CN114742297B - Method for processing power battery - Google Patents
Method for processing power battery Download PDFInfo
- Publication number
- CN114742297B CN114742297B CN202210376149.4A CN202210376149A CN114742297B CN 114742297 B CN114742297 B CN 114742297B CN 202210376149 A CN202210376149 A CN 202210376149A CN 114742297 B CN114742297 B CN 114742297B
- Authority
- CN
- China
- Prior art keywords
- target
- initial
- power battery
- data
- abscissa
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 78
- 238000012545 processing Methods 0.000 title claims description 37
- 239000011159 matrix material Substances 0.000 claims abstract description 97
- 238000000265 homogenisation Methods 0.000 claims description 42
- 238000012549 training Methods 0.000 claims description 34
- 238000005070 sampling Methods 0.000 claims description 24
- 238000004140 cleaning Methods 0.000 claims description 15
- 230000002159 abnormal effect Effects 0.000 claims description 14
- 230000008859 change Effects 0.000 claims description 14
- 239000012634 fragment Substances 0.000 claims description 11
- 238000003672 processing method Methods 0.000 abstract description 10
- 238000005516 engineering process Methods 0.000 abstract description 6
- 230000008569 process Effects 0.000 description 36
- 238000013528 artificial neural network Methods 0.000 description 15
- 238000010606 normalization Methods 0.000 description 15
- 238000013527 convolutional neural network Methods 0.000 description 11
- 238000011176 pooling Methods 0.000 description 10
- 238000010586 diagram Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 5
- WHXSMMKQMYFTQS-UHFFFAOYSA-N Lithium Chemical compound [Li] WHXSMMKQMYFTQS-UHFFFAOYSA-N 0.000 description 4
- 239000010410 layer Substances 0.000 description 4
- 229910052744 lithium Inorganic materials 0.000 description 4
- 239000000178 monomer Substances 0.000 description 4
- 230000035772 mutation Effects 0.000 description 4
- HBBGRARXTFLTSG-UHFFFAOYSA-N Lithium ion Chemical compound [Li+] HBBGRARXTFLTSG-UHFFFAOYSA-N 0.000 description 3
- 230000005856 abnormality Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000001914 filtration Methods 0.000 description 3
- 229910001416 lithium ion Inorganic materials 0.000 description 3
- 238000003062 neural network model Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000013145 classification model Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 238000002372 labelling Methods 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 239000012792 core layer Substances 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000003550 marker Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Business, Economics & Management (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computational Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Biophysics (AREA)
- Pure & Applied Mathematics (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Health & Medical Sciences (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Quality & Reliability (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- Algebra (AREA)
- Operations Research (AREA)
- Databases & Information Systems (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Tourism & Hospitality (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
Abstract
The invention discloses a processing method of a power battery. Wherein the method comprises the following steps: acquiring target characteristic data of a target power battery, wherein the target characteristic data is data corresponding to target characteristics of the target power battery; matrixing the target feature data to generate a target matrix; and identifying the target matrix by using the fault identification model to obtain an identification result of the target power battery, wherein the identification result is used for indicating whether the target power battery has faults or not. The invention solves the technical problem of lower accuracy of fault identification of the power battery in the related technology.
Description
Technical Field
The invention relates to the field of new energy automobiles, in particular to a processing method of a power battery.
Background
The existing fault identification technology of the lithium ion power battery is to search battery characteristics to conduct numerical conversion, and further identify whether the power battery has faults or not through a threshold value, but the threshold value judgment cannot solve all problems at one time, when the conditions of different internal proportions of the battery and different battery manufacturers occur, the threshold value needs to be repeatedly updated, and the problem of inaccurate identification exists in the threshold value identification.
In view of the above problems, no effective solution has been proposed at present.
Disclosure of Invention
The embodiment of the invention provides a processing method of a power battery, which at least solves the technical problem of low accuracy of fault identification of the power battery in the related technology.
According to an aspect of an embodiment of the present invention, there is provided a method for processing a power battery, including: acquiring target characteristic data of a target power battery, wherein the target characteristic data is data corresponding to target characteristics of the target power battery; matrixing the target feature data to generate a target matrix; and identifying the target matrix by using the fault identification model to obtain an identification result of the target power battery, wherein the identification result is used for indicating whether the target power battery has faults or not.
Optionally, matrixing the target feature data to generate a target matrix, including: dividing target feature data based on a preset battery state to generate target state data, wherein the target state data is data corresponding to the preset battery state in the target feature data; and carrying out matrixing treatment on the target state data to generate a target matrix.
Optionally, matrixing the target state data to generate a target matrix, including: acquiring a target graph corresponding to the target state data; performing grid division on the target curve graph by using preset granularity to generate an initial grid curve graph; carrying out homogenization treatment on the initial grid graph to generate a target grid graph; and matrixing the target grid graph to generate a target matrix.
Optionally, normalizing the grid curve image to generate a target grid graph, including: obtaining a target segment which accords with a preset rule in a grid curve image; and carrying out homogenization treatment on the target coordinates of the target segment to generate a target grid graph.
Optionally, performing a homogenization process on the target coordinates of the target segment to generate a target grid graph, including: carrying out homogenization treatment on the initial abscissa of the target segment to obtain a target abscissa, wherein the initial abscissa is used for representing the initial duration of the target segment; carrying out homogenization treatment on an initial ordinate of the target segment to obtain a target ordinate, wherein the initial ordinate is used for representing an initial feature range corresponding to the target feature; a target grid graph is generated based on the target abscissa and the target ordinate.
Optionally, performing a homogenization process on the initial abscissa of the target segment based on a preset duration to obtain a target abscissa, including: acquiring the duration of a target fragment; comparing the preset time length with the duration time length, and determining a comparison result, wherein the comparison result is used for indicating whether the preset time length is longer than the duration time length; sampling an initial abscissa based on a preset proportion under the condition that the preset duration is smaller than the duration, and obtaining a sampling result, wherein the preset proportion is the proportion of the preset duration to the initial duration of the target segment; and carrying out homogenization treatment on the initial abscissa based on the sampling result to obtain the target abscissa.
Optionally, the method further comprises: under the condition that the preset time length is longer than or equal to the duration time length, interpolating the initial abscissa based on the preset proportion to obtain an interpolation result; and carrying out homogenization treatment on the initial abscissa based on the interpolation result to obtain the target abscissa.
Optionally, performing a homogenization process on the initial ordinate of the target segment based on the target feature to obtain a target ordinate, including: determining a target cleaning rule corresponding to the target feature; and removing the target coordinates in the initial ordinate of the target segment based on the target cleaning rule to obtain the target ordinate, wherein the target coordinates are abnormal coordinates in the initial ordinate.
