CN116451038A - Power battery thermal runaway early warning method and system based on singular spectrum entropy - Google Patents
Power battery thermal runaway early warning method and system based on singular spectrum entropy Download PDFInfo
- Publication number
- CN116451038A CN116451038A CN202310308492.XA CN202310308492A CN116451038A CN 116451038 A CN116451038 A CN 116451038A CN 202310308492 A CN202310308492 A CN 202310308492A CN 116451038 A CN116451038 A CN 116451038A
- Authority
- CN
- China
- Prior art keywords
- power battery
- singular spectrum
- singular
- spectrum entropy
- entropy
- 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
- 238000001228 spectrum Methods 0.000 title claims abstract description 158
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000010606 normalization Methods 0.000 claims abstract description 60
- 239000011159 matrix material Substances 0.000 claims abstract description 52
- 238000012545 processing Methods 0.000 claims abstract description 52
- 239000000178 monomer Substances 0.000 claims abstract description 48
- 108010076504 Protein Sorting Signals Proteins 0.000 claims abstract description 45
- 238000004458 analytical method Methods 0.000 claims abstract description 38
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 24
- 238000010183 spectrum analysis Methods 0.000 claims abstract description 15
- 230000008054 signal transmission Effects 0.000 claims abstract description 5
- 238000004364 calculation method Methods 0.000 claims description 9
- 238000010276 construction Methods 0.000 claims description 6
- 230000000877 morphologic effect Effects 0.000 claims description 6
- 230000005484 gravity Effects 0.000 claims description 3
- 238000012795 verification Methods 0.000 claims description 3
- 238000011158 quantitative evaluation Methods 0.000 abstract description 5
- 230000035945 sensitivity Effects 0.000 abstract description 5
- 230000002159 abnormal effect Effects 0.000 description 9
- 230000003595 spectral effect Effects 0.000 description 5
- 238000005070 sampling Methods 0.000 description 4
- 238000012360 testing method Methods 0.000 description 3
- 230000001052 transient effect Effects 0.000 description 3
- 238000001514 detection method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004146 energy storage Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/60—Other road transportation technologies with climate change mitigation effect
- Y02T10/70—Energy storage systems for electromobility, e.g. batteries
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Secondary Cells (AREA)
Abstract
The invention provides a power battery thermal runaway early warning method and system based on singular spectrum entropy, which comprises the steps of firstly acquiring output signals of sensors of a power battery by a vehicle end, constructing a mode matrix according to a time domain signal sequence formed by a time sequence of signal transmission, then carrying out singular value decomposition on the mode matrix, arranging singular values to form a singular spectrum of the signal, then carrying out singular spectrum analysis on the signal, further calculating to obtain singular spectrum entropy of the time domain signal sequence, carrying out singular spectrum entropy quantitative evaluation alarm coefficient of normalization processing, comparing the alarm coefficient with an alarm threshold value, and thus judging the thermal runaway risk of the power battery and positioning the fault position. According to the invention, analysis is performed from the angle of information content in the signal, the singular spectrum entropy value is utilized to early warn the thermal runaway of the power battery, so that the sensitivity to fault signals is greatly improved, the moment of risk occurrence can be detected more rapidly, the fault monomers can be positioned, and the time point of the thermal runaway early warning is effectively advanced.
Description
Technical Field
The invention relates to the technical field of power batteries, in particular to a power battery thermal runaway early warning method and system based on singular spectrum entropy.
Background
To alleviate the increasingly serious environmental problems and the scarcity of non-renewable energy sources, the motorization of automobile power systems has become one of the main development directions of automobiles in the future. The power battery is also the only power source as the energy storage device of the electric automobile, and plays a central role in the development of the electric automobile. However, during the use of the power battery, safety accidents often occur, but the accidents of the power battery generally occur in the form of thermal runaway, which mainly show that the temperature of the battery cell suddenly rises to cause fire, gas leakage, explosion and the like, and the casualties and property losses are extremely easy to cause.
The existing method for early warning the thermal runaway risk mainly is based on a temperature threshold value or a temperature change rate threshold value, and sends out an alarm electric signal after the temperature or the temperature change rate of the power battery exceeds the set temperature threshold value or the temperature change rate threshold value, so that the thermal runaway early warning is executed, but when the temperature or the temperature change rate of the power battery exceeds the set temperature threshold value or the temperature change rate threshold value, the thermal runaway is basically unavoidable, and the loss is difficult to recover, so that the scheme has great limitation. In addition, the conventional signal feature extraction method has the following disadvantages: (1) It is difficult to accurately describe the nonlinear characteristics of the system and the non-stationary signal characteristics of the signal, and (2) it is difficult to solve the quantitative evaluation problem of the signal characteristics. Therefore, a power battery thermal runaway early warning scheme capable of accurately and quantitatively evaluating signal thermal runaway characteristics and carrying out early warning more timely is needed to be proposed.
