CN112183999A - Power distribution main equipment sensor reliability evaluation index feature extraction method - Google Patents

Power distribution main equipment sensor reliability evaluation index feature extraction method Download PDF

Info

Publication number
CN112183999A
CN112183999A CN202011015626.1A CN202011015626A CN112183999A CN 112183999 A CN112183999 A CN 112183999A CN 202011015626 A CN202011015626 A CN 202011015626A CN 112183999 A CN112183999 A CN 112183999A
Authority
CN
China
Prior art keywords
index
indexes
main equipment
weight
power distribution
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202011015626.1A
Other languages
Chinese (zh)
Inventor
张世栋
房牧
李帅
张鹏平
王峰
刘洋
刘合金
黄敏
苏国强
孙勇
张林利
由新红
李立生
邵志敏
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.)
State Grid Corp of China SGCC
Electric Power Research Institute of State Grid Shandong Electric Power Co Ltd
Original Assignee
State Grid Corp of China SGCC
Electric Power Research Institute of State Grid Shandong Electric Power 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 State Grid Corp of China SGCC, Electric Power Research Institute of State Grid Shandong Electric Power Co Ltd filed Critical State Grid Corp of China SGCC
Priority to CN202011015626.1A priority Critical patent/CN112183999A/en
Publication of CN112183999A publication Critical patent/CN112183999A/en
Pending legal-status Critical Current

Links

Images

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
    • G06Q10/06—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
    • G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
    • 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
    • G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
    • G06F18/24133—Distances to prototypes
    • G06F18/24143—Distances to neighbourhood prototypes, e.g. restricted Coulomb energy networks [RCEN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2113/00—Details relating to the application field
    • G06F2113/04—Power grid distribution networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/02—Reliability analysis or reliability optimisation; Failure analysis, e.g. worst case scenario performance, failure mode and effects analysis [FMEA]
    • 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
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E40/00—Technologies for an efficient electrical power generation, transmission or distribution
    • Y02E40/70—Smart grids as climate change mitigation technology in the energy generation sector
    • 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
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/80—Management or planning
    • Y02P90/82—Energy audits or management systems therefor
    • 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
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Economics (AREA)
  • Educational Administration (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Strategic Management (AREA)
  • Data Mining & Analysis (AREA)
  • Development Economics (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Game Theory and Decision Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Computer Hardware Design (AREA)
  • Geometry (AREA)
  • Complex Calculations (AREA)

Abstract

The invention provides a power distribution main equipment sensor reliability evaluation index feature extraction method, which is used for extracting features of a large number of power distribution Internet of things main equipment sensor evaluation indexes, and mapping high-dimensional indexes by using extracted low-dimensional indexes: firstly, a hierarchical structure model is constructed by collecting various factors related to the reliability of a main equipment sensor device, the weight of each index is quantified by adopting a G1-CV method, and then the characteristic extraction of the index is carried out by utilizing an ITriMap algorithm to obtain an index system after reduction. The method adopts the Mahalanobis distance to measure the distance between samples so as to improve the local precision of the TriMap, the new algorithm model is named as the itriMap, the itriMap algorithm can improve the local precision of the original algorithm on the basis of keeping the global precision of the original algorithm, the algorithm has smaller reconstruction error compared with the traditional algorithms such as t-SNE, Largr-Visas, UMAP and the like, the indexes after characteristic extraction can scientifically and objectively evaluate the reliability of the power distribution network main equipment sensor device, and the method has higher practical value in engineering.