Optionally, the method further comprises: obtaining a target training set, wherein the target training set comprises: the system comprises a sample matrix, at least one piece of label information corresponding to the sample matrix, at least one piece of label information and at least one piece of label information, wherein the at least one piece of label information is used for indicating whether at least one area in the sample matrix has faults or not, and the at least one piece of label information corresponds to the at least one area; constructing an initial model based on a plurality of preset convolution kernels; training the initial model based on the target training set to obtain a fault identification model.
According to another aspect of the embodiment of the present invention, there is also provided a computer readable storage medium, where the computer readable storage medium includes a stored program, and when the program runs, the device in which the computer readable storage medium is controlled to execute the above-mentioned method for processing a power battery.
According to another aspect of the embodiment of the present invention, there is also provided a computer terminal including: the power battery processing device comprises a memory and a processor, wherein the processor is used for running a program stored in the memory, and the processing method of the power battery is executed when the program runs.
In the embodiment of the invention, the target characteristic data of the target power battery is acquired, wherein the target characteristic data is data corresponding to the target characteristic of the target power battery; matrixing the target feature data to generate a target matrix; the method comprises the steps of utilizing a fault recognition model to recognize a target matrix to obtain a recognition result of a target power battery, wherein the recognition result is used for representing whether the target power battery has a fault or not, and by carrying out matrixing processing on characteristic data of the power battery and utilizing the fault recognition model to recognize the target matrix, the purpose of accurately early warning cloud safety of the power battery is achieved, so that the technical effect of improving the early warning accuracy of the cloud safety of the power battery is achieved, and the technical problem that the accuracy rate of fault recognition of the power battery in related technologies is low is solved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the application and do not constitute a limitation on the application. In the drawings:
Fig. 1 is a flowchart of a processing method of a power battery according to an embodiment of the present invention;
FIG. 2 is a matrix diagram of an alternative embodiment of a convolution kernel curve level change in accordance with an embodiment of the present invention;
FIG. 3 is a matrix diagram that optionally embodies the vertical variation of the convolution kernel curve in accordance with an embodiment of the present invention;
FIG. 4 is a matrix diagram of an alternative embodiment of a downward extending variation of a convolution kernel curve in accordance with an embodiment of the present invention;
FIG. 5 is a matrix diagram of an alternative embodiment of a convolution kernel curve upward extension variation in accordance with an embodiment of the present invention;
FIG. 6 is a matrix diagram that optionally embodies the upward abrupt change of the convolution kernel curve in accordance with an embodiment of the present invention;
FIG. 7 is a matrix diagram of an alternative embodiment of a downward abrupt change in a convolution kernel curve in accordance with an embodiment of the present invention;
FIG. 8 is a flow chart of an alternative power cell voltage temperature signature graphical fault identification method in accordance with an embodiment of the present invention;
Fig. 9 is a schematic structural view of a processing device for a power battery according to an embodiment of the present invention.
Detailed Description
In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and the claims of the present invention and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the invention described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
Example 1
According to an embodiment of the present invention, there is provided a processing method embodiment of a power battery, it being noted that the steps shown in the flowchart of the drawings may be performed in a computer system such as a set of computer executable instructions, and that although a logical order is shown in the flowchart, in some cases the steps shown or described may be performed in an order different from that herein.
Fig. 1 is a flowchart of a processing method of a power battery according to an embodiment of the present invention, as shown in fig. 1, the method including the steps of:
Step S102, target characteristic data of the target power battery is obtained, wherein the target characteristic data is data corresponding to target characteristics of the target power battery.
The target power battery may be a lithium power battery, or may be another type of power battery, and the present invention is not limited to this, and the present embodiment is described by taking a lithium power battery as an example.
The target characteristics described above may include, but are not limited to, characteristics such as voltage, current, temperature, etc. of the target power cell.
The above-mentioned target feature data is data corresponding to a target feature of the target power battery set by the user in advance, and the target feature data may be stored in the processor, so as to accurately identify a fault of the power battery.
In an alternative embodiment, in order to accurately identify the fault of the target power battery, first, target characteristic data of the power battery set by a user in advance needs to be acquired, that is, data corresponding to voltage, current and temperature of a single day of the target power battery needs to be acquired in advance.
Step S104, matrixing the target feature data to generate a target matrix.
The above-mentioned matrixing process may be a normalization process of a matrix, a symmetry process of a matrix, or other types of processes of a matrix, where the normalization process may be a process of converting a value range of target feature data into a range of 0 to 1, where the normalization process may eliminate an influence of a larger value in the target feature data on a smaller value, for example, a threshold may be set for voltage data in the target feature data, and if a certain voltage data far exceeds the threshold, in order to eliminate an influence of the voltage data on other voltage data in the threshold range, the normalization process may be performed on the voltage data, so that the value of the voltage data is in the range of 0 to 1, that is, an influence of the larger value in the target feature data on the smaller value is eliminated. Similarly, in order to accurately identify a failure of the power battery, in the present embodiment, the matrixing process is described by taking a normalization process of the matrix as an example.
The target matrix may be a matrix obtained by performing matrix normalization processing on target feature data, where values of the target feature data in the target matrix are all in a range from 0to 1.
In an alternative embodiment, in order to eliminate the influence of the larger value on the smaller value in the target feature data, the target feature data may be subjected to matrix normalization processing, so as to obtain a target matrix with the value of the target data in the interval of 0 to 1.
In another alternative embodiment, in order to implement the matrixing process of the target feature data, the target feature data may be further subjected to state division to obtain target state data, where the target state may include, but is not limited to, a driving state, a charging state and a stopping state of the power battery, and the target state data may include, but is not limited to, data of a voltage, a current and a temperature of the power battery in the driving state, data of a voltage, a current and a temperature of the power battery in the charging state, and data of a voltage, a current and a temperature of the power battery in the stopping state. After the target state data is divided, the target state data can be subjected to matrix normalization processing, and then a target matrix can be obtained.
And S106, identifying the target matrix by using a fault identification model to obtain an identification result of the target power battery, wherein the identification result is used for indicating whether the target power battery has a fault or not.
The fault recognition model may be a cyclic neural network model, a convolutional neural network (Convolutional Neural Network, abbreviated as CNN) model, or the like, and the user may select the model according to his own needs, which is not specifically limited in the present invention, and in order to obtain a recognition result capable of accurately recognizing whether the power battery is faulty, the CNN neural network is selected as the fault recognition model in this embodiment.
The above-described recognition result is a result that can be presented in the form of text, voice, or image. It is mainly used for explaining whether the target power battery fails.
In an alternative embodiment, the target matrix may be input into the fault recognition model, and the fault recognition model is used to recognize the target matrix, so as to determine whether the target feature corresponding to the target matrix is abnormal, and further determine whether the target power battery has a fault.