Disclosure of Invention
In order to solve the problems that the prior power battery thermal runaway early warning is not timely enough, a nonlinear system is difficult to accurately describe, quantitative evaluation of signal characteristics is difficult to solve and the like, the invention provides a power battery thermal runaway early warning method based on singular spectrum entropy. The invention also relates to a power battery thermal runaway early warning system based on singular spectrum entropy.
The technical scheme of the invention is as follows:
a power battery thermal runaway early warning method based on singular spectrum entropy is characterized by comprising the following steps:
a step of constructing a mode matrix, in which the output signals of the sensors of the power battery are collected and fed by a vehicle end, and the mode matrix is constructed according to a time domain signal sequence formed by the time sequence of signal feeding;
singular value decomposition, namely performing singular value decomposition on the pattern matrix, and arranging singular values to form a singular spectrum of the signal, wherein the number of non-zero singular values is the total number of patterns contained in each column of the pattern matrix;
calculating singular spectrum entropy, namely performing singular spectrum analysis of the signal, calculating the proportion of each mode in all modes, and further calculating the singular spectrum entropy of the time domain signal sequence to reflect the uncertainty degree of each mode of the time domain signal sequence under singular spectrum division and quantitatively describe the time domain complex morphological characteristics of signal energy distribution;
and a normalization processing early warning analysis step, wherein the obtained singular spectrum entropy of the time domain signal sequence is subjected to normalization processing based on comparison with the singular spectrum entropy of white noise, and an alarm coefficient is quantitatively evaluated according to the singular spectrum entropy after normalization processing and is compared with an alarm threshold value, so that the thermal runaway risk of the power battery is judged and the fault position is positioned.
Preferably, in the step of constructing the pattern matrix, the pattern matrix is constructed by setting an analysis window length and a delay constant and intercepting the time domain signal sequence in window order.
Preferably, in the step of calculating the singular spectrum entropy, the information entropy of the singular spectrum is calculated according to the definition of shannon entropy based on the proportion of each mode in all modes, so as to obtain the singular spectrum entropy of the time domain signal sequence.
Preferably, in the step of normalization processing early warning analysis, a Z-score method is adopted to normalize singular spectrum entropy after normalization processing, a difference value between the singular spectrum entropy after normalization processing of a certain power battery monomer and an average value of the singular spectrum entropy after normalization of all power battery monomers is divided by a standard deviation of the singular spectrum entropy after normalization of all power battery monomers to obtain an alarm coefficient of each power battery monomer, and the alarm coefficient of each power battery monomer is compared with an alarm threshold value, so that a thermal runaway risk of the power battery is judged and a fault position is located.
Preferably, in the step of constructing the mode matrix, the output signals of the sensors of the power battery include voltage, current and temperature signals under a plurality of actual working conditions or in a charging state when the vehicle is running.
Preferably, in the step of normalization processing early warning analysis, analysis and verification adjustment are performed based on driving data of a power battery cloud big data platform so as to determine an alarm threshold range.
A power battery thermal runaway early warning system based on singular spectrum entropy is characterized by comprising a construction mode matrix module, a singular value decomposition module, a singular spectrum entropy calculation module and a normalization processing early warning analysis module which are connected in sequence,
the mode matrix constructing module is used for acquiring and uploading output signals of each sensor of the driving force battery by a vehicle end and constructing a mode matrix according to a time domain signal sequence formed by a time sequence of signal transmission;
the singular value decomposition module is used for carrying out singular value decomposition on the pattern matrix, arranging singular values to form a singular spectrum of the signal, wherein the number of non-zero singular values is the total number of patterns contained in each column of the pattern matrix;
the singular spectrum entropy calculating module is used for carrying out singular spectrum analysis on signals and calculating the proportion of each mode in all modes, so as to calculate singular spectrum entropy of the time domain signal sequence, and reflect the uncertainty degree of each mode of the time domain signal sequence under singular spectrum division and quantitatively describe the time domain complex morphological characteristics of signal energy distribution;
the normalization processing early warning analysis module performs normalization processing on the obtained singular spectrum entropy of the time domain signal sequence based on comparison with the singular spectrum entropy of white noise, quantitatively evaluates an alarm coefficient according to the singular spectrum entropy after normalization processing and compares the alarm coefficient with an alarm threshold value, thereby judging the thermal runaway risk of the power battery and positioning the fault position.
Preferably, the construction pattern matrix module constructs the pattern matrix by setting an analysis window length and a delay constant and intercepting the time domain signal sequence in window order.
Preferably, the singular spectrum entropy calculating module calculates the information entropy of the singular spectrum according to the definition of shannon entropy based on the proportion of each mode in all modes, so as to obtain the singular spectrum entropy of the time domain signal sequence.