Description

Power distribution main equipment sensor reliability evaluation index feature extraction method
Technical Field
The invention belongs to the technical field of high-reliability self-sensing of a power distribution Internet of things, and particularly relates to a power distribution main equipment sensor reliability evaluation index feature extraction method.
Background
Along with the large-scale information construction of power enterprises in recent years, the mode of the traditional power distribution network faces huge challenges: firstly, the variety and the number of the social service demands of the user side are increased rapidly, and the role of the power distribution network is required to be changed from the traditional one-way power provider to the two-way energy flow and the advanced service; secondly, the cluster effect of high-proportion distributed renewable energy sources is obvious, and heterogeneous energy source system fusion and the like are needed for the power distribution network. Therefore, the concept of the power distribution internet of things is developed. The power distribution internet of things is used as an important component of a ubiquitous power internet of things, the deep integration of the internet of things and a power distribution network is realized, along with the stable development of the construction of the power distribution internet of things, the functions of the power distribution network needing to be carried are changed, and as a complex large system with the deep integration of the internet of things technology, the big data technology and the optimized operation regulation and control technology, the power distribution internet of things presents the characteristic of interactive coupling of energy flow, information flow and service flow.
Distribution main equipment such as a transformer, a circuit breaker, a fuse, an isolating switch, a load switch and a voltage transformer are arranged in the distribution internet of things, and various distribution main equipment are connected with sensor devices such as a temperature and humidity sensor, a pressure sensor, a vibration sensor, a noise sensor, an infrared sensor and a harmonic sensor. In order to stably run the whole distribution internet of things, reliability research is carried out on the distribution main equipment sensor device from multiple aspects, and a main equipment state sensor reliability evaluation system is constructed and is worthy of research.
At present, most of the existing work is to establish a reliability evaluation system for a wireless sensor network, for example, the reliability evaluation system is established by using methods such as a multi-factor analysis process and a neural network. Few learners evaluate the reliability of the sensor device, and the sensor device also relates to a plurality of indexes influencing the reliability of the sensor device and can influence the safe operation of the whole power distribution internet of things.
Disclosure of Invention
The invention discloses an evaluation index feature extraction method, aiming at solving the defects of the reliability evaluation of a power distribution main equipment sensor device, the defects that the traditional evaluation system is too high in calculation complexity, too strong in subjectivity, and the feature extraction method cannot guarantee the local precision of an algorithm, and the like.
A method for extracting reliability evaluation index features of a power distribution main equipment sensor at least comprises the following steps:
1) constructing a comprehensive evaluation index system with 4 criterion layers of technical evaluation indexes, device energy efficiency evaluation indexes, safety evaluation indexes and device operation condition evaluation indexes;
2) determining an initial weight of each index by expert scoring by using a G1 method;
3) aiming at the condition that each index unit and magnitude of the index are different, a Min-max standardization processing method is adopted to carry out standardization processing on the index;
4) determining the final weight of each index by using a variation coefficient method;
5) extracting characteristics of the indexes, and reducing high-level indexes to obtain a power distribution Internet of things main equipment sensor reliability evaluation index system subjected to characteristic extraction;
wherein, step 5) also includes the step:
inputting: high dimensional sample data X ═ X1,x2,＝xn]∈RDWherein D is decision space, the number of adjacent points is k, and the dimension of the low-dimensional space is D, wherein D is more than D;
and (3) outputting: low dimensional sample data Y ═ Y1,y2,…yn]∈RD；
(i) Mahalanobis distance metric learning process: firstly, defining a homogeneous set Qw＝{(xi,xj|xiAnd xjIs homogeneous), then a non-homogeneous set Q is definedb＝{(xi,xj|xiAnd xjIs heterogeneous) }, with the goal that A is at QwThe distance between the middle point pair is as small as possible, and is in QbThe distance between the middle point pairs is required to be as large as possible, and the data sample set is set to be omega ═ x1,x2,…xNIn which xi∈Rn1,2, …, N, using
