CN114490836A - Data mining processing method suitable for electric vehicle charging fault - Google Patents

Data mining processing method suitable for electric vehicle charging fault Download PDF

Info

Publication number
CN114490836A
CN114490836A CN202210392883.XA CN202210392883A CN114490836A CN 114490836 A CN114490836 A CN 114490836A CN 202210392883 A CN202210392883 A CN 202210392883A CN 114490836 A CN114490836 A CN 114490836A
Authority
CN
China
Prior art keywords
data
fault
data mining
electric vehicle
charging facility
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.)
Granted
Application number
CN202210392883.XA
Other languages
Chinese (zh)
Other versions
CN114490836B (en
Inventor
张剑
徐科
王迎秋
张�浩
贺春
祖国强
刘洪�
李思维
郝爽
李少雄
袁新润
谢秦
李硕
赵越
张卫欣
唐庆华
甘智勇
张弛
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tianjin University
State Grid Corp of China SGCC
State Grid Information and Telecommunication Co Ltd
State Grid Tianjin Electric Power Co Ltd
Electric Power Research Institute of State Grid Tianjin Electric Power Co Ltd
Nari Technology Co Ltd
Original Assignee
Tianjin University
State Grid Corp of China SGCC
State Grid Information and Telecommunication Co Ltd
State Grid Tianjin Electric Power Co Ltd
Electric Power Research Institute of State Grid Tianjin Electric Power Co Ltd
Nari Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Tianjin University, State Grid Corp of China SGCC, State Grid Information and Telecommunication Co Ltd, State Grid Tianjin Electric Power Co Ltd, Electric Power Research Institute of State Grid Tianjin Electric Power Co Ltd, Nari Technology Co Ltd filed Critical Tianjin University
Priority to CN202210392883.XA priority Critical patent/CN114490836B/en
Publication of CN114490836A publication Critical patent/CN114490836A/en
Application granted granted Critical
Publication of CN114490836B publication Critical patent/CN114490836B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/24—Querying
    • G06F16/245—Query processing
    • G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
    • G06F16/2465—Query processing support for facilitating data mining operations in structured databases
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • 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
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06395—Quality analysis or management
    • 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
    • G06Q10/20—Administration of product repair or maintenance
    • 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
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Economics (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • General Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Development Economics (AREA)
  • Educational Administration (AREA)
  • Health & Medical Sciences (AREA)
  • Probability & Statistics with Applications (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Fuzzy Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Computational Linguistics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Game Theory and Decision Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Electric Propulsion And Braking For Vehicles (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

The invention relates to a data mining processing method suitable for charging faults of electric vehicles, which is characterized in that the output of the data mining processing method is input into a charging facility running state data mining model, complex data is simplified, a proper charging facility data mining scheme is selected, a reasonable data mining model is designed, safety fault data characteristics are analyzed, data mining results are displayed, safety fault data characteristics are analyzed, a charging facility integrated fault tree is established, and the relation between each layer of fault source and a specific fault type is analyzed through extraction rules. The improved bad data screening algorithm removes errors caused by random generation of a reference number set and adoption of sample estimation in the original gap statistical algorithm, so that statistics can reflect facts more objectively. The algorithm does not need to estimate samples, calculate standard deviation and simulation error, reduces the calculation amount of the algorithm, improves the calculation speed of the algorithm and enhances the practicability of the algorithm.