In another alternative embodiment, the recognition result may be outputted in the form of text to describe whether the power battery is malfunctioning, for example: "an abnormality has occurred in a certain segment or a certain data of the matrix"; the recognition result can be output in a voice form to describe whether the power battery fails, and voice can be output to ensure that a user needs to process the matrix as soon as possible; the recognition result can be output in the form of an image to describe whether the power battery fails, and the specific operation is that the target matrix can be converted into the image, and the abnormal part is marked on the image, so that a user can process the input target matrix based on the marked abnormal part. Of course, the fault recognition model can output not only these results, but also other types of results, and the embodiment is merely illustrative and not particularly limited.
Because the target matrix corresponds to the graph corresponding to the target feature of the power battery, whether the target feature is abnormal or not can be monitored by identifying the target matrix through the fault identification model, and therefore the identification efficiency of the fault of the target power battery is improved.
In an alternative embodiment, in order to obtain a recognition result capable of accurately recognizing whether the power battery is faulty, the target matrix may be recognized by using the fault recognition model, so as to obtain a recognition result of the power battery, for example, a text, a voice and an image result, and then the user may know whether the target power battery is faulty through the recognition result.
Through the embodiment, the target characteristic data of the target power battery is obtained, wherein the target characteristic data is data corresponding to the target characteristic of the target power battery; matrixing the target feature data to generate a target matrix; the method comprises the steps of utilizing a fault recognition model to recognize a target matrix to obtain a recognition result of a target power battery, wherein the recognition result is used for representing whether the target power battery has faults or not, and by carrying out matrixing processing on characteristic data of the power battery and utilizing the fault recognition model to recognize the target matrix, the technical effect of improving the fault recognition accuracy of the power battery can be achieved, and the technical problem that the accuracy rate of carrying out fault recognition on the power battery in the related technology is lower is solved.
Optionally, matrixing the target feature data to generate a target matrix, including: dividing target feature data based on a preset battery state to generate target state data, wherein the target state data is data corresponding to the preset battery state in the target feature data; and carrying out matrixing treatment on the target state data to generate a target matrix.
The preset battery state is a battery state set by a user, and is not particularly limited, and in order to accurately identify the faults of the power battery through the characteristics of voltage, current, temperature and the like of the power battery, the preset battery state in the embodiment of the invention can include, but is not limited to, a driving state, a charging state and a parking state of the power battery on a single day.
The above-mentioned division may be a processing method set by the user, and is not particularly limited, and in this embodiment, in order to unify all the voltage-time two-dimensional grid patterns or the temperature-time two-dimensional grid patterns by interpolation or sampling, the target feature data may be divided by setting equal length, equal width, and equal density.
The target state data may include, but is not limited to, voltage, current, temperature data of the target power battery during traveling, voltage, current, temperature data of the target power battery during charging, voltage, current, temperature data of the target power battery during stopping.
The specific partitioning is referred to in the following table 1:
TABLE 1
In an alternative embodiment, in order to perform matrixing processing on the target feature data to generate a target matrix so as to accurately identify a fault of the power battery, the target feature data may be firstly subjected to state division based on a preset battery state to generate target state data, and after the target state data is obtained, the target state data is subjected to matrix normalization processing, so that the target matrix may be obtained.
Optionally, matrixing the target state data to generate a target matrix, including: acquiring a target graph corresponding to the target state data; performing grid division on the target curve graph by using preset granularity to generate an initial grid curve graph; carrying out homogenization treatment on the initial grid graph to generate a target grid graph; and matrixing the target grid graph to generate a target matrix.
The target graph may be a graph with time as an abscissa and target state data as an ordinate, where the target graph may intuitively reflect trends including, but not limited to, voltage data of the target power battery during driving, voltage data of the target power battery during charging, voltage data of the target power battery during parking, temperature data of the target power battery during driving, temperature data of the target power battery during charging, and temperature data of the target power battery during parking, which change with time.
The preset granularity may be a minimum granularity set by a user in advance, where the preset granularity may include, but is not limited to, granularity that divides voltage, temperature, and time into minimum units, and in this embodiment, voltage may be millivolts (mV), temperature is degrees celsius (deg.c), and time is seconds(s) as the preset granularity.
The initial grid graph may be a voltage-time two-dimensional grid graph and a temperature-time two-dimensional grid graph, which are generated by dividing the target state data according to a preset granularity and then sorting according to a time sequence.
The target grid graph may be a grid graph obtained by normalizing the initial grid graph.
In an alternative embodiment, the target graph corresponding to the target state data may be obtained first, that is, a trend graph that may intuitively reflect the change over time of the voltage data of the target power battery during driving, the voltage data of the target power battery during charging, the voltage data of the target power battery during stopping, the temperature data of the target power battery during driving, the temperature data of the target power battery during charging, the temperature data of the target power battery during stopping, and the like, and then the target graph may be divided according to preset granularity set in advance by the user and ordered according to time, so as to obtain an initial grid graph, and the target graph may be divided according to preset granularity set in advance, so that the normalization process performed by the user may be faster and simpler, then the target grid graph may be obtained by performing the normalization process on the initial grid graph, the data of the initial grid graph may be all defined in a range from 0 to 1, the calculation amount of the data may be reduced for the matrix process performed for the subsequent user, and finally the normalization process may be performed on the target graph, so as to obtain the target matrix.
Optionally, normalizing the grid curve image to generate a target grid graph, including: obtaining a target segment which accords with a preset rule in a grid curve image; and carrying out homogenization treatment on the target coordinates of the target segment to generate a target grid graph.
The preset rule may be a rule set by a user in advance for screening the segments, and is mainly used for reducing the problem of data accuracy reduction caused by overlarge difference between the segments, which is not particularly limited. In this embodiment, the segments are classified by taking segments with similar durations as preset rules, and in this embodiment, segments with a duration of about 1 hour may be taken as examples of segments with similar durations, where the segments may be data of the voltage and the temperature of the target power battery in a period of time in driving, charging and parking states.
The target segment may be a segment obtained by screening the segment according to a preset rule, and is not specifically limited, and in this embodiment, the ratio of the minimum duration to the maximum duration is selected to be 4:5, the target segment can be used for normalizing the grid curve image to generate a target grid curve graph, for example, the abscissa and the ordinate in the target segment can be respectively normalized to obtain the target grid curve graph.
The above-mentioned target coordinates may be an ordinate and an abscissa in the target segment, wherein the ordinate may be voltage data and temperature data, and the abscissa may be time data.
In an alternative embodiment, in order to perform normalization processing on the grid curve image to generate a target grid curve, first, a target segment in the grid curve image, which meets a preset rule, needs to be acquired, that is, first, a ratio of a minimum duration to a maximum duration in the grid curve image needs to be acquired, where the ratio is 4: fragments within 5, and then carrying out homogenization treatment on the abscissa and the ordinate of the target fragments, so as to generate a target grid graph.