Figure BDA0002698957990000021
Solved optimization matrix W*Then, the Marek's metric matrix A ═ W is obtained*(W*)TFinding a sample point x using the mahalanobis distanceiK neighbors ofPoint xi1,xi2,…,xikWherein W is an integrated weight matrix of each index, S is an alternative scheme set, SWRepresents QwCovariance matrix of all point pairs in, SbRepresents QbCovariance matrices of all the point pairs in the list;
(ii) by using
Figure BDA0002698957990000022
Calculate each sample point xiWeight w of k neighborsijkWhere is a small constant;
(iii) and (3) carrying out dimensionality reduction on the training sample: linearly approximating itself with k neighbors of each sample such that the approximation error is minimal under the mahalanobis metric;
the cost function is set as follows:
Figure BDA0002698957990000031
make it
Figure BDA0002698957990000032
And (3) solving an optimal coefficient by using a maximum likelihood method:
Figure BDA0002698957990000033
wherein the content of the first and second substances,
Figure BDA0002698957990000034
ensure wijkReconstructing sample data in a low-dimensional space under the premise of no change, wherein a cost function is as follows:
Figure BDA0002698957990000035
is provided with
Figure BDA0002698957990000036
M＝(I-W)TA (I-W), the matrix M is arranged in ascending order, the smallest d2A non-zero feature vector of
Figure BDA0002698957990000037
Then finally obtain
Figure BDA0002698957990000038
(iv) And (3) carrying out dimension reduction on the test sample: (iv) similar to step (iii), for the sample x to be measuredtestFinding k adjacent points x by using Mahalanobis distancetest1,xtest2,…,xtestkWhich corresponds to y in the lower dimensional spacetest1,ytest2,…,ytestkThen calculating the reconstruction coefficient
Figure BDA0002698957990000039
The new sample is then represented in the low-dimensional space as:
Figure BDA00026989579900000310
(v) finding distance y in low-dimensional space by using nearest neighbor algorithmtestThe most recent data sample has a category of ytestAnd (4) belonging classification, finishing the feature extraction and finishing the algorithm.
Preferably, step 2) further comprises at least:
(i) selecting an authoritative expert and determining an order relation;
(ii) determining the relative importance degree between adjacent indexes: if the cumulative importance in the same stage exceeds 1.8, i.e., is extremely important, it is necessary to be in the original RjkMultiplying the value by a scaling factor p to ensure
Figure BDA0002698957990000041
R′jk＝RjkP, in the formula, RjkThe values of the weight ratio between the adjacent indexes are judged for the alpha-th expert from 1.0, 1.2, 1.4, 1.6 and 1.8, which respectively represent the same importance, slightly importance, obvious importance, strong importance and extreme importance.
(iii) And determining subjective weight in a layering manner, wherein in the evaluation of the ith expert, the index subjective weight with lower relative importance degree is as follows:
Figure BDA0002698957990000042
by
Figure BDA0002698957990000043
The subjective weight of the evaluation of other index experts in the criterion layer is obtained in a recursion way
Figure BDA0002698957990000044
The subjective weight of each index of each level can be obtained by repeating the steps, so that the expert evaluation subjective weight of each index is as follows:
Figure BDA0002698957990000045
then, the subjective weight of each index is solved by adopting a weighted arithmetic mean method
Figure BDA0002698957990000046
Figure BDA0002698957990000047
Wherein l represents the maximum value of i;
in summary, the index set X ═ X1,x2,…,xnThe subjective weight coefficient set of each index
Figure BDA0002698957990000048
Preferably, step 4) further comprises at least:
(i) the coefficient of variation of each index is calculated by the following formula:
Figure BDA0002698957990000049
wherein σiIndicates the standard deviation of the i-th index,
Figure BDA00026989579900000410
represents the average of the i-th index,
objective weight of each index
Figure BDA00026989579900000411
Comprises the following steps:
(ii) the integrated weight ωt(t ═ 1,2, …, n) both subjective and objective weights are considered, defined as:
Figure BDA0002698957990000051
Figure BDA0002698957990000052
refers to the subjective weight of each index,
Figure BDA0002698957990000053
the objective weight of each index is indicated;
in summary, for index set X ═ X1,x2,…,xnThe comprehensive weight coefficient set is V ═ omega1,ω2,…,ωnCorresponding to the comprehensive weight matrix D ═ omega0 ω1 … ωn]T；
(iii) And obtaining a normalized weighting decision matrix Z.
Preferably, the constant of (ii) in step 5) is set to 10-4。
Preferably, the technical evaluation indexes in step 1 include the following 16 indexes:
the method comprises the steps of node redundancy, channel redundancy, sampling frequency, source node link reliability, relay node link reliability, signal transmission performance, power supply stability, product working time, data testing time, average voltage deviation rate, flow benefit, fault statistical range, component response time, quality quantification, equipment duty ratio and maximum retransmission times.