Description

Data mining processing method suitable for electric vehicle charging fault
Technical Field
The invention belongs to the technical field of charging equipment data processing, and particularly relates to a data mining processing method suitable for charging faults of an electric vehicle.
Background
Compared with the traditional fuel automobile, the electric automobile is environment-friendly and noiseless, the development of the electric automobile is beneficial to relieving energy shortage and environmental pollution, and the popularization of the electric automobile charging pile is the premise of the development of the electric automobile. When the charging pile is in a working state, massive data can be generated, various data types, attributes and the like of an initial data set are diversified, and difficulties are brought to follow-up statistics and analysis of the data. Meanwhile, charging safety accidents of the electric automobile, particularly charging facility faults, frequently occur, so that the property safety of electric automobile users and operators is influenced, and the development of the electric automobile industry is restricted to a certain extent. Therefore, how to adopt data mining technology to analyze the safety failure characteristics of the charging facility operation becomes the development trend of the electric automobile, but now only stays in the idea and is not capable of performing calculation.
Disclosure of Invention
The invention aims to overcome the defects of the prior art, and provides a data mining processing method suitable for electric vehicle charging faults.
The technical problem to be solved by the invention is realized by adopting the following technical scheme:
a data mining processing method suitable for electric vehicle charging faults comprises the following steps:
step 1: collecting running state data of an electric vehicle charging facility through the Internet, a battery management system and an operation monitoring platform;
step 2: identifying bad data in the operation state data collected in the step 1;
and step 3: correcting the bad data identified in the step 2 by using an echo state network;
and 4, step 4: performing quality evaluation on the running state data of the electric vehicle charging facility based on the correction result of the bad data in the step 3, and meanwhile, constructing a quality evaluation system and a quality evaluation model of the running state data of the charging facility according to the characteristics of the running state data;
and 5: performing data conversion of data generalization, data normalization and data attribute construction on the charging facility running state data quality evaluation system and the quality evaluation model in the step 4;
step 6: constructing a charging facility running state data mining model, inputting the output data after data conversion in the step 5 into the charging facility running state data mining model for feature data mining, and simplifying complex data;
and 7: and 6, constructing a charging facility integrated fault tree according to the simplified characteristic data in the step 6, and analyzing the relation between each layer of fault source and the specific fault type through extracting rules.
Further, the operation state data in step 1 includes: voltage, current, and temperature.
Furthermore, the identifying of the bad data in the step 2 includes the following steps:
step 2.1, clustering the running state data in the step 1, wherein the clustering number is
Figure 100002_DEST_PATH_IMAGE001
And is and
Figure 86590DEST_PATH_IMAGE001
=1、2、……
Figure 98671DEST_PATH_IMAGE002
，
Figure 347249DEST_PATH_IMAGE002
calculating cluster dispersion corresponding to different cluster numbers for the total number of clusters
Figure 100002_DEST_PATH_IMAGE003
；
Step 2.2, calculating the expectation of clustering dispersion corresponding to different clustering numbers by adopting uniform distribution as reference distributionE(W r,k ) WhereinrFor generating based on characteristics of the raw datarA uniformly distributed reference data set of individuals;
step 2.3, calculating the clustering number
Figure 601513DEST_PATH_IMAGE001
=1、2、……
Figure 884727DEST_PATH_IMAGE002
Intermediate variables of time correspondence
Figure 35086DEST_PATH_IMAGE004
A value;
step 2.4, calculated according to step 2.3
Figure 770960DEST_PATH_IMAGE004
And obtaining the optimal clustering number and identifying bad data.
Also, intermediate variables in said step 2.3
Figure 563336DEST_PATH_IMAGE004
The calculation method comprises the following steps:
Figure 100002_DEST_PATH_IMAGE005
wherein the content of the first and second substances,
Figure 966635DEST_PATH_IMAGE006
the mathematical expectation value in the case where the sample size is n is distributed for the reference value.
Moreover, the specific implementation method of the step 2.4 is as follows: for those calculated in step 2.3
Figure 57869DEST_PATH_IMAGE001
An
Figure 281040DEST_PATH_IMAGE004
Make a comparison if
Figure 100002_DEST_PATH_IMAGE007
Then, then
Figure 814790DEST_PATH_IMAGE008
Is the best cluster number, otherwise order
Figure DEST_PATH_IMAGE009
And returns to step 2.1.
Furthermore, the reserve battery state updating formula of the echo state network in the step 3
Figure 197229DEST_PATH_IMAGE010
Comprises the following steps:
Figure DEST_PATH_IMAGE011
output of the echo state network:
Figure 689391DEST_PATH_IMAGE012
wherein the content of the first and second substances,
Figure DEST_PATH_IMAGE013
the specific value of the adjustment factor for the reserve battery depends on the actual situation,
Figure 494798DEST_PATH_IMAGE014
is the function of the activation of the function,
Figure DEST_PATH_IMAGE015
is the data that is input to the computer,
Figure 566659DEST_PATH_IMAGE016
is either the network state or the reserve battery state,
Figure DEST_PATH_IMAGE017
is the input of a matrix of weights,
Figure 944551DEST_PATH_IMAGE018
is the intermediate weight matrix and the intermediate weight matrix,
Figure DEST_PATH_IMAGE019
is the output of the computer system,