Optionally, performing a homogenization process on the target coordinates of the target segment to generate a target grid graph, including: carrying out homogenization treatment on the initial abscissa of the target segment to obtain a target abscissa, wherein the initial abscissa is used for representing the initial duration of the target segment; carrying out homogenization treatment on an initial ordinate of the target segment to obtain a target ordinate, wherein the initial ordinate is used for representing an initial feature range corresponding to the target feature; a target grid graph is generated based on the target abscissa and the target ordinate.
The initial abscissa may be an abscissa divided by segments, which may represent a duration of time of the target segment in a period of time, and the target abscissa may be an abscissa after the initial abscissa is subjected to a homogenization process.
The ordinate of the target may be an ordinate of the target after the division of the segments, which may represent an initial feature range of the target feature corresponding to the target segment in the initial duration, and the ordinate of the target may be an ordinate of the target after the initial ordinate is subjected to the homogenization treatment.
The initial duration may be a duration of the target segment within a period of time after the segment division, and the initial feature range may be a variation range of feature data of the target feature within the initial duration, which is not specifically limited, and in this embodiment, a voltage in a range of 2.5-4.2 volts (V) and a temperature in a range of-20-60 degrees celsius (deg.c) may be exemplified.
In an alternative embodiment, in order to perform a homogenization process on the target coordinates of the target segment to generate a target grid graph, first, the initial abscissa of the target segment may be subjected to a homogenization process to obtain the target abscissa, second, the initial ordinate of the target segment may be subjected to a homogenization process to obtain the target ordinate, and finally, the target grid graph may be generated according to the target abscissa and the target ordinate.
Optionally, performing a homogenization process on the initial abscissa of the target segment based on a preset duration to obtain a target abscissa, including: acquiring the duration of a target fragment; comparing the preset time length with the duration time length, and determining a comparison result, wherein the comparison result is used for indicating whether the preset time length is longer than the duration time length; sampling an initial abscissa based on a preset proportion under the condition that the preset duration is smaller than the duration, and obtaining a sampling result, wherein the preset proportion is the proportion of the preset duration to the initial duration of the target segment; and carrying out homogenization treatment on the initial abscissa based on the sampling result to obtain the target abscissa.
The preset time period is a time period set in advance by the user, and is not particularly limited, and in this embodiment, since the quick charge in the charging section is all about 1 hour, in this embodiment, 1 hour is taken as the preset time period.
The above-mentioned preset ratio is a ratio value set in advance by the user, and is not particularly limited, and in this embodiment, the ratio of "preset duration/initial duration of the target segment" is taken as the preset ratio.
The sampling method described above is not particularly limited, and in this embodiment, an equal-proportion sampling method is selected.
The sampling result may be a result of performing equal proportion sampling on the initial abscissa based on a preset proportion, and the equal proportion sampling on the initial abscissa may obtain abscissas with the same density.
In an alternative embodiment, in order to perform a homogenization process on an initial abscissa of a target segment based on a preset duration to obtain a target abscissa, first, time length data (i.e., duration) of the target segment within a period of time may be obtained, the duration is compared with the preset duration, if it is confirmed that the preset duration is smaller than the duration, an equal proportion sampling is performed on the initial abscissa based on a preset proportion set in advance, the density of the obtained abscissa is the same as that of the preset duration, and finally, a homogenization process is performed on the initial abscissa based on a sampling result to obtain the target abscissa.
Optionally, the method further comprises: under the condition that the preset time length is longer than or equal to the duration time length, interpolating the initial abscissa based on the preset proportion to obtain an interpolation result; and carrying out homogenization treatment on the initial abscissa based on the interpolation result to obtain the target abscissa.
The interpolation method is not particularly limited, and in this embodiment, the interpolation method may include, but is not limited to, average difference, equal ratio interpolation, and the like.
The interpolation result may be a result of performing an equal-ratio interpolation on the initial abscissa based on a preset ratio, and the equal-ratio interpolation on the initial abscissa may be performed to obtain abscissas with the same density.
In an alternative embodiment, in order to perform a homogenization process on an initial abscissa of a target segment based on a preset duration to obtain a target abscissa, first, time length data (i.e., duration) of the target segment within a period of time may be obtained, the duration is compared with the preset duration, if the preset duration is confirmed to be greater than or equal to the duration, the initial abscissa may be interpolated based on a preset proportion set in advance, and then, the obtained abscissa density is the same as the density of the preset duration, and similarly, the initial abscissa is subjected to a homogenization process based on the interpolation result to obtain the target abscissa.
Optionally, performing a homogenization process on the initial ordinate of the target segment based on the target feature to obtain a target ordinate, including: determining a target cleaning rule corresponding to the target feature; and removing the target coordinates in the initial ordinate of the target segment based on the target cleaning rule to obtain the target ordinate, wherein the target coordinates are abnormal coordinates in the initial ordinate.
The target cleaning rule may be a data cleaning rule set in advance by a user, which is not specifically limited. In this embodiment, the voltage data of greater than 4.2V and less than 2.5V in the initial ordinate may be removed based on a common data cleaning rule, and the temperature data of greater than 60 ℃ and less than-20 ℃ in the initial ordinate may be removed, where the data cleaning rule may be a rule for deleting incomplete data, abnormal data and repeated data in the data.
The above-mentioned target coordinates are coordinates in which an abnormality occurs in the initial ordinate, and in this embodiment, may be voltage data of more than 4.2V and less than 2.5V, temperature data of more than 60 ℃ and less than-20 ℃.
In an alternative embodiment, in order to perform a process of homogenizing the initial ordinate of the target segment based on the target feature to obtain the target ordinate, the target cleaning rule may be set in advance, and then voltage data greater than 4.2V and less than 2.5V in the initial ordinate of the target segment based on the target cleaning rule may be removed, and temperature data (i.e. the target coordinate) greater than 60 ℃ and less than-20 ℃ may be obtained.
Optionally, the method further comprises: obtaining a target training set, wherein the target training set comprises: the system comprises a sample matrix, at least one piece of label information corresponding to the sample matrix, at least one piece of label information and at least one piece of label information, wherein the at least one piece of label information is used for indicating whether at least one area in the sample matrix has faults or not, and the at least one piece of label information corresponds to the at least one area; constructing an initial model based on a plurality of preset convolution kernels; training the initial model based on the target training set to obtain a fault identification model.
The sample matrix may be a matrix corresponding to the sample voltage or the sample temperature, and it has been determined in advance whether different areas in the matrix corresponding to the sample voltage or the sample temperature have faults.