Preferably, the normalization processing early warning analysis module adopts a Z-score method to normalize singular spectrum entropy after normalization processing, divides the difference value of the singular spectrum entropy after normalization processing of a certain power battery monomer and the average value of the singular spectrum entropy of normalization of all power battery monomers by the standard deviation of the singular spectrum entropy of normalization of all power battery monomers to obtain the alarm coefficient of each power battery monomer, and compares the alarm coefficient of each power battery monomer with an alarm threshold value, thereby judging the thermal runaway risk of the power battery and positioning the fault position.
The beneficial effects of the invention are as follows:
the invention provides a power battery thermal runaway early warning method based on singular spectrum entropy, which is an analysis method combining singular spectrum and information entropy, wherein a singular value decomposition result is used as an information entropy solving object, namely, the information entropy of the singular spectrum (singular spectrum entropy for short) is calculated, and the method is particularly suitable for application occasions with fewer sampling points and noise in acquired signals. The singular spectrum analysis of signals is a modern spectrum analysis technology based on dynamic analysis, and the basic idea is to acquire the intrinsic complexity characteristics of the power battery system by carrying out phase space reconstruction and singular value decomposition on the time domain signal sequence of the power battery system. On the basis of singular spectrum analysis of signals, singular spectrum entropy is calculated, so that complex state characteristics of a time sequence can be quantitatively described. In the information theory, the information entropy is a description of the uncertainty degree of the system and is used for measuring the change condition of the signal and the size of the information content. The singular spectrum entropy is an information quantity measure value essentially, reflects the uncertainty degree of each mode of a time domain signal sequence under singular spectrum division, and also reflects the time domain complexity of signal energy distribution. Under normal working conditions, the complexity of the signal is small, the more the energy is concentrated in a few modes, and the singular spectrum entropy value is small; in contrast, when the battery monomer is in thermal runaway, the complexity of the incoming signal is increased, the energy of the complex transient signal is dispersed, the singular spectrum entropy is larger, the singular spectrum entropy is very sensitive to abnormal changes of the input signal, the current fault monomer can be rapidly detected and positioned, compared with the traditional algorithm for judging whether the data such as voltage, temperature and the like exceeds the threshold value by directly setting the threshold value, the thermal runaway alarm of the battery can be greatly advanced, precious time is strived for fault processing, and the safety is greatly improved. Therefore, the invention can rapidly reflect the internal information change of the power battery system and the characteristic of independent signal stability by utilizing singular spectrum entropy, and can rapidly judge whether the working state of the current battery is abnormal or not.
The invention constructs the mode matrix by combining the information acquired and uploaded by the sensor in the step of constructing the mode matrix and forming a time domain signal sequence according to the time sequence of signal transmission, realizes phase space reconstruction, carries out singular value decomposition by the step of singular value decomposition to form a singular spectrum of the signal, calculates the singular spectrum entropy in the step of calculating the singular spectrum entropy on the basis of the singular spectrum analysis of the signal, detects fault information by utilizing the property of the information entropy at the angle of information quantity, describes out-of-control information by using an information measurement degree, and can carry out early recognition and early warning of thermal runaway by taking the information quantity measurement as a characteristic. The information acquisition at the vehicle end is inevitably interfered by noise, the sampling frequency is limited by performance, compared with the laboratory condition, the method is lower, and the singular spectrum entropy is just suitable for the application occasions with fewer sampling points and noise in the acquired signals. The value of the singular spectrum entropy has obvious change when abnormal information exists, and whether the working state of the current battery is abnormal or not can be rapidly judged according to the change characteristics. The method comprises the steps of carrying out normalization processing on the singular spectrum entropy of the obtained time domain signal sequence based on comparison with the singular spectrum entropy of white noise, quantitatively evaluating an alarm coefficient according to the singular spectrum entropy after normalization processing, defining a parameter alarm coefficient based on the singular spectrum entropy value, describing the internal information of a power battery through the value of the alarm coefficient, quantitatively describing the complex state characteristics of a time sequence, quantifying the thermal runaway risk by utilizing the alarm coefficient parameter, comparing the alarm coefficient value of each power battery monomer with a preset alarm threshold value to determine whether the thermal runaway state occurs in the battery monomer and position the battery monomer, thereby quickly and accurately finding out the power battery with the thermal runaway fault, quantitatively analyzing the power battery with the thermal runaway fault, quickly positioning the power battery monomer with the abnormality at the current moment when reminding the occurrence of the thermal runaway risk at the certain moment, greatly improving the sensitivity of the system to the fault signal, quickly detecting the occurrence of the risk and positioning the fault monomer, effectively advancing the time point of the thermal runaway, greatly improving the safety, radically solving the problem that the thermal runaway fault has difficult to quantitatively describe the characteristics of the system, and solving the problem that the thermal runaway fault has difficult to be accurately and has good prospect.