Preferably, the device energy efficiency assessment index in step 1 includes the following 17 indexes:
the method comprises the steps of data packet sending energy consumption, data packet receiving energy consumption, power factor, load factor, winding temperature rise, lead temperature rise, no-load excitation current percentage, energy-saving hardware proportion, total harmonic distortion rate, three-phase load unbalance rate, rated no-load loss, rated load loss, link energy availability, wiring length exceeding rate, wiring sectional area exceeding rate, configuration and change and medium management.
Preferably, the safety assessment indicators in step 1 include the following 16 indicators:
drift deviation fault, accuracy drop fault, fixed deviation fault, complete failure fault, local information security, patch security, information leakage probability, influence of denial of service, network attack frequency, communication network interference rate, channel packet loss rate, security mechanism perfection, signal transmission interruption probability, data transmission security, level authority completeness, and encrypted transmission.
Preferably, the device operation condition evaluation indexes in step 1 include the following 17 indexes:
the method comprises the steps of sensor precision, sensor installation position, protocol distance applicability, end-to-end time delay, source node link operation condition, relay node link operation condition, real-time received data volume, service success rate, node energy availability, mean time to failure, task profile period, node connectivity probability, node capacity, link path flow, task transmission stability and routing signaling overhead.
The invention provides a power distribution main equipment sensor reliability evaluation index feature extraction method, which is used for extracting features of a large number of power distribution Internet of things main equipment sensor evaluation indexes, and mapping high-dimensional indexes by using extracted low-dimensional indexes: firstly, a hierarchical structure model is constructed by collecting various factors related to the reliability of a main equipment sensor device, the weight of each index is quantified by adopting a G1-CV method, and then the characteristic extraction of the index is carried out by utilizing an ITriMap algorithm to obtain an index system after reduction.
The method adopts the Mahalanobis distance to measure the distance between samples so as to improve the local precision of the TriMap, the new algorithm model is named as the itriMap, the itriMap algorithm can improve the local precision of the original algorithm on the basis of keeping the global precision of the original algorithm, the algorithm has smaller reconstruction error compared with the traditional algorithms such as t-SNE, Largr-Visas, UMAP and the like, the indexes after characteristic extraction can scientifically and objectively evaluate the reliability of the power distribution network main equipment sensor device, and the method has higher practical value in engineering.
Drawings
FIG. 1 is a schematic composition diagram of the comprehensive evaluation index system in example 1;
FIG. 2 is a distribution diagram of indexes in the index criterion layer of the technical evaluation in example 1;
fig. 3 is an index distribution diagram of an index criterion layer of the energy efficiency evaluation of the apparatus in embodiment 1;
FIG. 4 is an index distribution diagram of a safety evaluation index criterion layer in embodiment 1;
FIG. 5 is an index distribution diagram of an index criterion layer for evaluating the operation condition of the apparatus in embodiment 1;
FIG. 6 is a flow chart of feature extraction for the index system based on the G1-CV method and the itriMap algorithm in example 1.
Detailed Description
The detailed description of index feature extraction and establishment of an evaluation index system for the reliability of the main equipment sensor of the power distribution internet of things is provided with reference to the accompanying drawings 1 to 6 and specific embodiments. It must be emphasized that the following description is merely exemplary in nature and is not intended to limit the scope or application of the present invention.
As shown in fig. 1 to 5, the power distribution main equipment sensor reliability evaluation index hierarchical structure model constructed in this embodiment includes four criteria layer primary indexes, such as a technical evaluation index, an apparatus energy efficiency evaluation index, a security evaluation index, and an apparatus operation condition evaluation, and totally 66 secondary indexes.