The tag information may be information indicating whether or not a region corresponding to the tag information in the sample matrix has a failure.
The above-mentioned multiple preset convolution kernels may be convolution kernels set by a user in advance, and an initial identification model may be constructed based on the preset convolution kernels to achieve an effect of accurately identifying a fault of the power battery, which is not particularly limited, and may be a 3*3 convolution kernel in this embodiment, where the 6 convolution kernels are shown in fig. 2 to 7, fig. 2 represents a curve level, fig. 3 represents a curve vertical, fig. 4 represents a curve downward extension, fig. 5 represents a curve upward extension, fig. 6 represents a curve upward mutation, fig. 7 represents a curve downward mutation, and a basic curve change may be described.
The initial model is a training model which can judge whether the region of the sample matrix has faults or not and output the result.
In an optional embodiment, in order to realize accurate early warning for cloud security of the target power battery, firstly, a target training set may be obtained, where the target training set includes a sample matrix, at least one piece of tag information corresponding to the sample matrix corresponds to at least one area, the at least one piece of tag information may indicate whether a fault occurs in a corresponding area in the sample matrix, and secondly, an initial model may be constructed based on a plurality of convolution kernels set in advance, the target training set is input to the initial model for multiple training and output results, and then a fault recognition model may be obtained.
An embodiment of the present invention is described in detail below with reference to fig. 8:
FIG. 8 is a flowchart of a power battery voltage and temperature characteristic graphical fault recognition method according to an embodiment of the invention, which is characterized by three sub-steps, namely segment division, characteristic setting, learning and classification, wherein the main purpose of the segment division is to divide data of a single day into segments according to the use state of a battery and realize equal-length, equal-width and equal-density graphical gridding division; the characteristic setting mainly aims at setting a convolution kernel, namely a most basic filtering label, of a CNN convolution neural network and manually marking each fragment image in a training set to form a training set; the method has the main purposes of learning and classifying, namely CNN neural network learning similar to image recognition is adopted for the acquired fragment images, the steps of convolution and pooling are included, the newly extracted images can be directly classified through a classification model of the neural network learning, whether the newly extracted images belong to abnormal curves or not is judged, and engineers are reminded, and the steps have the advantages that the computational complexity is greatly reduced through convolution and pooling of the convolution neural network, and the neural network construction is carried out on the newly extracted images, so that intelligent classification is realized. The specific steps are shown in fig. 8:
step S801: and carrying out segment division on the single-day driving data of the lithium battery to obtain segment data.
Optionally, the main purpose of this step is to divide the data of a single day into segments according to the usage state of the battery, and to realize equal-length, equal-width and equal-density grid division of the graph: dividing the single-day driving data of the lithium battery into segments; ordering data over a period of time (typically 1 day) per trolley or per battery pack; the working state of each trolley or power battery is divided into: travel, charge, park, signals specified by GBT32960, specific divisions may be referenced in table 1.
Step S802: and dividing grid lines of the temperature/voltage interval in the abscissa time axis and in the ordinate time axis to obtain a two-dimensional grid diagram of the segment data.
Optionally, the grid lines of the abscissa and ordinate axes are divided, and the main purpose of this step is to obtain the minimum granularity of the segments, and to perform gridding division: voltage ordinate, for voltage data, ordinate net line division is performed in units of minimum voltage resolution, typically mV; temperature ordinate, for temperature data, ordinate network line division is performed in units of minimum temperature resolution, typically, in degrees celsius; the time abscissa, the abscissa uses the minimum recognition unit of time, usually uses 10s as the unit to divide the abscissa network line; at this time, a voltage two-dimensional grid pattern with time on the abscissa and voltage on the ordinate and a temperature two-dimensional grid pattern with time on the abscissa and temperature on the ordinate are realized.
Step S803: and sampling grid lines and interpolating the equal density distribution to obtain a two-dimensional grid graph after homogenization.
Optionally, the equal length, equal width and equal density are set, and the purpose of the step is to homogenize all the voltage two-dimensional grid patterns or the temperature two-dimensional grid patterns by an interpolation or sampling method: the segments of driving, charging and stopping are respectively screened and classified, and segments with similar time length are classified, for example, quick charging in the charging segments is generally about 1 hour, and the segments with the minimum time length and the maximum time length in the ratio of 4:5 are generally selected as alternative segments; carrying out homogenization treatment on the ordinate of the alternative fragment, wherein the voltage of the monomer is generally 2.5-4.2V, the temperature is generally between-20 ℃ and-60 ℃, and abnormal points are removed according to a common data cleaning rule; homogenizing the abscissa of the alternative segment, taking the median round according to the length of the segment, taking 1 hour as a standard for quick charging of the segment, sampling the segment with the time longer than the standard segment according to the proportion of the standard length/the segment length, and sampling the segment with equal proportion, wherein the obtained abscissa density is the same as the standard segment density; similarly, the time is shorter than the standard segment, the segment is interpolated according to the proportion of the standard time length/the segment time length, the interpolation method is not limited to the modes of average interpolation, equal ratio interpolation and the like, and the obtained abscissa density is the same as the standard segment density. At this time, all the fragments to be analyzed are homogenized on the abscissa
Step S804: the function curve graph is gridded to 0-1 to obtain a uniform matrix.
Optionally, the graphically gridding the function curve 0-1 representation further comprises: the imaging representation of the function curve generally analyzes the abnormality of the voltage curve, and when the function curve is judged manually, the change of the highest value and the lowest value of the single voltage is analyzed to find the voltage fault trend of the battery, such as consistency deterioration, abnormal internal resistance, abnormal capacity and the like, and the aim of the step is to realize the drawing of the voltage or temperature data to be analyzed, so as to provide a basis for realizing automatic graph judgment for machine learning: taking voltage curve change as an example, in the quick charge segment, the curves of the highest, lowest and median of the monomer voltages along with time are selected as curves to be judged. Extracting the highest, lowest and median values of the monomer voltages in the corresponding time, wherein the minimum unit of the values is the same as the minimum unit of the abscissa, and the excess part is rounded; the corresponding position of the segment in step 803 is marked as 1, the rest positions are marked as 0, and a matrix with voltage as ordinate and time as abscissa is formed, wherein the coordinate positions corresponding to the highest, lowest and median values of the monomer voltages are 1, and the rest positions are 0.
The method is used for describing the segment selection, the abscissa normalization method and the ordinate normalization method which are needed to be carried out on the segment division in detail, and most of battery use scenes can be contained and classified through the method, and artificial intelligence recognition can be carried out through a follow-up scheme.
Step S805: and setting a convolution kernel so that the convolution kernel can screen out curve characteristics to be analyzed in the curve image.