Preferably, the invention constructs the mode matrix by setting the analysis window length and the time delay constant and intercepting the time domain signal sequence in window sequence, when constructing the mode matrix, the value of the time delay constant can ensure that the signal does not lose information, the characteristics of data can be completely reflected, and after the singular spectrum entropy value is obtained by calculation, the normalization processing is carried out based on the comparison with the singular spectrum entropy of white noise, thereby being convenient for visually comparing the complexity of the signal and eliminating the influence of the selection of the analysis window length on the calculation result. When the alarm coefficient is quantitatively evaluated by singular spectrum entropy in the step of normalization processing early warning analysis and is compared with the alarm threshold value to judge the thermal runaway risk of the power battery, the range can be determined by analyzing and checking a large amount of driving data of a large data platform of the cloud of the power battery for the selection of the alarm threshold value, so that the application range is enlarged, and the false alarm rate is reduced.
The invention provides a power battery thermal runaway early warning system based on singular spectrum entropy, which corresponds to the power battery thermal runaway early warning method based on singular spectrum entropy, and can be understood as an implementation system of the power battery thermal runaway early warning method based on singular spectrum entropy. The scheme greatly improves the sensitivity of the system to fault signals, can more rapidly detect the occurrence time of risks and position fault monomers, effectively advances the time point of thermal runaway early warning, and greatly improves the safety.
Drawings
FIG. 1 is a flow chart of the power battery thermal runaway early warning method based on singular spectrum entropy.
Fig. 2 is a flowchart of the calculation of singular spectrum entropy optimization in the power battery thermal runaway early warning method based on singular spectrum entropy.
Fig. 3 is a graph showing the voltage distribution of all the cells at the time of failure in a certain group of power cells according to the present invention.
FIG. 4 is a graph of the singular spectral entropy distribution of all the cells at the time of failure in a certain group of power cells according to the present invention.
Fig. 5 is a graph of the characteristic value S of a faulty cell and two normal cells of the present invention over time.
Detailed Description
The present invention will be described below with reference to the accompanying drawings.
A power battery thermal runaway early warning method based on singular spectrum entropy is shown in figure 1, and comprises the steps of constructing a mode matrix, wherein output signals of sensors of a power battery are collected by a vehicle end and are input into the power battery, and data preprocessing (voltage, temperature, current and the like) can be performed, so that the mode matrix is constructed according to a time domain signal sequence formed by a time sequence of signal input; singular value decomposition, namely performing singular value decomposition on the pattern matrix, and arranging singular values to form a singular spectrum of the signal, wherein the number of non-zero singular values is the total number of patterns contained in each column of the pattern matrix; calculating singular spectrum entropy, namely performing singular spectrum analysis of signals, calculating the proportion of each mode in all modes, and further calculating the singular spectrum entropy (namely calculating entropy value) of the time domain signal sequence so as to reflect the uncertainty degree of each mode of the time domain signal sequence under singular spectrum division and quantitatively describe the time domain complex morphological characteristics of signal energy distribution; and a normalization processing early warning analysis step, wherein the obtained singular spectrum entropy of the time domain signal sequence is subjected to normalization processing based on comparison with the singular spectrum entropy of white noise, an alarm coefficient is quantitatively evaluated according to the singular spectrum entropy after normalization processing and is compared with an alarm threshold, a fault monomer is positioned and a dangerous moment is reminded when the singular spectrum entropy is larger than the alarm threshold, and otherwise, a diagnosis result is output, so that the thermal runaway risk of the power battery is judged and the fault position is positioned. The power battery thermal runaway early warning method based on singular spectrum entropy can be understood as adopting the singular spectrum entropy method, in short, the singular spectrum entropy method is an analysis method combining singular spectrum and information entropy, and singular value decomposition results are used as information entropy solving objects. The method is suitable for application occasions with fewer sampling points and noise in the acquired signals. The singular spectrum analysis of signals is a modern spectrum analysis technology based on dynamic analysis, and the basic idea is to acquire the intrinsic complexity characteristics of the signal by carrying out phase space reconstruction and singular value decomposition on a time domain signal sequence of a system. On the basis of singular spectrum analysis of signals, the information entropy (singular spectrum entropy for short) of the singular spectrum is calculated, so that the complex state characteristics of the time sequence can be quantitatively described. When the thermal runaway of the power battery is judged, the abnormal signal is found and analyzed by extracting and processing the signal transmitted by the sensor. The scheme analyzes from the angle of information content of the signals, and early warning is carried out on thermal runaway by comparing the difference between the information content of the fault signals and the information content of the normal signals. The simpler the signal, the smaller the singular spectral entropy: the more the complex transient signal is dispersed in energy, the greater the singular spectrum entropy. The singular spectrum entropy value is very sensitive to abnormal changes of input signals, the current fault monomer can be rapidly detected and positioned, the sensitivity of the system to the fault signal is greatly improved, the moment of risk occurrence can be more rapidly detected, the fault monomer is positioned, compared with the traditional algorithm for judging whether data such as voltage, temperature and the like exceed the threshold value directly by setting the threshold value, the thermal runaway alarm of the battery can be greatly advanced, precious time is striven for fault processing, and the safety is greatly improved.