As shown in the attached figure 6, the specific process of the invention is as follows: firstly, a four-layer 66-dimensional reliability evaluation index hierarchical structure model is constructed by referring to a large amount of literature data, then weight assignment is carried out on each index by using a G1-CV method, then standardization processing is carried out on the index, and finally, the provided ITRIMap algorithm is utilized to carry out feature extraction on an index system, so that the effect that a high-dimensional space can be mapped by using a low-dimensional space is achieved.
The ITRIMap algorithm-based power distribution main equipment sensor reliability evaluation index feature extraction method comprises the following specific steps:
inputting: high dimensional sample data X ═ X1,x2,…xn]∈RDThe number of adjacent points is k, the low-dimensional space dimension is D, wherein D is more than D;
and (3) outputting: low dimensional sample data Y ═ Y1,y2,…yn]∈RD；
(i) Mahalanobis distance metric learning process: let the set of data samples be Ω ═ x1,x2,…xNIn which xi∈Rn1,2, …, N, using
Figure BDA0002698957990000071
Solved optimization matrix W*Then, the Marek's metric matrix A ═ W is obtained*(W*)TFinding a sample point x using the mahalanobis distanceiK number of neighboring points xi1,xi2,…,xik；
(ii) By using
Figure BDA0002698957990000072
Calculate each sample point xiWeight w of k neighborsijk；
(iii) And (3) carrying out dimensionality reduction on the training sample: linearly approximating itself with k neighbors of each sample such that the approximation error is minimal under the mahalanobis metric;
the cost function is set as follows:
Figure BDA0002698957990000073
make it
Figure BDA0002698957990000074
And (3) solving an optimal coefficient by using a maximum likelihood method:
Figure BDA0002698957990000075
wherein the content of the first and second substances,
Figure BDA0002698957990000076
ensure wijkReconstructing sample data in a low-dimensional space under the premise of no change, wherein a cost function is as follows:
Figure BDA0002698957990000081
is provided with
Figure BDA0002698957990000082
M＝(I-W)TA (I-W), the matrix M is arranged in ascending order, the smallest d2A non-zero feature vector of
Figure BDA0002698957990000083
Then finally obtain
Figure BDA0002698957990000084
(iv) And (3) carrying out dimension reduction on the test sample: (iv) similar to step (iii), for the sample x to be measuredtestFinding k adjacent points x by using Mahalanobis distancetest1,xtest2,…,xtestkWhich corresponds to y in the lower dimensional spacetest1,ytest2,＝,ytestkThen calculating the reconstruction coefficient
Figure BDA0002698957990000085
The new sample is then represented in the low-dimensional space as:
Figure BDA0002698957990000086
(v) finding distance y in low-dimensional space by using nearest neighbor algorithmtestThe most recent data sample has a category of ytestAnd (4) belonging classification, finishing the feature extraction and finishing the algorithm.
The step of carrying out weight assignment on each index by the G1 method comprises the following steps:
(i) selecting an authoritative expert and determining an order relation;
(ii) determining the relative importance degree between adjacent indexes: if the cumulative importance in the same stage exceeds 1.8, i.e., is extremely important, it is necessary to be in the original RjkMultiplying the value by a scaling factor p to ensure
Figure BDA0002698957990000087
R′jk＝Rjk·p
(iii) And determining subjective weight in a layering manner, wherein in the evaluation of the ith expert, the index subjective weight with lower relative importance degree is as follows:
Figure BDA0002698957990000088
by
Figure BDA0002698957990000089
The subjective weight of the evaluation of other index experts in the criterion layer is obtained in a recursion way
Figure BDA0002698957990000091
The subjective weight of each index of each level can be obtained by repeating the steps, so that the expert evaluation subjective weight of each index is as follows:
Figure BDA0002698957990000092
then, the subjective weight of each index is solved by adopting a weighted arithmetic mean method
Figure BDA0002698957990000099
Figure BDA0002698957990000093
In summary, the index set X ═ X1,x2,…,xnThe subjective weight coefficient set of each index
Figure BDA0002698957990000094
The final determination of the objective weight of each index by the Coefficient of Variation (CV) method further comprises the steps of:
(i) the coefficient of variation of each index is calculated by the following formula:
Figure BDA0002698957990000095
wherein σiIndicates the standard deviation of the i-th index,
Figure BDA0002698957990000096
represents the average of the i-th index,
objective weight of each index
Figure BDA0002698957990000097
Comprises the following steps:
(ii) the integrated weight ωt(t ═ 1,2, …, n) both subjective and objective weights are considered, defined as:
Figure BDA0002698957990000098
in summary, for index set X ═ X1,x2,…,xnThe comprehensive weight coefficient set is V ═ omega1,ω2,…,ωnCorresponding to the comprehensive weight matrix D ═ omega0 ω1 … ωn]T；
(iii) And obtaining a normalized weighting decision matrix Z.
The above-mentioned method is only a preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention should be covered by the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims (8)