Alternatively, the main purpose of this step is to set the convolution kernel, i.e. the most basic filter label, through the CNN convolutional neural network, and to manually label each segment image in the training set to form a training set. The convolution kernel setting, the purpose of this step is to set the convolution kernel passing through the CNN convolution neural network, i.e. the most basic filtering label, to screen out the curve characteristic to be analyzed, and at the same time, the convolution kernel setting size is to be matched with the segment image, to benefit the complete coverage, and the convolution kernel setting includes the size and matrix characteristic of the convolution kernel, in this embodiment, the convolution kernel of 3*3 is taken as an example, and has 6 convolution kernels, which respectively represent the curve level, the curve vertical, the curve downward extension, the curve upward mutation, the curve downward mutation, and can describe the basic curve change.
Step S806: whether the curve image is a training set is determined, if so, step S807 is executed, and if not, step S811 is executed.
Optionally, determining whether the training set further includes: and (3) manually marking each segment image in the training set, and manually classifying and marking the segment images in each training set, wherein the labeling treatment is performed on the segment images of each training set, such as the phenomenon that the battery cell SOC deviates due to self-discharge, the phenomenon that the cell is high or low, and the like.
The method is used for describing the convolution kernel setting method and operation examples aiming at the battery voltage and temperature characteristics in detail, and comprises the size and matrix characteristics of convolution kernels, and the significance of each convolution kernel filter.
Step S807: and convolving the curve image to obtain the local features in the curve image.
Optionally, learning and classifying, wherein the main purpose of the step is to perform CNN neural network learning similar to image recognition on the obtained fragment images, including convolution and pooling steps, and the classification model through the neural network learning can directly classify the newly extracted images, judge whether the newly extracted images belong to abnormal curves or not, and remind engineers. The curve image is convolved, the convolution layer is a core layer for constructing a convolutional neural network, the convolution kernels are scanned across the whole image one by one according to the corresponding size by taking a minimum time unit and a minimum vertical axis unit as step length, and each convolution kernel is convolved with a local area of input data respectively. Thus, the characteristic values of the small areas are obtained, if the characteristic values have 6 convolution kernels, 6 characteristic value matrixes are generated, and the local characteristics in the picture are extracted by the convolution kernel filtering of the convolution layer.
Step S808: the feature matrix is pooled, so that the inner dimension of the feature matrix can be reduced, and the operation amount is reduced.
Optionally, in order to further reduce the dimension of the feature matrix, reduce the operand, effectively avoid overfitting, and perform further pooling operation on the feature matrix. The pooling method generally comprises the following steps of covering the whole feature matrix with N specific matrices and taking the feature values by adopting a specific square matrix, usually 1/N of the feature matrix, and adopting the following steps: max pooling, mean pooling, gao Sichi pooling, trainable pooling.
Step S809: and performing neural network learning on the images in the training set to obtain a complete neural network model.
Optionally, the neural network learning further includes: the fully connected neural network, which is the same as a conventional fully connected neural network, can be converted between a fully connected layer and a convolutional layer, and maps the learned feature representation to a marker space of the sample. Training is carried out according to the classified images of the training set, and finally a complete neural network model is obtained.
Step S810: and inputting the image to be evaluated into a model and classifying, so that the fault type of the power battery can be obtained, and the process is ended.
Optionally, the image to be evaluated is input into a model, classified, a fault of which temperature or voltage the power battery is obtained, and then the flow is ended.
Step S811: and (5) manually judging whether the risk graph is or not, and labeling.
The method describes the convolution and pooling operations in detail, can reduce the image calculation dimension, greatly improves the operation efficiency, and realizes the classification and identification of AI through the fully connected neural network.
The invention provides a power battery voltage and temperature characteristic graphical fault identification method, which comprises the steps of firstly dividing the use state of a lithium ion battery into segments, marking the segments, manually judging whether faults and fault types exist or not, meshing the data, and identifying the faults in a neural network mode. The fault identification of the lithium ion battery which does not depend on the threshold characteristic is realized.
Example 2
According to an embodiment of the present invention, there is further provided a computer readable storage medium, where the computer readable storage medium includes a stored program, and when the program runs, the apparatus in which the computer readable storage medium is controlled to execute the method for processing a power battery according to embodiment 1.
Example 3
According to an embodiment of the present invention, there is also provided a computer terminal including: the power battery processing device comprises a memory and a processor, wherein the processor is used for running a program stored in the memory, and the processing method of the power battery described in the embodiment 1 is executed when the program runs.
Example 4
According to the embodiment of the present invention, a processing device for a power battery is provided, which can execute the processing method for a power battery provided in the foregoing embodiment 1, and a specific implementation manner and a preferred application scenario are the same as those of the foregoing embodiment 1, and are not described herein.
Fig. 9 is a schematic structural view of a processing device for a power battery according to an embodiment of the present invention, as shown in fig. 9, the device including: the first obtaining module 92 is configured to obtain target feature data of the target power battery, where the target feature data is data corresponding to a target feature of the target power battery; a first processing module 94, configured to matrix the target feature data to generate a target matrix; and the recognition module 96 is configured to recognize the target matrix by using the fault recognition model, so as to obtain a recognition result of the target power battery, where the recognition result is used to indicate whether the target power battery has a fault.
Optionally, the first processing module includes: the first dividing unit is used for dividing the target feature data based on the preset battery state to generate target state data, wherein the target state data is data corresponding to the preset battery state in the target feature data; the first processing unit is used for carrying out matrixing processing on the target state data to generate a target matrix.
Optionally, the first processing module further comprises: the acquisition unit is used for acquiring a target graph corresponding to the target state data; the second dividing unit is used for carrying out grid division on the target curve graph by utilizing preset granularity and generating an initial grid curve graph; the second processing unit is used for carrying out homogenization processing on the initial grid graph and generating a target grid graph; and the third processing unit is used for matrixing the target grid graph to generate a target matrix.
Optionally, the second processing unit includes: the acquisition subunit is used for acquiring target fragments which accord with preset rules in the grid curve image; and the processing subunit is used for carrying out homogenization processing on the target coordinates of the target segment to generate a target grid graph.
Optionally, the processing subunit is further configured to perform a homogenization process on an initial abscissa of the target segment to obtain a target abscissa, where the initial abscissa is used to represent an initial duration of the target segment; carrying out homogenization treatment on an initial ordinate of the target segment to obtain a target ordinate, wherein the initial ordinate is used for representing an initial feature range corresponding to the target feature; a target grid graph is generated based on the target abscissa and the target ordinate.
Optionally, the processing subunit is further configured to obtain a duration of the target segment; comparing the preset time length with the duration time length, and determining a comparison result, wherein the comparison result is used for indicating whether the preset time length is longer than the duration time length; sampling an initial abscissa based on a preset proportion under the condition that the preset duration is smaller than the duration, and obtaining a sampling result, wherein the preset proportion is the proportion of the preset duration to the initial duration of the target segment; and carrying out homogenization treatment on the initial abscissa based on the sampling result to obtain the target abscissa.