1. Step of constructing a pattern matrix
Firstly, data are collected and uploaded by a vehicle end, the time interval for uploading the data is 30s, and the data cover various actual working conditions and various states such as charging when the vehicle runs. Taking voltage as an example to calculate normalized singular spectrum entropy values of all power battery monomers, as shown in a flow chart shown in fig. 2, x= { X k K=1, 2,..n is a discrete variable of the original voltage sequence.
The voltage modeling matrix is constructed in accordance with the time sequence of signal input. The pattern matrix is preferably constructed by setting the analysis window length and the delay constant and intercepting the time domain signal sequence in window order, the length of the analysis window being chosen to be M and the delay constant delta. Sample data is truncated in window order of (M, δ), i.e., x= { X k Sequentially divided into lambda segments, and thereby form a pattern matrix a:
wherein,, representing an upward rounding. The window length M is generally not preferably greater than N/3. At the same time be
The delay constant is typically δ=1, with no information missing from the signal.
2. Singular value decomposition step
Singular value decomposition is carried out on the mode matrix A, the number of obtained singular values is set as M, and the singular value sigma is calculated i (i=1, 2,., M) in descending order, then σ 1 ≥σ 2 ≥...≥σ M At this time sigma i A singular value spectrum of the measured signal is constructed. If sigma i The number of non-zero singular values in (a) is j, then the j value represents the total number of patterns contained in each column of pattern matrix a.
3. Step of calculating singular spectrum entropy
And carrying out singular spectrum analysis of the signals, calculating the proportion of each mode in all modes, and further calculating the singular spectrum entropy of the time domain signal sequence. From singular values sigma i The correspondence with j patterns in the pattern matrix A can be known, the singular value spectrum { sigma } i Actually, it is a division of the measured signal in the time domain. If order
Then p is i The specific gravity of the ith mode in all modes. At the same time, p i The i-th pattern of pattern matrix a is also represented as the weight of all patterns. Thus, the singular spectral entropy of the signal in the time domain is obtainable according to the definition of the information entropy (shannon entropy) as:
the singular spectrum entropy essentially reflects the uncertainty of each mode of the time domain signal sequence under the singular spectrum division. The singular spectral entropy also represents the time domain complexity of the signal energy distribution if considered from the perspective of complexity analysis. When the method is applied to judging the working state of the power battery monomer, the complexity of signals is small, the more the energy is concentrated in a few modes, and the singular spectrum entropy value is small under the normal working condition. On the contrary, when the battery monomer is out of control, the complexity of the incoming signal is increased, the energy of the complex transient signal is dispersed, and the singular spectrum entropy value is larger.
4. Normalization processing early warning analysis step
According to the theory of information, when the entropy takes the maximum value, the system is in a steady state, and the energy distribution is most uniform at this time. While the energy distribution difference of each mode of the white noise signal is minimal, the singular spectrum is basically a straight line. Therefore, the singular spectrum entropy of the white noise signal is the largest, and is expressed by the following formula:
singular spectrum entropy H obtained in step of calculating singular spectrum entropy s Singular spectrum entropy based on white noiseThe comparison is normalized, so that the complexity of the signals can be intuitively compared, and the influence of the selection of the length of the analysis window on the calculation result can be eliminated. Normalized singular spectral entropy value +.>The calculated expression of (2) is:
and quantitatively evaluating an alarm coefficient according to the singular spectrum entropy after normalization processing and comparing the alarm coefficient with an alarm threshold value, thereby judging the thermal runaway risk of the power battery and positioning the fault position.
As shown in fig. 3, which shows a graph of voltage distribution of all the cells at the time of failure in a certain group of cells, the voltage variation trend of the cell can be seen, and when the cells have thermal runaway failure, there is a significant voltage drop variation relative to other normal cells. Fig. 4 illustrates the magnitude of singular spectrum entropy values of each single cell at the fault time calculated according to the voltage, and it can be intuitively seen that the single cell at the fault time has a phenomenon of obvious entropy value surge. The invention provides a method for early warning of thermal runaway of a power battery based on the characteristic of singular spectrum entropy values on description of the power battery monomers.