1. A method for extracting reliability evaluation index features of a power distribution main equipment sensor at least comprises the following steps:
1) constructing a comprehensive evaluation index system with 4 criterion layers of technical evaluation indexes, device energy efficiency evaluation indexes, safety evaluation indexes and device operation condition evaluation indexes;
2) determining an initial weight of each index by expert scoring by using a G1 method;
3) aiming at the condition that each index unit and magnitude of the index are different, a Min-max standardization processing method is adopted to carry out standardization processing on the index;
4) determining the final weight of each index by using a variation coefficient method;
5) extracting characteristics of the indexes, and reducing high-level indexes to obtain a power distribution Internet of things main equipment sensor reliability evaluation index system subjected to characteristic extraction;
wherein, step 5) also includes the step:
inputting: high dimensional sample data X ═ X1,x2,…xn]∈RDWherein D is decision space, the number of adjacent points is k, and the dimension of the low-dimensional space is D, wherein D is more than D;
and (3) outputting: low dimensional sample data Y ═ Y1,y2,…yn]∈RD；
(i) Mahalanobis distance metric learning process: firstly, defining a homogeneous set Qw＝{(xi,xj|xiAnd xjIs homogeneous), then a non-homogeneous set Q is definedb＝{(xi,xj|xiAnd xjIs heterogeneous) }, with the goal that A is at QwThe distance between the middle point pair is as small as possible, and is in QbThe distance between the middle point pairs is required to be as large as possible, and the data sample set is set to be omega ═ x1,x2,…xNIn which xi∈Rn1,2, …, N, using
Figure FDA0002698957980000011
Solved optimization matrix W*Then, the Marek's metric matrix A ═ W is obtained*(W*)TFinding a sample point x using the mahalanobis distanceiK number of neighboring points xi1,xi2,＝,xikWherein W is an integrated weight matrix of each index, S is an alternative scheme set, SWRepresents QwCovariance matrix of all point pairs in, SbRepresents QbCovariance matrices of all the point pairs in the list;
(ii) by using
Figure FDA0002698957980000012
Calculate each sample point xiWeight w of k neighborsijkWhere is a small constant;
(iii) and (3) carrying out dimensionality reduction on the training sample: linearly approximating itself with k neighbors of each sample such that the approximation error is minimal under the mahalanobis metric;
the cost function is set as follows:
Figure FDA0002698957980000021
make it
Figure FDA0002698957980000022
And (3) solving an optimal coefficient by using a maximum likelihood method:
Figure FDA0002698957980000023
wherein the content of the first and second substances,
Figure FDA0002698957980000024
ensure wijkReconstructing sample data in a low-dimensional space under the premise of no change, wherein a cost function is as follows:
Figure FDA0002698957980000025
is provided with
Figure FDA0002698957980000026
M＝(I-W)TA (I-W), the matrix M is arranged in ascending order, the smallest d2A non-zero feature vector of
Figure FDA0002698957980000027
Then finally obtain
Figure FDA0002698957980000028
(iv) And (3) carrying out dimension reduction on the test sample: (iv) similar to step (iii), for the sample x to be measuredtestFinding k adjacent points x by using Mahalanobis distancetest1,xtest2,…,xtestkWhich corresponds to y in the lower dimensional spacetest1,ytest2,…,ytestkThen calculating the reconstruction coefficient
Figure FDA0002698957980000029
The new sample is then represented in the low-dimensional space as:
Figure FDA00026989579800000210
(v) finding distance y in low-dimensional space by using nearest neighbor algorithmtestThe most recent data sample has a category of ytestAnd (4) belonging classification, finishing the feature extraction and finishing the algorithm.
2. The method for extracting the reliability evaluation index feature of the power distribution main equipment sensor according to claim 1, wherein the step 2) further comprises at least:
(i) selecting an authoritative expert and determining an order relation;
(ii) determining the relative importance degree between adjacent indexes: if the cumulative importance in the same stage exceeds 1.8, i.e., is extremely important, it is necessary to be in the original RjkMultiplying the value by a scaling factor p to ensure
Figure FDA0002698957980000031
R′jk＝RjkP, in the formula, RjkThe values of the weight ratio between the adjacent indexes are judged for the alpha-th expert from 1.0, 1.2, 1.4, 1.6 and 1.8, which respectively represent the same importance, slightly importance, obvious importance, strong importance and extreme importance.
(iii) And determining subjective weight in a layering manner, wherein in the evaluation of the ith expert, the index subjective weight with lower relative importance degree is as follows:
Figure FDA0002698957980000032
by
Figure FDA0002698957980000033
The subjective weight of the evaluation of other index experts in the criterion layer is obtained in a recursion way
Figure FDA0002698957980000034
The subjective weight of each index of each level can be obtained by repeating the steps, so that the expert evaluation subjective weight of each index is as follows:
Figure FDA0002698957980000035
then, the subjective weight of each index is solved by adopting a weighted arithmetic mean method
Figure FDA0002698957980000036
Figure FDA0002698957980000037
Wherein l represents the maximum value of i;
in summary, the index set X ═ X1,x2,…,xnThe subjective weight coefficient set of each index
Figure FDA0002698957980000038
3. The method for extracting the reliability evaluation index feature of the power distribution main equipment sensor according to claim 1, wherein the step 4) further comprises at least:
(i) the coefficient of variation of each index is calculated by the following formula:
Figure FDA0002698957980000041
wherein σiIndicates the standard deviation of the i-th index,
Figure FDA0002698957980000042
represents the average of the i-th index,
objective weight of each index
Figure FDA0002698957980000043
Comprises the following steps:
(ii) the integrated weight ωt(t ═ 1,2, …, n) both subjective and objective weights are considered, defined as:
Figure FDA0002698957980000044
Figure FDA0002698957980000045
refers to the subjective weight of each index,
Figure FDA0002698957980000046
the objective weight of each index is indicated;
in summary, for index set X ═ X1,x2,＝,xnThe comprehensive weight coefficient set is V ═ omega1,ω2,…,ωnCorresponding to the comprehensive weight matrix D ═ omega0 ω1 … ωn]T；
(iii) And obtaining a normalized weighting decision matrix Z.
4. The distribution main equipment sensor reliability evaluation index feature extraction method according to claim 1, wherein a constant of (ii) in step 5) is set to 10-4。
5. The method for extracting the reliability evaluation index feature of the power distribution main equipment sensor according to claim 1, wherein the technical evaluation index of the step 1) comprises the following 16 indexes:
the method comprises the steps of node redundancy, channel redundancy, sampling frequency, source node link reliability, relay node link reliability, signal transmission performance, power supply stability, product working time, data testing time, average voltage deviation rate, flow benefit, fault statistical range, component response time, quality quantification, equipment duty ratio and maximum retransmission times.
6. The method for extracting the reliability evaluation index features of the power distribution main equipment sensor according to claim 1, wherein the device energy efficiency evaluation index of step 1) comprises the following 17 indexes:
the method comprises the steps of data packet sending energy consumption, data packet receiving energy consumption, power factor, load factor, winding temperature rise, lead temperature rise, no-load excitation current percentage, energy-saving hardware proportion, total harmonic distortion rate, three-phase load unbalance rate, rated no-load loss, rated load loss, link energy availability, wiring length exceeding rate, wiring sectional area exceeding rate, configuration and change and medium management.
7. The method for extracting the reliability evaluation index feature of the power distribution main equipment sensor according to claim 1, wherein the safety evaluation index of the step 1) comprises the following 16 indexes:
drift deviation fault, accuracy drop fault, fixed deviation fault, complete failure fault, local information security, patch security, information leakage probability, influence of denial of service, network attack frequency, communication network interference rate, channel packet loss rate, security mechanism perfection, signal transmission interruption probability, data transmission security, level authority completeness, and encrypted transmission.
8. The method for extracting the reliability evaluation index feature of the power distribution main equipment sensor according to claim 1, wherein the device operation condition evaluation index of step 1) comprises the following 17 indexes:
the method comprises the steps of sensor precision, sensor installation position, protocol distance applicability, end-to-end time delay, source node link operation condition, relay node link operation condition, real-time received data volume, service success rate, node energy availability, mean time to failure, task profile period, node connectivity probability, node capacity, link path flow, task transmission stability and routing signaling overhead.
CN202011015626.1A 2020-09-24 2020-09-24 Power distribution main equipment sensor reliability evaluation index feature extraction method Pending CN112183999A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011015626.1A CN112183999A (en) 2020-09-24 2020-09-24 Power distribution main equipment sensor reliability evaluation index feature extraction method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011015626.1A CN112183999A (en) 2020-09-24 2020-09-24 Power distribution main equipment sensor reliability evaluation index feature extraction method