Optionally, the apparatus further comprises: the interpolation module is used for interpolating the initial abscissa based on a preset proportion under the condition that the preset time length is longer than or equal to the duration time length to obtain an interpolation result; and the second processing module is used for carrying out homogenization processing on the initial abscissa based on the interpolation result to obtain the target abscissa.
Optionally, the processing subunit is further configured to determine a target cleaning rule corresponding to the target feature; and removing the target coordinates in the initial ordinate of the target segment based on the target cleaning rule to obtain the target ordinate, wherein the target coordinates are abnormal coordinates in the initial ordinate.
Optionally, the apparatus further comprises: the second acquisition module is configured to acquire a target training set, where the target training set includes: the system comprises a sample matrix, at least one piece of label information corresponding to the sample matrix, at least one piece of label information and at least one piece of label information, wherein the at least one piece of label information is used for indicating whether at least one area in the sample matrix has faults or not, and the at least one piece of label information corresponds to the at least one area; the generating module is used for constructing an initial model based on a plurality of preset convolution kernels; and the training module is used for training the initial model based on the target training set to obtain a fault identification model.
The foregoing embodiment numbers of the present invention are merely for the purpose of description, and do not represent the advantages or disadvantages of the embodiments.
In the foregoing embodiments of the present invention, the descriptions of the embodiments are emphasized, and for a portion of this disclosure that is not described in detail in this embodiment, reference is made to the related descriptions of other embodiments.
In the several embodiments provided in the present application, it should be understood that the disclosed technology may be implemented in other manners. The above-described embodiments of the apparatus are merely exemplary, and the division of the units, for example, may be a logic function division, and may be implemented in another manner, for example, a plurality of units or components may be combined or may be integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be through some interfaces, units or modules, or may be in electrical or other forms.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.
The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied essentially or in part or all of the technical solution or in part in the form of a software product stored in a storage medium, including instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a usb disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a removable hard disk, a magnetic disk, or an optical disk, or other various media capable of storing program codes.
The foregoing is merely a preferred embodiment of the present invention and it should be noted that modifications and adaptations to those skilled in the art may be made without departing from the principles of the present invention, which are intended to be comprehended within the scope of the present invention.
Claims (7)
1. A method for identifying a failure of a power cell, comprising:
Obtaining target feature data of a target power battery, wherein the target feature data is data corresponding to target features of the target power battery, and the target feature data comprises: characteristic data of the target power battery when the vehicle is running, characteristic data of the target power battery when the vehicle is charging, characteristic data of the target power battery when the vehicle is in a stopped state;
performing matrixing treatment on the target characteristic data to generate a target matrix;
Identifying the target matrix by using a fault identification model to obtain an identification result of the target power battery, wherein the identification result is used for indicating whether the target power battery has a fault or not;
The method for identifying the target matrix by using the fault identification model to obtain the identification result of the target power battery comprises the following steps: and identifying the curve change characteristics of the curve image in the target matrix by utilizing six convolution kernels in the fault identification model to obtain the identification result, wherein the six convolution kernels respectively correspond to the curve change characteristics, and the curve change characteristics comprise: curve horizontal, curve vertical, curve downward extending, curve upward abrupt change, curve downward abrupt change;
The matrixing processing is performed on the target feature data to generate a target matrix, which comprises the following steps: dividing the target feature data based on a preset battery state to generate target state data, wherein the target state data is data corresponding to the preset battery state in the target feature data; performing matrixing treatment on the target state data to generate a target matrix;
wherein, the matrixing processing is performed on the target state data to generate a target matrix, which comprises: acquiring a target graph corresponding to the target state data; performing grid division on the target graph by using preset granularity to generate an initial grid graph; carrying out homogenization treatment on the initial grid graph to generate a target grid graph; matrixing the target grid graph to generate the target matrix;
The fault identification method of the power battery further comprises the following steps: obtaining a target training set, wherein the target training set comprises: the system comprises a sample matrix, at least one piece of label information corresponding to the sample matrix, at least one piece of label information and a plurality of pieces of label information, wherein the at least one piece of label information is used for indicating whether at least one area in the sample matrix has a fault or not, and the at least one piece of label information corresponds to the at least one area; constructing an initial model based on the six convolution kernels; and training the initial model based on the target training set to obtain the fault recognition model.
2. The method of claim 1, wherein homogenizing the initial grid graph to generate a target grid graph comprises:
Obtaining a target segment which accords with a preset rule in the initial grid graph;
And carrying out homogenization treatment on the target coordinates of the target segment to generate the target grid graph.
3. The method of claim 2, wherein homogenizing the target coordinates of the target segment to generate the target grid graph comprises:
Carrying out homogenization treatment on the initial abscissa of the target segment to obtain a target abscissa, wherein the initial abscissa is used for representing the initial duration of the target segment;
Carrying out homogenization treatment on the initial ordinate of the target segment to obtain a target ordinate, wherein the initial ordinate is used for representing an initial feature range corresponding to the target feature;
the target grid graph is generated based on the target abscissa and the target ordinate.
4. A method according to claim 3, wherein the homogenizing the initial abscissa of the target segment to obtain the target abscissa comprises:
Acquiring the duration of the target fragment;
Comparing a preset time length with the duration time length, and determining a comparison result, wherein the comparison result is used for indicating whether the preset time length is larger than the duration time length or not;
sampling the initial abscissa based on a preset proportion under the condition that the preset duration is smaller than the duration to obtain a sampling result, wherein the preset proportion is the proportion of the preset duration to the initial duration of the target segment;
and carrying out homogenization treatment on the initial abscissa based on the sampling result to obtain the target abscissa.
5. The method according to claim 4, wherein the method further comprises:
Under the condition that the preset time length is greater than or equal to the duration, interpolating the initial abscissa based on the preset proportion to obtain an interpolation result;
And carrying out homogenization treatment on the initial abscissa based on the interpolation result to obtain the target abscissa.
6. A method according to claim 3, wherein the homogenizing the initial ordinate of the target segment to obtain the target ordinate comprises:
determining a target cleaning rule corresponding to the target feature;
and removing the target coordinates in the initial ordinate of the target segment based on the target cleaning rule to obtain the target ordinate, wherein the target coordinates are abnormal coordinates in the initial ordinate.