Further, in the step of normalization processing early warning analysis, a Z-score method is adopted to normalize singular spectrum entropy after normalization processing, the difference value of the singular spectrum entropy after normalization processing of a certain power battery monomer and the average value of the singular spectrum entropy of normalization of all power battery monomers is divided by the standard deviation of the singular spectrum entropy of normalization of all power battery monomers to obtain the alarm coefficient of each power battery monomer, and the alarm coefficient is set for quantitative evaluation to realize the identification and early warning of a thermal runaway event, wherein the expression is as follows:
wherein S is a normalized characteristic value, namely an alarm coefficient quantitative evaluation value,for the current entropy value, H ave For all monomers in the current test site +.>Mean, sigma of E Is the standard deviation of the entropy of all the monomers in the current test point.
In all vehicles, the fault vehicle and the normal vehicle obey normal distribution, and the occurrence of the thermal runaway phenomenon of the vehicle is a small probability event, and the fault vehicle and the normal vehicle are restrained by utilizing the 3 sigma principle. The 3σ principle can be described simply as: under the assumption of normal distribution, the probability of occurrence of a value other than the average value of three times σ is small, and thus can be regarded as an abnormal value, that is, |s|=3 in the above expression. Through the test of a large amount of measured data, if 3 sigma is selected, the detection is excessively sensitive, and a higher false alarm rate exists, when the detection is set to be |S| or more than 7.5, the situation that the battery is about to generate thermal runaway phenomenon is judged to have a relatively objective effect, such as a curve of the change of the characteristic values S of a fault single body and two normal single bodies along with time shown in fig. 5, the data is the alarm moment of the traditional method when the data is cut off, the abnormal state is judged by the singular spectrum entropy value before the data, and compared with the traditional method, the method provided by the invention can be used for judging the thermal runaway state of the power battery more quickly. When the threshold value is set, the threshold value needs to be freely adjusted according to the requirements, and the alarm threshold value range can be determined based on the driving data of the power battery cloud big data platform for analysis and verification adjustment, so that the application range is enlarged, and the false alarm rate is reduced.
The invention also relates to a power battery thermal runaway early warning system based on singular spectrum entropy, which corresponds to the power battery thermal runaway early warning method based on singular spectrum entropy, and can be understood as an implementation system of the power battery thermal runaway early warning method based on singular spectrum entropy, and comprises a construction mode matrix module, a singular value decomposition module, a calculation singular spectrum entropy module and a normalization processing early warning analysis module which are sequentially connected, wherein the construction mode matrix module is used for constructing a time domain signal sequence formed by the output signals of all sensors of a vehicle-end acquisition and transmission power battery according to the time sequence of signal transmission; the singular value decomposition module is used for carrying out singular value decomposition on the pattern matrix, arranging singular values to form a singular spectrum of the signal, wherein the number of non-zero singular values is the total number of patterns contained in each column of the pattern matrix; the singular spectrum entropy module is used for carrying out singular spectrum analysis of the signals and calculating the proportion of each mode in all modes, so as to obtain the singular spectrum entropy of the time domain signal sequence by calculation, thereby reflecting the uncertainty degree of each mode of the time domain signal sequence under the singular spectrum division and quantitatively describing the time domain complex morphological characteristics of the signal energy distribution; the normalization processing early warning analysis module performs normalization processing on the singular spectrum entropy of the obtained time domain signal sequence based on comparison with the singular spectrum entropy of white noise, quantitatively evaluates an alarm coefficient according to the singular spectrum entropy after normalization processing and compares the alarm coefficient with an alarm threshold value, thereby judging the thermal runaway risk of the power battery and positioning the fault position. When the thermal runaway of the power battery is judged, the abnormal signal is found and analyzed by extracting and processing the signal transmitted by the sensor. According to the scheme, the information quantity contained in the signals is analyzed, the thermal runaway is warned by comparing the difference between the information quantities of the fault signals and the normal signals, the sensitivity of the system to the fault signals is greatly improved, the moment of risk occurrence can be detected more rapidly, the fault monomers are positioned, the time point of the thermal runaway warning is effectively advanced, and the safety is greatly improved.
Preferably, the construction pattern matrix module constructs the pattern matrix by setting an analysis window length and a delay constant and intercepting the time domain signal sequence in window order.
Preferably, the singular spectrum entropy calculating module calculates the information entropy of the singular spectrum according to the definition of shannon entropy based on the proportion of each mode in all modes, so as to obtain the singular spectrum entropy of the time domain signal sequence.