Figure 342034DEST_PATH_IMAGE020
is the output weight matrix.
Moreover, the quality evaluation system of the charging facility running state data in the step 4 comprises the inherent quality, accessible quality, context quality and representation quality of the running state data; the charging facility operation state data quality evaluation model comprises a decision relation, a detection relation, a business rule, an algorithm library and a rule weight.
Moreover, the feature data mining in step 6 includes the following steps:
step 6.1, expanding the classification information to the output data after the data conversion in the step 5;
Figure DEST_PATH_IMAGE021
as the number of the categories,
Figure 867693DEST_PATH_IMAGE022
for the output data after data conversionXIf each classification information of (1) is output data
Figure DEST_PATH_IMAGE023
Belong to the class
Figure 304097DEST_PATH_IMAGE024
Then, then
Figure DEST_PATH_IMAGE025
Class label of
Figure 864392DEST_PATH_IMAGE026
Is set to
Figure DEST_PATH_IMAGE027
Otherwise
Figure 432776DEST_PATH_IMAGE028
Is arranged as
Figure DEST_PATH_IMAGE029
(ii) a Determining
Figure 445732DEST_PATH_IMAGE030
And
Figure 593816DEST_PATH_IMAGE029
after that, make
Figure 415142DEST_PATH_IMAGE026
For all
Figure 452630DEST_PATH_IMAGE023
All have normal distributions; at the same time classifying information
Figure DEST_PATH_IMAGE031
Expansion to output data
Figure 687302DEST_PATH_IMAGE025
Need to output data
Figure 576761DEST_PATH_IMAGE025
Performing standardization to output dataXEach factor of (a) is normally distributed, and the normalized output data is
Figure 642806DEST_PATH_IMAGE032
Then expand the data
Figure DEST_PATH_IMAGE033
Comprises the following steps:
Figure 490676DEST_PATH_IMAGE034
wherein the normalized output data
Figure 9382DEST_PATH_IMAGE032
Has the dimension of
Figure DEST_PATH_IMAGE035
The number of input and output data;
Figure 523104DEST_PATH_IMAGE031
has the dimension of
Figure 53442DEST_PATH_IMAGE036
Has the dimension of
Figure DEST_PATH_IMAGE037
As data
Figure 196848DEST_PATH_IMAGE025
The number of classification categories of (a);
step 6.2, expanding data in the step 6.1 based on a principal component analysis method
Figure 78216DEST_PATH_IMAGE038
Carrying out feature extraction;
the initial principal component PCs dimension is
Figure DEST_PATH_IMAGE039
If the dimensionality of the input and output data is to be reduced
Figure 902952DEST_PATH_IMAGE040
Until it is less than
Figure DEST_PATH_IMAGE041
Then need to select
Figure 146852DEST_PATH_IMAGE040
A principal component PCs, and obtaining it
Figure 602104DEST_PATH_IMAGE040
Maximum variance, wherein in a certain principal component PCs, the matrix is transformed
Figure 596867DEST_PATH_IMAGE042
Comprises the following steps:
Figure DEST_PATH_IMAGE043
wherein the content of the first and second substances,
Figure 959715DEST_PATH_IMAGE044
is the current principal component
Figure DEST_PATH_IMAGE045
The number of the categories is one,
Figure 323701DEST_PATH_IMAGE044
has the dimension of
Figure 949854DEST_PATH_IMAGE046
In order to transform the matrix, the matrix is,
Figure 805815DEST_PATH_IMAGE042
has the dimension of
Figure DEST_PATH_IMAGE047
To correspond toIn order to expand the amount of input/output data,
Figure 972354DEST_PATH_IMAGE048
to correspond to the class of the augmented input-output data,
Figure DEST_PATH_IMAGE049
the number of main components of the data after the characteristic extraction is adopted;
step 6.3, transforming the matrix
Figure 423801DEST_PATH_IMAGE042
For extended data
Figure 486435DEST_PATH_IMAGE033
Carrying out coordinate transformation;
by transforming matrices
Figure 688747DEST_PATH_IMAGE042
And extended data
Figure 331080DEST_PATH_IMAGE050
Multiplying to obtain the conversion data on the characteristic space
Figure DEST_PATH_IMAGE051
，
Figure 138499DEST_PATH_IMAGE052
And designing a data mining model through charging facility data mining, analyzing safety fault data characteristics and displaying a data mining result.
Moreover, the step 7 of establishing the fault tree includes the following steps:
7.1, establishing a fault tree, wherein the root of all faults in the running state data of the electric vehicle charging facility collected through the Internet, the battery management system and the operation monitoring platform is used as a statement of a top event;
7.2, finding out bad data in the running state data as an input event, and obtaining converted characteristic data as an output event;
step 7.3, repeating step 7.2 until a bottom event is obtained;
and 7.4, connecting the events of all levels by using a logic symbol to form a fault tree.
The invention has the advantages and positive effects that:
according to the invention, the output of the charging facility is input into the charging facility running state data mining model, the complex data is simplified, the charging facility data mining scheme is selected, the data mining model is designed, the safety fault data characteristics are analyzed to show the data mining result, the safety fault data characteristics are analyzed, the charging facility integrated fault tree is established, and the relation between each layer of fault source and the specific fault type is analyzed through extracting rules. The improved bad data screening algorithm removes errors caused by random generation of a reference number set and adoption of sample estimation in the original gap statistical algorithm, so that statistics can reflect facts more objectively. The algorithm does not need to estimate samples, calculate standard deviation and simulation error, reduces the calculation amount of the algorithm, improves the calculation speed of the algorithm and enhances the practicability of the algorithm.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a flow chart of the algorithm for improving GSA for identifying bad data in step 2 according to the present invention;
FIG. 3 is a data quality evaluation index system diagram of the operating state of a charging facility according to the present invention;
FIG. 4 is a model diagram of the electric vehicle charge and discharge data quality evaluation according to the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings.
A data mining processing method suitable for electric vehicle charging faults is shown in FIG. 1, and comprises the following steps:
step 1: the running state data of the electric automobile charging facility is collected through the internet, the battery management system and the operation monitoring platform.
The operating state data includes: voltage, current, and temperature.
Step 2: and (3) identifying bad data in the operation state data collected in the step (1).
In the traditional calculation method, the gap statistical algorithm is combined with clustering analysis to have many advantages in the aspect of detecting and identifying bad data of the power system, but when massive data in an interconnected large power grid are encountered, the calculation amount is large, the calculation speed is influenced, and therefore the practicability of the algorithm in the aspect of real-time calculation is influenced. In order to make the algorithm suitable for the calculation of mass data, the invention improves the algorithm as shown in fig. 2.
The identification of bad data comprises the following steps:
step 2.1, clustering the running state data in the step 1, wherein the clustering number is
Figure 699931DEST_PATH_IMAGE001
And is and
Figure 264904DEST_PATH_IMAGE001
=1、2、……
Figure 242088DEST_PATH_IMAGE002
，
Figure 202216DEST_PATH_IMAGE002
calculating cluster dispersion corresponding to different cluster numbers for the total number of clusters
Figure 809914DEST_PATH_IMAGE003
。
Step 2.2, calculating the expectation of clustering dispersion corresponding to different clustering numbers by adopting uniform distribution as reference distribution
Figure DEST_PATH_IMAGE053
WhereinrFor generating based on characteristics of the raw datarA uniformly distributed reference data set of individuals.
Step 2.3, calculating the clustering number
Figure 190080DEST_PATH_IMAGE001
=1、2、……
Figure 33271DEST_PATH_IMAGE002
Intermediate variables of time correspondence
Figure 487386DEST_PATH_IMAGE004
The value of the one or more of the one,
Figure 390620DEST_PATH_IMAGE004
as an intermediate variable for finally determining the minimum k value, no independent definition is made;
Figure 726924DEST_PATH_IMAGE005
wherein, the first and the second end of the pipe are connected with each other,
Figure 249172DEST_PATH_IMAGE006
indicating the mathematical expectation for a reference value distribution sample size of n (the magnitude of this expectation is related to n because of the calculation
Figure 186822DEST_PATH_IMAGE054
Implies the information related to n)
Step 2.4, calculated according to step 2.3
Figure 870744DEST_PATH_IMAGE004
And obtaining the optimal clustering number and identifying bad data.
For those calculated in step 2.3
Figure DEST_PATH_IMAGE055
An
Figure 225502DEST_PATH_IMAGE004
Make a comparison if
Figure 348179DEST_PATH_IMAGE007
Then, then
Figure 635941DEST_PATH_IMAGE008
Is the best cluster number, otherwise order
Figure 756343DEST_PATH_IMAGE009
And returns to step 2.1.
The improved algorithm removes errors caused by randomly generating a reference number set and adopting sample estimation in the original GSA method, so that the statistics can reflect the fact more objectively. The algorithm does not need to estimate samples, calculate standard deviation and simulation error, reduces the calculation amount of the algorithm, improves the calculation speed of the algorithm and enhances the practicability of the algorithm.
And step 3: and (3) correcting the bad data identified in the step (2) by using an echo state network. It is necessary to generate a sufficiently complex dynamic space by using the reserve battery in the network structure, and linearly combine the required corresponding outputs with the change of the input, so as to correct the bad data.
Reserve battery state update formula for echo state network
Figure 395135DEST_PATH_IMAGE010
Comprises the following steps:
Figure 993607DEST_PATH_IMAGE011
output of the echo state network:
Figure 637340DEST_PATH_IMAGE012
wherein the content of the first and second substances,
Figure 194223DEST_PATH_IMAGE013
the specific value of the regulating factor for the reserve battery depends on the actual situation,
Figure 726836DEST_PATH_IMAGE014
is the function of the activation of the function,
Figure 519211DEST_PATH_IMAGE015
is the data that is input to the computer,
Figure 391352DEST_PATH_IMAGE016
is either the network state or the reserve battery state,
Figure 978191DEST_PATH_IMAGE017
is the input of a matrix of weights,
Figure 201362DEST_PATH_IMAGE018
is the intermediate weight matrix and the intermediate weight matrix,
Figure 203954DEST_PATH_IMAGE019
is the output of the computer system,
Figure 320814DEST_PATH_IMAGE020
is the output weight matrix.
And 4, step 4: and 3, performing quality evaluation on the running state data of the electric vehicle charging facility based on the correction result of the bad data in the step 3, and constructing a charging facility running state data quality evaluation system and a quality evaluation model according to the running state data characteristics.
As shown in fig. 3, the quality evaluation system of the charging facility operating state data in step 4 of the present invention is mainly defined in terms of the inherent quality, accessible quality, context quality, and representation quality of the operating state data. As shown in fig. 4, the charging facility operating state data quality evaluation model includes a decision relationship, a detection relation, a business rule, an algorithm library, a rule weight, and the like. The decision relationship mainly describes the dependency relationship among the charging facility operation state data such as the voltage, the current and the like of each module; the detection relation is a data internal detection relation established according to the dependency relation among the charging facility operation state data; the business rules are a rule set related to the quality detection of the running state data of the charging facility and the problem data processing; the algorithm library is an algorithm set related to the quality control process of the data of the operating state of the charging facility; the rule weight represents the importance of the rule.