7. A computer-readable storage medium, characterized in that the computer-readable storage medium comprises a stored program, wherein the program, when run, controls a device in which the computer-readable storage medium is located to perform the fault identification method of the power battery according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210376149.4A CN114742297B (en) | 2022-04-11 | 2022-04-11 | Method for processing power battery |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210376149.4A CN114742297B (en) | 2022-04-11 | 2022-04-11 | Method for processing power battery |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114742297A CN114742297A (en) | 2022-07-12 |
| CN114742297B true CN114742297B (en) | 2024-05-24 |
Family
ID=82281117
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210376149.4A Active CN114742297B (en) | 2022-04-11 | 2022-04-11 | Method for processing power battery |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114742297B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116154239B (en) * | 2023-04-14 | 2023-11-21 | 湖南省计量检测研究院 | Multi-level implementation-based hydrogen fuel cell energy conversion method and device |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108445406A (en) * | 2018-03-13 | 2018-08-24 | 桂林电子科技大学 | A kind of power battery health status method of estimation |
| KR20210016154A (en) * | 2019-07-31 | 2021-02-15 | 주식회사 에스제이 테크 | Battery diagnostic methods using machine learning |
| CN112611976A (en) * | 2020-12-10 | 2021-04-06 | 华东理工大学 | Power battery state of health estimation method based on double differential curves |
| CN112668715A (en) * | 2020-12-28 | 2021-04-16 | 卡斯柯信号有限公司 | Turnout switch machine abnormity diagnosis method and system based on machine learning |
| CN113343633A (en) * | 2021-06-10 | 2021-09-03 | 上海交通大学 | Thermal runaway fault classification and risk prediction method and system for power lithium battery |
| WO2021185308A1 (en) * | 2020-03-18 | 2021-09-23 | 北京理工大学 | Online determination method and system for state of health of power battery pack of electric vehicle |
| CN113608140A (en) * | 2021-06-25 | 2021-11-05 | 国网山东省电力公司泗水县供电公司 | Battery fault diagnosis method and system |
| CN113821875A (en) * | 2021-09-26 | 2021-12-21 | 湖南大学 | Intelligent vehicle fault real-time prediction method and system based on end cloud cooperation |
| CN113900035A (en) * | 2021-09-28 | 2022-01-07 | 深圳市科陆电子科技股份有限公司 | Battery detection method, device, equipment and storage medium |
| CN113917337A (en) * | 2021-10-13 | 2022-01-11 | 国网福建省电力有限公司 | Battery health state estimation method based on charging data and LSTM neural network |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20200210824A1 (en) * | 2018-12-28 | 2020-07-02 | Utopus Insights, Inc. | Scalable system and method for forecasting wind turbine failure with varying lead time windows |
-
2022
- 2022-04-11 CN CN202210376149.4A patent/CN114742297B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108445406A (en) * | 2018-03-13 | 2018-08-24 | 桂林电子科技大学 | A kind of power battery health status method of estimation |
| KR20210016154A (en) * | 2019-07-31 | 2021-02-15 | 주식회사 에스제이 테크 | Battery diagnostic methods using machine learning |
| WO2021185308A1 (en) * | 2020-03-18 | 2021-09-23 | 北京理工大学 | Online determination method and system for state of health of power battery pack of electric vehicle |
| CN112611976A (en) * | 2020-12-10 | 2021-04-06 | 华东理工大学 | Power battery state of health estimation method based on double differential curves |
| CN112668715A (en) * | 2020-12-28 | 2021-04-16 | 卡斯柯信号有限公司 | Turnout switch machine abnormity diagnosis method and system based on machine learning |
| CN113343633A (en) * | 2021-06-10 | 2021-09-03 | 上海交通大学 | Thermal runaway fault classification and risk prediction method and system for power lithium battery |
| CN113608140A (en) * | 2021-06-25 | 2021-11-05 | 国网山东省电力公司泗水县供电公司 | Battery fault diagnosis method and system |
| CN113821875A (en) * | 2021-09-26 | 2021-12-21 | 湖南大学 | Intelligent vehicle fault real-time prediction method and system based on end cloud cooperation |
| CN113900035A (en) * | 2021-09-28 | 2022-01-07 | 深圳市科陆电子科技股份有限公司 | Battery detection method, device, equipment and storage medium |
| CN113917337A (en) * | 2021-10-13 | 2022-01-11 | 国网福建省电力有限公司 | Battery health state estimation method based on charging data and LSTM neural network |
Non-Patent Citations (2)
| Title |
|---|
| A novel deep learning framework for state of health estimation of lithium-ion battery;Fan, YX 等;《JOURNAL OF ENERGY STORAGE》;20201231;第1-6页 * |
| 锂离子电池状态估计机器学习方法综述;谢奕展 等;《汽车文摘》;第43卷(第11期);第26-35页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114742297A (en) | 2022-07-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10896351B2 (en) | Active machine learning for training an event classification | |
| KR20180118596A (en) | Semi-automatic labeling of data sets | |
| CN114529873A (en) | Target detection method and city violation event monitoring method applying same | |
| CN106570439B (en) | Vehicle detection method and device | |
| CN110008082B (en) | Abnormal task intelligent monitoring method, device, equipment and storage medium | |
| CN110826429A (en) | Scenic spot video-based method and system for automatically monitoring travel emergency | |
| CN111274926B (en) | Image data screening method, device, computer equipment and storage medium | |
| CN111046910A (en) | Image classification, relation network model training and image annotation method and device | |
| CN114742297B (en) | Method for processing power battery | |
| CN112101114B (en) | Video target detection method, device, equipment and storage medium | |
| CN110599453A (en) | Panel defect detection method and device based on image fusion and equipment terminal | |
| CN111582032A (en) | Pedestrian detection method and device, terminal equipment and storage medium | |
| US11120308B2 (en) | Vehicle damage detection method based on image analysis, electronic device and storage medium | |
| CN108182444A (en) | The method and device of video quality diagnosis based on scene classification | |
| CN116872961B (en) | Control system for intelligent driving vehicle | |
| CN113401130B (en) | Driving style recognition method and device based on environmental information and storage medium | |
| US11256950B2 (en) | Image feature amount output device, image recognition device, the image feature amount output program, and image recognition program | |
| CN111382638A (en) | Image detection method, device, equipment and storage medium | |
| CN111339834B (en) | Method for identifying vehicle driving direction, computer device and storage medium | |
| CN110579708B (en) | Battery capacity identification method and device, computing equipment and computer storage medium | |
| US20240092394A1 (en) | Generating unknown-unsafe scenarios, improving automated vehicles, and computer system | |
| CN114118413A (en) | Network training and equipment control method, device, equipment and storage medium | |
| CN113537087A (en) | Intelligent traffic information processing method and device and server | |
| CN113269678A (en) | Fault point positioning method for contact network transmission line | |
| Bäuerle et al. | Training de-confusion: an interactive, network-supported visual analysis system for resolving errors in image classification training data |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |