023-07-18 | 2023-07-18 | Battery multi-fault detection method, device and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117031294A true CN117031294A (en) | 2023-11-10 |
Family
ID=88601392
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310886095.0A Pending CN117031294A (en) | 2023-07-18 | 2023-07-18 | Battery multi-fault detection method, device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117031294A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117310543A (en) * | 2023-11-29 | 2023-12-29 | 中国华能集团清洁能源技术研究院有限公司 | Battery abnormality diagnosis method and device |
-
2023
- 2023-07-18 CN CN202310886095.0A patent/CN117031294A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117310543A (en) * | 2023-11-29 | 2023-12-29 | 中国华能集团清洁能源技术研究院有限公司 | Battery abnormality diagnosis method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111931868B (en) | Time series data abnormity detection method and device | |
| US20040153815A1 (en) | Methodology for temporal fault event isolation and identification | |
| CN104216350A (en) | System and method for analyzing sensed data | |
| CN112147512A (en) | Diagnosis and separation method for short-circuit and abuse faults of lithium ion battery | |
| CN111967535B (en) | Fault diagnosis method and device for temperature sensor of grain storage management scene | |
| CN117031294A (en) | Battery multi-fault detection method, device and storage medium | |
| CN109800446B (en) | Method and device for estimating voltage inconsistency in discharging process of lithium ion battery | |
| CN112966879A (en) | Environmental test chamber fault prediction method and device, computer equipment and storage medium | |
| CN108717496B (en) | Radar antenna array surface fault detection method and system | |
| CN115327417A (en) | Early warning method and system for abnormity of power battery monomer and electronic equipment | |
| CN115800272A (en) | Power grid fault analysis method, system, terminal and medium based on topology identification | |
| CN110221590B (en) | Industrial process multi-fault diagnosis method based on discriminant analysis | |
| CN109116243A (en) | A kind of method for detecting abnormality and device of power battery | |
| CN112464297B (en) | Hardware Trojan detection method, device and storage medium | |
| CN116400248B (en) | Method and device for determining short circuit fault in battery | |
| Won et al. | Performance analysis of machine learning based fault detection for cloud infrastructure | |
| CN115659271A (en) | Sensor abnormality detection method, model training method, system, device, and medium | |
| CN115629323A (en) | Battery pack fault detection method and device and vehicle | |
| CN109471766B (en) | Sequential fault diagnosis method and device based on testability model | |
| CN113721121A (en) | Fault detection method and device for semiconductor process | |
| CN112148536A (en) | Method and device for detecting deep learning chip, electronic equipment and computer storage medium | |
| CN117572159B (en) | Power failure detection method and system based on big data analysis | |
| CN115308631B (en) | Fault diagnosis method and system for new energy automobile power battery pack | |
| CN110542851A (en) | Fault diagnosis method and device for circuit breaker operating mechanism, computer and storage medium | |
| CN116992308B (en) | Data and knowledge fusion process fluctuation analysis and optimization method and device |
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 |