Preferably, the normalization processing early warning analysis module adopts a Z-score method to normalize singular spectrum entropy after normalization processing, divides the difference value of the singular spectrum entropy after normalization processing of a certain power battery monomer and the average value of the singular spectrum entropy of normalization of all power battery monomers by the standard deviation of the singular spectrum entropy of normalization of all power battery monomers to obtain the alarm coefficient of each power battery monomer, and compares the alarm coefficient of each power battery monomer with an alarm threshold value, thereby judging the thermal runaway risk of the power battery and positioning the fault position.
It should be noted that the above-described embodiments will enable those skilled in the art to more fully understand the invention, but do not limit it in any way. Therefore, although the present invention has been described in detail with reference to the drawings and examples, it will be understood by those skilled in the art that the present invention may be modified or equivalent, and in all cases, all technical solutions and modifications which do not depart from the spirit and scope of the present invention are intended to be included in the scope of the present invention.
Claims (10)
1. A power battery thermal runaway early warning method based on singular spectrum entropy is characterized by comprising the following steps:
a step of constructing a mode matrix, in which the output signals of the sensors of the power battery are collected and fed by a vehicle end, and the mode matrix is constructed according to a time domain signal sequence formed by the time sequence of signal feeding;
singular value decomposition, namely performing singular value decomposition on the pattern matrix, and arranging singular values to form a singular spectrum of the signal, wherein the number of non-zero singular values is the total number of patterns contained in each column of the pattern matrix;
calculating singular spectrum entropy, namely performing singular spectrum analysis of the signal, calculating the proportion of each mode in all modes, and further calculating the singular spectrum entropy of the time domain signal sequence to reflect the uncertainty degree of each mode of the time domain signal sequence under singular spectrum division and quantitatively describe the time domain complex morphological characteristics of signal energy distribution;
and a normalization processing early warning analysis step, wherein the obtained singular spectrum entropy of the time domain signal sequence is subjected to normalization processing based on comparison with the singular spectrum entropy of white noise, and an alarm coefficient is quantitatively evaluated according to the singular spectrum entropy after normalization processing and is compared with an alarm threshold value, so that the thermal runaway risk of the power battery is judged and the fault position is positioned.
2. The thermal runaway warning method of a power battery according to claim 1, wherein in the step of constructing a pattern matrix, the pattern matrix is constructed by setting an analysis window length and a time delay constant and intercepting a time domain signal sequence in a window order.
3. The thermal runaway warning method of a power battery according to claim 1 or 2, wherein in the step of calculating singular spectrum entropy, the information entropy of the singular spectrum is calculated according to the definition of shannon entropy based on the specific gravity of each mode in all modes, so as to obtain the singular spectrum entropy of the time domain signal sequence.
4. The power battery thermal runaway warning method according to claim 3, wherein in the normalization processing early warning analysis step, a Z-score method is adopted to normalize singular spectrum entropy after normalization processing, a difference value between the singular spectrum entropy after normalization processing of a certain power battery monomer and an average value of the singular spectrum entropy after normalization of all power battery monomers is divided by a standard deviation of the singular spectrum entropy after normalization of all power battery monomers to obtain an alarm coefficient of each power battery monomer, and the alarm coefficient of each power battery monomer is compared with an alarm threshold value, so that thermal runaway risks of the power battery are judged and fault positions are located.
5. The method of claim 4, wherein in the step of constructing the pattern matrix, the output signals of the sensors of the power battery include voltage, current, and temperature signals under several actual conditions or in a charged state when the vehicle is running.
6. The power battery thermal runaway early warning method according to claim 1, wherein in the normalization processing early warning analysis step, analysis and verification adjustment are performed based on driving data of a power battery cloud big data platform so as to determine an alarm threshold range.
7. A power battery thermal runaway early warning system based on singular spectrum entropy is characterized by comprising a construction mode matrix module, a singular value decomposition module, a singular spectrum entropy calculation module and a normalization processing early warning analysis module which are connected in sequence,
the mode matrix constructing module is used for acquiring and uploading output signals of each sensor of the driving force battery by a vehicle end and constructing a mode matrix according to a time domain signal sequence formed by a time sequence of signal transmission;
the singular value decomposition module is used for carrying out singular value decomposition on the pattern matrix, arranging singular values to form a singular spectrum of the signal, wherein the number of non-zero singular values is the total number of patterns contained in each column of the pattern matrix;
the singular spectrum entropy calculating module is used for carrying out singular spectrum analysis on signals and calculating the proportion of each mode in all modes, so as to calculate singular spectrum entropy of the time domain signal sequence, and reflect the uncertainty degree of each mode of the time domain signal sequence under singular spectrum division and quantitatively describe the time domain complex morphological characteristics of signal energy distribution;
the normalization processing early warning analysis module performs normalization processing on the obtained singular spectrum entropy of the time domain signal sequence based on comparison with the singular spectrum entropy of white noise, quantitatively evaluates an alarm coefficient according to the singular spectrum entropy after normalization processing and compares the alarm coefficient with an alarm threshold value, thereby judging the thermal runaway risk of the power battery and positioning the fault position.