And 5: and 4, performing data generalization, data normalization and data attribute construction data conversion on the charging facility running state data quality evaluation system and the quality evaluation model in the step 4.
Step 6: and (5) constructing a charging facility running state data mining model, inputting the output data after data conversion in the step 5 into the charging facility running state data mining model for feature data mining, and simplifying complex data.
Step 6.1, expanding the classification information to the output data after the data conversion in the step 5;
Figure 219500DEST_PATH_IMAGE021
as the number of the categories,
Figure 287557DEST_PATH_IMAGE022
for the output data after data conversionXIf each classification information of (1) is output data
Figure 31522DEST_PATH_IMAGE023
Belong to the class
Figure 268468DEST_PATH_IMAGE024
Then, then
Figure 72476DEST_PATH_IMAGE025
Class label of
Figure 66977DEST_PATH_IMAGE026
Is set to
Figure 739267DEST_PATH_IMAGE027
Otherwise
Figure 440506DEST_PATH_IMAGE028
Is arranged as
Figure 540049DEST_PATH_IMAGE029
(ii) a Determining
Figure 225109DEST_PATH_IMAGE030
And
Figure 373193DEST_PATH_IMAGE029
after that, make
Figure 820617DEST_PATH_IMAGE026
For all
Figure 232007DEST_PATH_IMAGE023
All have normal distributions.
For example,Xare each
Figure 997838DEST_PATH_IMAGE056
The class label corresponding to the category label is 1, 2, 2, and the classification information thereof is as follows:
Figure DEST_PATH_IMAGE057
for having a normal distribution
Figure 480772DEST_PATH_IMAGE058
First element
Figure DEST_PATH_IMAGE059
And
Figure 687762DEST_PATH_IMAGE060
should be 0 and the standard deviation should be 1, on the basis of which it can be determined
Figure DEST_PATH_IMAGE061
Is determined, and then
Figure 597949DEST_PATH_IMAGE062
The value of (c).
At the same time classifying information
Figure DEST_PATH_IMAGE063
Expansion to output data
Figure 836031DEST_PATH_IMAGE025
Need to output data
Figure 653814DEST_PATH_IMAGE025
Performing standardization to output data
Figure 246470DEST_PATH_IMAGE025
Each factor of (a) is normally distributed, and the normalized output data is
Figure 999662DEST_PATH_IMAGE032
Then expand the data
Figure 740085DEST_PATH_IMAGE033
Comprises the following steps:
Figure 971346DEST_PATH_IMAGE034
wherein the normalized output data
Figure 746404DEST_PATH_IMAGE032
Has a dimension of
Figure 670498DEST_PATH_IMAGE035
Number of input and output data;
Figure 835900DEST_PATH_IMAGE031
has the dimension of
Figure 496951DEST_PATH_IMAGE036
Has the dimension of
Figure 1881DEST_PATH_IMAGE037
As data
Figure 955931DEST_PATH_IMAGE025
The number of classification categories of (a);
step 6.2, expanding data in the step 6.1 based on principal component analysis method
Figure 811891DEST_PATH_IMAGE038
Outputting data for feature extractionTaking;
the initial principal component PCs dimension is
Figure 509589DEST_PATH_IMAGE039
If the dimensionality of the input and output data is to be reduced
Figure 134605DEST_PATH_IMAGE040
Until it is less than
Figure 197239DEST_PATH_IMAGE041
Then need to select
Figure 399550DEST_PATH_IMAGE040
A principal component PCs, and obtaining it
Figure 776305DEST_PATH_IMAGE040
Maximum variance, wherein in a certain principal component PCs, the matrix is transformed
Figure 878997DEST_PATH_IMAGE042
Comprises the following steps:
Figure 315795DEST_PATH_IMAGE043
wherein the content of the first and second substances,
Figure 677506DEST_PATH_IMAGE044
is the current principal component
Figure 717006DEST_PATH_IMAGE045
The number of the categories is one,
Figure 51035DEST_PATH_IMAGE044
has a dimension of
Figure 783368DEST_PATH_IMAGE046
In order to transform the matrix, the matrix is,
Figure 835638DEST_PATH_IMAGE042
has the dimension of
Figure 678829DEST_PATH_IMAGE047
To correspond to the amount of the augmented input-output data,
Figure 132944DEST_PATH_IMAGE048
to correspond to the class of the augmented input-output data,
Figure 708282DEST_PATH_IMAGE049
the number of main components of the data after the characteristic extraction is adopted;
step 6.3, transforming the matrix QUOTE
Figure DEST_PATH_IMAGE065
Figure 77208DEST_PATH_IMAGE065
For extended data
Figure 724090DEST_PATH_IMAGE033
Carrying out coordinate transformation on the original data;
by transforming matrices
Figure 32712DEST_PATH_IMAGE042
And extended data
Figure 106847DEST_PATH_IMAGE050
Multiplying to obtain the conversion data on the characteristic space
Figure 664867DEST_PATH_IMAGE051
，
Figure 990807DEST_PATH_IMAGE052
And designing a data mining model through charging facility data mining, analyzing safety fault data characteristics and displaying a data mining result.
And 7: and constructing an integrated fault tree of the charging facility, and analyzing the relation between each layer of fault source and the specific fault type through extracting rules.
7.1, establishing a fault tree, wherein the root of all faults in the running state data of the electric vehicle charging facility collected through the Internet, the battery management system and the operation monitoring platform is used as a statement of a top event;
7.2, finding out bad data in the running state data as an input event, and obtaining converted characteristic data as an output event;
step 7.3, repeating step 7.2 until a bottom event is obtained;
and 7.4, connecting the events of all levels by using a logic symbol to form a fault tree.
It should be emphasized that the embodiments described herein are illustrative rather than restrictive, and thus the present invention is not limited to the embodiments described in the detailed description, but also includes other embodiments that can be derived from the technical solutions of the present invention by those skilled in the art.

Claims (9)

1. A data mining processing method suitable for electric vehicle charging faults is characterized by comprising the following steps: the method comprises the following steps:
step 1: collecting running state data of an electric vehicle charging facility through the Internet, a battery management system and an operation monitoring platform;
step 2: identifying bad data in the operation state data collected in the step 1;
and step 3: correcting the bad data identified in the step 2 by using an echo state network;
and 4, step 4: performing quality evaluation on the running state data of the electric vehicle charging facility based on the correction result of the bad data in the step 3, and meanwhile, constructing a quality evaluation system and a quality evaluation model of the running state data of the charging facility according to the characteristics of the running state data;
and 5: performing data conversion of data generalization, data normalization and data attribute construction on the charging facility running state data quality evaluation system and the quality evaluation model in the step 4;
step 6: constructing a charging facility running state data mining model, inputting the output data after data conversion in the step 5 into the charging facility running state data mining model for feature data mining, and simplifying complex data;
and 7: and 6, constructing an integrated fault tree of the charging facility according to the simplified characteristic data in the step 6, and analyzing the relation between each layer of fault source and the specific fault type through extracting rules.
2. The data mining processing method suitable for the charging fault of the electric vehicle as claimed in claim 1, wherein: the operation state data in the step 1 comprises: voltage, current, and temperature.
3. The data mining processing method suitable for the charging fault of the electric vehicle as claimed in claim 1, wherein: the identification of the bad data in the step 2 comprises the following steps:
step 2.1, clustering the running state data in the step 1, wherein the clustering number is
Figure DEST_PATH_IMAGE001
And is and
Figure 818409DEST_PATH_IMAGE001
=1、2、……
Figure 938812DEST_PATH_IMAGE002
，
Figure 249708DEST_PATH_IMAGE002
calculating cluster dispersion corresponding to different cluster numbers for the total number of clusters
Figure DEST_PATH_IMAGE003
；
Step 2.2, calculating the expectation of clustering dispersion corresponding to different clustering numbers by adopting uniform distribution as reference distributionE(W r,k ) WhereinrFor generating based on characteristics of the raw datarA uniformly distributed reference data set of individuals;
step 2.3, calculating the clustering number
Figure 208699DEST_PATH_IMAGE001
=1、2、……
Figure 85388DEST_PATH_IMAGE002
Intermediate variables of time correspondence
Figure 642271DEST_PATH_IMAGE004
A value;
step 2.4, calculated according to step 2.3
Figure 502780DEST_PATH_IMAGE004
And obtaining the optimal clustering number and identifying bad data.
4. The data mining processing method suitable for the charging fault of the electric vehicle as claimed in claim 3, wherein: intermediate variables in said step 2.3
Figure 904942DEST_PATH_IMAGE004
The calculation method comprises the following steps:
Figure DEST_PATH_IMAGE005
wherein the content of the first and second substances,
Figure 636138DEST_PATH_IMAGE006
the mathematical expectation value in the case where the sample size is n is distributed for the reference value.
5. The data mining processing method suitable for the charging fault of the electric vehicle as claimed in claim 3, wherein: the specific implementation method of the step 2.4 is as follows: for those calculated in step 2.3
Figure 160660DEST_PATH_IMAGE001
An
Figure 741421DEST_PATH_IMAGE004
Make a comparison if
Figure DEST_PATH_IMAGE007
Then, then
Figure 806329DEST_PATH_IMAGE008
Is the best cluster number, otherwise order
Figure 532976DEST_PATH_IMAGE009
And returns to step 2.1.
6. The data mining processing method suitable for the charging fault of the electric vehicle as claimed in claim 1, wherein: the reserve battery state updating formula of the echo state network in the step 3
Figure 290717DEST_PATH_IMAGE010
Comprises the following steps:
Figure 797921DEST_PATH_IMAGE011
output of the echo state network:
Figure 807466DEST_PATH_IMAGE012
wherein the content of the first and second substances,
Figure 778833DEST_PATH_IMAGE013
the specific value of the adjustment factor for the reserve battery depends on the actual situation,
Figure 848420DEST_PATH_IMAGE014
is the function of the activation of the function,
Figure 141123DEST_PATH_IMAGE015
is the data that is input to the computer,
Figure 954358DEST_PATH_IMAGE016
is either the network state or the reserve battery state,
Figure 717915DEST_PATH_IMAGE017
is the input of a matrix of weights,
Figure 20720DEST_PATH_IMAGE018
is the intermediate weight matrix and the intermediate weight matrix,
Figure 502517DEST_PATH_IMAGE019
is the output of the computer system,
Figure 978498DEST_PATH_IMAGE020
is the output weight matrix.
7. The data mining processing method suitable for the charging fault of the electric vehicle as claimed in claim 1, wherein: the quality evaluation system of the charging facility running state data in the step 4 comprises the inherent quality, accessible quality, context quality and representation quality of the running state data; the charging facility operation state data quality evaluation model comprises a decision relation, a detection relation, a business rule, an algorithm library and a rule weight.
8. The data mining processing method suitable for the charging fault of the electric vehicle as claimed in claim 1, wherein: the feature data mining in the step 6 comprises the following steps:
step 6.1, expanding the classification information to the output data after the data conversion in the step 5;
Figure 799823DEST_PATH_IMAGE021
as the number of the categories,
Figure 804689DEST_PATH_IMAGE022
for the output data after data conversionXIf each classification information of (1) is output data
Figure 711465DEST_PATH_IMAGE023
Belong to the class
Figure 495531DEST_PATH_IMAGE024
Then, then
Figure 171363DEST_PATH_IMAGE025
Class label of
Figure 878288DEST_PATH_IMAGE026
Is set to
Figure 69098DEST_PATH_IMAGE027
Otherwise
Figure 762247DEST_PATH_IMAGE028
Is arranged as
Figure 682799DEST_PATH_IMAGE029
(ii) a Determining
Figure 170412DEST_PATH_IMAGE030
And
Figure 910835DEST_PATH_IMAGE029
then make it
Figure 407675DEST_PATH_IMAGE026
For all
Figure 684198DEST_PATH_IMAGE023
All have normal distributions; at the same time classifying information
Figure 608292DEST_PATH_IMAGE031
Expansion to output data
Figure 773694DEST_PATH_IMAGE025
Need to output data
Figure 933280DEST_PATH_IMAGE025
Performing standardization to output dataXEach factor of (a) is normally distributed, and the normalized output data is
Figure 438210DEST_PATH_IMAGE032
Then expand the data
Figure 657839DEST_PATH_IMAGE033
Comprises the following steps:
Figure 248220DEST_PATH_IMAGE034
wherein the normalized output data
Figure 211497DEST_PATH_IMAGE032
Has a dimension of
Figure 836513DEST_PATH_IMAGE035
The number of input and output data;
Figure 459999DEST_PATH_IMAGE031
has the dimension of
Figure 537677DEST_PATH_IMAGE036
Has the dimension of
Figure 39065DEST_PATH_IMAGE037
As data
Figure 518588DEST_PATH_IMAGE025
The number of classification categories of (a);
step 6.2, expanding data in the step 6.1 based on a principal component analysis method
Figure 752123DEST_PATH_IMAGE038
Carrying out feature extraction;
the initial principal component PCs dimension is
Figure 441731DEST_PATH_IMAGE039
If the dimensionality of the input and output data is to be reduced
Figure 622176DEST_PATH_IMAGE040
Until it is less than
Figure 346419DEST_PATH_IMAGE041
Then need to select
Figure 954118DEST_PATH_IMAGE040
A principal component PCs, and obtaining it
Figure 101327DEST_PATH_IMAGE040
Maximum variance, wherein in a certain principal component PCs, a transformation matrix
Figure 85464DEST_PATH_IMAGE042
Comprises the following steps:
Figure 664213DEST_PATH_IMAGE043
wherein the content of the first and second substances,
Figure 442813DEST_PATH_IMAGE044
is the current principal component
Figure 841433DEST_PATH_IMAGE045
Class IIn addition, the method comprises the following steps of,
Figure 363682DEST_PATH_IMAGE044
has the dimension of
Figure 469041DEST_PATH_IMAGE046
In order to transform the matrix, the matrix is,
Figure 543176DEST_PATH_IMAGE042
has the dimension of
Figure 570038DEST_PATH_IMAGE047
To correspond to the amount of the augmented input-output data,
Figure 247707DEST_PATH_IMAGE048
to correspond to the class of the augmented input-output data,
Figure 410835DEST_PATH_IMAGE049
the number of main components of the data after the characteristic extraction;
step 6.3, transforming the matrix
Figure 655872DEST_PATH_IMAGE042
For extended data
Figure 904451DEST_PATH_IMAGE033
Carrying out coordinate transformation;
by transforming matrices
Figure 565239DEST_PATH_IMAGE042
And extended data
Figure 973087DEST_PATH_IMAGE050
Multiplying to obtain the conversion data on the characteristic space
Figure 264391DEST_PATH_IMAGE051
，
Figure 124899DEST_PATH_IMAGE052
And designing a data mining model through charging facility data mining, analyzing safety fault data characteristics and displaying a data mining result.
9. The data mining processing method suitable for the charging fault of the electric vehicle as claimed in claim 1, wherein: the step 7 of establishing the fault tree comprises the following steps:
7.1, establishing a fault tree, wherein the root of all faults in the running state data of the electric vehicle charging facility collected through the Internet, the battery management system and the operation monitoring platform is used as a statement of a top event;
7.2, finding out bad data in the running state data as an input event, and obtaining converted characteristic data as an output event;
step 7.3, repeating step 7.2 until a bottom event is obtained;
and 7.4, connecting the events of all levels by using a logic symbol to form a fault tree.
CN202210392883.XA 2022-04-15 2022-04-15 Data mining processing method suitable for electric vehicle charging fault Active CN114490836B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210392883.XA CN114490836B (en) 2022-04-15 2022-04-15 Data mining processing method suitable for electric vehicle charging fault

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210392883.XA CN114490836B (en) 2022-04-15 2022-04-15 Data mining processing method suitable for electric vehicle charging fault

Publications (2)

Publication Number Publication Date
CN114490836A true CN114490836A (en) 2022-05-13
CN114490836B CN114490836B (en) 2022-08-02

Family

ID=81487646

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210392883.XA Active CN114490836B (en) 2022-04-15 2022-04-15 Data mining processing method suitable for electric vehicle charging fault

Country Status (1)

Country Link
CN (1) CN114490836B (en)

Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103077325A (en) * 2013-01-30 2013-05-01 西安交通大学 Bad data detection method of smart power grid based on self-adaptive partition state estimation
CN108229661A (en) * 2018-01-03 2018-06-29 华南理工大学 A kind of depth echo state network model building method based on multiple coding re-projection
CN108876057A (en) * 2018-07-24 2018-11-23 合肥阳光新能源科技有限公司 A kind of failure prediction method of micro-capacitance sensor, device and electronic equipment
CN110263846A (en) * 2019-06-18 2019-09-20 华北电力大学 The method for diagnosing faults for being excavated and being learnt based on fault data depth
CN110298369A (en) * 2018-03-21 2019-10-01 中国电力科学研究院有限公司 A kind of discrimination method and system of electric system bad data
CN110544047A (en) * 2019-09-10 2019-12-06 东北电力大学 Bad data identification method
CN111007401A (en) * 2019-12-16 2020-04-14 国网江苏省电力有限公司电力科学研究院 Electric vehicle battery fault diagnosis method and device based on artificial intelligence
CN111038291A (en) * 2019-12-16 2020-04-21 国网江苏省电力有限公司电力科学研究院 Intelligent diagnosis system and method for charging fault of electric automobile
CN111707943A (en) * 2020-05-15 2020-09-25 中国电力科学研究院有限公司 Battery simulation-based electric vehicle charging fault early warning method and system
CN112731159A (en) * 2020-12-23 2021-04-30 江苏省电力试验研究院有限公司 Method for pre-judging and positioning battery fault of battery compartment of energy storage power station
CN113592256A (en) * 2021-07-13 2021-11-02 北京全路通信信号研究设计院集团有限公司 Train operation plan adjusting method and system based on charging station faults
CN113705615A (en) * 2021-07-29 2021-11-26 南京邮电大学 Neural network-based electric vehicle charging process multistage equipment fault diagnosis method and system
CN113887606A (en) * 2021-09-28 2022-01-04 上海工业自动化仪表研究院有限公司 Electronic equipment control system fault diagnosis method based on fault tree establishment