8. The power cell thermal runaway warning system of claim 7, wherein the build mode matrix module builds the mode matrix by setting an analysis window length and a time delay constant and intercepting a time domain signal sequence in window order.
9. The power battery thermal runaway warning system according to claim 7 or 8, wherein the singular spectrum entropy calculation module calculates the information entropy of the singular spectrum according to the definition of shannon entropy based on the specific gravity of each mode in all modes, so as to obtain the singular spectrum entropy of the time domain signal sequence.
10. The power battery thermal runaway warning system according to claim 9, wherein the normalization processing early warning analysis module adopts a Z-score method to normalize singular spectrum entropy after normalization processing, divides a difference value between the singular spectrum entropy after normalization processing of a certain power battery monomer and an average value of the singular spectrum entropy of normalization of all power battery monomers by a standard deviation of the singular spectrum entropy of normalization of all power battery monomers to obtain an alarm coefficient of each power battery monomer, and compares the alarm coefficient of each power battery monomer with an alarm threshold value, thereby judging thermal runaway risk of the power battery and positioning a fault position.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310308492.XA CN116451038A (en) | 2023-03-27 | 2023-03-27 | Power battery thermal runaway early warning method and system based on singular spectrum entropy |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310308492.XA CN116451038A (en) | 2023-03-27 | 2023-03-27 | Power battery thermal runaway early warning method and system based on singular spectrum entropy |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116451038A true CN116451038A (en) | 2023-07-18 |
Family
ID=87126493
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310308492.XA Pending CN116451038A (en) | 2023-03-27 | 2023-03-27 | Power battery thermal runaway early warning method and system based on singular spectrum entropy |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116451038A (en) |
Cited By (2)
| 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 |
| CN117648589A (en) * | 2024-01-30 | 2024-03-05 | 云储新能源科技有限公司 | Energy storage battery thermal runaway early warning method, system, electronic equipment and medium |
-
2023
- 2023-03-27 CN CN202310308492.XA patent/CN116451038A/en active Pending
Cited By (3)
| 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 |
| CN117648589A (en) * | 2024-01-30 | 2024-03-05 | 云储新能源科技有限公司 | Energy storage battery thermal runaway early warning method, system, electronic equipment and medium |
| CN117648589B (en) * | 2024-01-30 | 2024-05-14 | 云储新能源科技有限公司 | Energy storage battery thermal runaway early warning method, system, electronic equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116451038A (en) | Power battery thermal runaway early warning method and system based on singular spectrum entropy | |
| CN110018425B (en) | Power battery fault diagnosis method and system | |
| CN113791350B (en) | Battery fault prediction method | |
| CN111641236A (en) | Dynamic threshold power battery charging voltage state judgment method based on big data AI | |
| CN113131019A (en) | New energy electric automobile battery module operation safety real-time supervision early warning system based on cloud calculates and thing networking | |
| CN117556364A (en) | Mining ore pressure safety intelligent monitoring system | |
| CN117517596A (en) | Method and system for monitoring combustible and toxic harmful gases in real time based on Internet of things | |
| CN116533903B (en) | Low-voltage early warning system and method based on automobile sensor | |
| CN116923188A (en) | Charging protection system of new energy vehicle | |
| CN113581015B (en) | Safety early warning method and device for fuel cell system | |
| CN109782184B (en) | Thermal runaway risk diagnosis method of Pack system and electronic equipment thereof | |
| CN116190818A (en) | Lithium battery device based on multidimensional sensing fusion, lithium battery monitoring device and method | |
| AU2021104319A4 (en) | A system for traction inverter fault detection and a method thereof | |
| CN115601913A (en) | Combustible dangerous material fire risk early warning method and device | |
| CN114683963B (en) | Battery safety early warning method, device and equipment | |
| US11842578B2 (en) | Vehicle safety component diagnosis apparatus | |
| CN118046804A (en) | Hybrid power energy management control system of fuel cell | |
| CN118171217B (en) | Hydraulic engineering water level early warning method and system based on BIM technology | |
| CN118245758B (en) | Steady-state analysis processing system and method for port lifting equipment | |
| CN118395222B (en) | Abnormality monitoring method and system for transformer | |
| CN117930027B (en) | Power battery capacity abnormality detection method, device and platform | |
| CN117207844A (en) | Dynamic alarm method, device, equipment and storage medium | |
| CN111680356B (en) | Unmanned aerial vehicle actuator real-time safety performance evaluation method based on two-stage observer | |
| CN114559774A (en) | Unmanned automobile tire monitoring method based on artificial intelligence | |
| CN117767506A (en) | New energy automobile power adapter power self-adaption method based on data driving |
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 |