Patent Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103077325A (en) * 2013-01-30 2013-05-01 西安交通大学 Bad data detection method of smart power grid based on self-adaptive partition state estimation
CN108229661A (en) * 2018-01-03 2018-06-29 华南理工大学 A kind of depth echo state network model building method based on multiple coding re-projection
CN110298369A (en) * 2018-03-21 2019-10-01 中国电力科学研究院有限公司 A kind of discrimination method and system of electric system bad data
CN108876057A (en) * 2018-07-24 2018-11-23 合肥阳光新能源科技有限公司 A kind of failure prediction method of micro-capacitance sensor, device and electronic equipment
CN110263846A (en) * 2019-06-18 2019-09-20 华北电力大学 The method for diagnosing faults for being excavated and being learnt based on fault data depth
CN110544047A (en) * 2019-09-10 2019-12-06 东北电力大学 Bad data identification method
CN111007401A (en) * 2019-12-16 2020-04-14 国网江苏省电力有限公司电力科学研究院 Electric vehicle battery fault diagnosis method and device based on artificial intelligence
CN111038291A (en) * 2019-12-16 2020-04-21 国网江苏省电力有限公司电力科学研究院 Intelligent diagnosis system and method for charging fault of electric automobile
CN111707943A (en) * 2020-05-15 2020-09-25 中国电力科学研究院有限公司 Battery simulation-based electric vehicle charging fault early warning method and system
CN112731159A (en) * 2020-12-23 2021-04-30 江苏省电力试验研究院有限公司 Method for pre-judging and positioning battery fault of battery compartment of energy storage power station
CN113592256A (en) * 2021-07-13 2021-11-02 北京全路通信信号研究设计院集团有限公司 Train operation plan adjusting method and system based on charging station faults
CN113705615A (en) * 2021-07-29 2021-11-26 南京邮电大学 Neural network-based electric vehicle charging process multistage equipment fault diagnosis method and system
CN113887606A (en) * 2021-09-28 2022-01-04 上海工业自动化仪表研究院有限公司 Electronic equipment control system fault diagnosis method based on fault tree establishment

Also Published As

Publication number Publication date
CN114490836B (en) 2022-08-02

Similar Documents

Publication Publication Date Title
CN107301296B (en) Data-based qualitative analysis method for circuit breaker fault influence factors
CN111478314B (en) Transient stability evaluation method for power system
CN108345670B (en) Service hotspot discovery method for 95598 power work order
CN112906764B (en) Communication safety equipment intelligent diagnosis method and system based on improved BP neural network
CN111667135B (en) Load structure analysis method based on typical feature extraction
CN113112090B (en) Space load prediction method based on principal component analysis of comprehensive mutual informativity
CN114021483A (en) Ultra-short-term wind power prediction method based on time domain characteristics and XGboost
CN111815026A (en) Multi-energy system load prediction method based on feature clustering
CN115526258A (en) Power system transient stability evaluation method based on Spearman correlation coefficient feature extraction
CN111832839A (en) Energy consumption prediction method based on sufficient incremental learning
CN111275485A (en) Power grid customer grade division method and system based on big data analysis, computer equipment and storage medium
CN114490836B (en) Data mining processing method suitable for electric vehicle charging fault
CN111585277B (en) Power system dynamic security assessment method based on hybrid integration model
CN112748331A (en) Circuit breaker mechanical fault identification method and device based on DS evidence fusion
CN116151799A (en) BP neural network-based distribution line multi-working-condition fault rate rapid assessment method
CN116400266A (en) Transformer fault detection method, device and medium based on digital twin model
CN116070458A (en) New wind power plant scene generation method based on RAC-GAN
CN115170304A (en) Method and device for extracting risk feature description
Li et al. Multi-source heterogeneous log fusion technology of power information system based on big data and imprecise reasoning theory
CN114372835A (en) Comprehensive energy service potential customer identification method, system and computer equipment
Xiao et al. Multi-variety and small-batch production quality forecasting by novel data-driven grey Weibull model
CN114826632A (en) Network attack classification method based on network security data cleaning fusion
Zheng Design and verification of use case generation algorithm based on multiple combination tests
CN113779109B (en) Electric power data preprocessing method based on context environment
Fu Design of Marketing Analysis System Based on Data Mining Technology

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
CP02 Change in the address of a patent holder
CP02 Change in the address of a patent holder

Address after: No. 8, Haitai Huake 4th Road, Huayuan Industrial Zone, High tech Zone, Binhai New Area, Tianjin, 300384

Patentee after: ELECTRIC POWER SCIENCE & RESEARCH INSTITUTE OF STATE GRID TIANJIN ELECTRIC POWER Co.

Patentee after: STATE GRID TIANJIN ELECTRIC POWER Co.

Patentee after: NARI TECHNOLOGY Co.,Ltd.

Patentee after: Tianjin University

Patentee after: STATE GRID INFORMATION & TELECOMMUNICATION GROUP Co.,Ltd.

Patentee after: STATE GRID CORPORATION OF CHINA

Address before: No.8, Haitai Huake 4th Road, Xiqing District, Tianjin 300384

Patentee before: ELECTRIC POWER SCIENCE & RESEARCH INSTITUTE OF STATE GRID TIANJIN ELECTRIC POWER Co.

Patentee before: STATE GRID TIANJIN ELECTRIC POWER Co.

Patentee before: NARI TECHNOLOGY Co.,Ltd.

Patentee before: Tianjin University

Patentee before: STATE GRID INFORMATION & TELECOMMUNICATION GROUP Co.,Ltd.

Patentee before: STATE GRID CORPORATION OF CHINA