Publications (1)

Publication Number Publication Date
CN112183999A true CN112183999A (en) 2021-01-05

Family

ID=73956617

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011015626.1A Pending CN112183999A (en) 2020-09-24 2020-09-24 Power distribution main equipment sensor reliability evaluation index feature extraction method

Country Status (1)

Country Link
CN (1) CN112183999A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113221435A (en) * 2021-02-08 2021-08-06 中国核电工程有限公司 Sensor screening method and device and sensor data reconstruction method and system
CN115774845A (en) * 2022-12-08 2023-03-10 德州易泰数据湖信息技术有限公司 Internet data comparison system
CN116489666A (en) * 2023-05-29 2023-07-25 河南鑫安利安全科技股份有限公司 Wireless sensor network construction method and system for safety monitoring
CN116629252A (en) * 2023-07-21 2023-08-22 无锡小净共享网络科技有限公司 Remote parameter adaptive configuration method and system based on Internet of things

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104156878A (en) * 2014-08-29 2014-11-19 国家电网公司 Method for determining weight of evaluation index of rural power grid renovation and upgrading project
CN104392390A (en) * 2014-11-13 2015-03-04 国家电网公司 TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) model-based intelligent substation secondary equipment evaluation method
CN105354764A (en) * 2015-11-16 2016-02-24 国网天津市电力公司 Method for evaluating power grid maintenance scheme
CN106611260A (en) * 2015-10-26 2017-05-03 国家电网公司 Establishment for evaluation model of residential area power distribution network engineering cost index system
CN111160719A (en) * 2019-12-10 2020-05-15 国网浙江省电力有限公司杭州供电公司 Two-network fused multi-index evaluation method
WO2020140560A1 (en) * 2019-05-31 2020-07-09 南京航空航天大学 Malfunction early-warning method for production logistics delivery equipment

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104156878A (en) * 2014-08-29 2014-11-19 国家电网公司 Method for determining weight of evaluation index of rural power grid renovation and upgrading project
CN104392390A (en) * 2014-11-13 2015-03-04 国家电网公司 TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) model-based intelligent substation secondary equipment evaluation method
CN106611260A (en) * 2015-10-26 2017-05-03 国家电网公司 Establishment for evaluation model of residential area power distribution network engineering cost index system
CN105354764A (en) * 2015-11-16 2016-02-24 国网天津市电力公司 Method for evaluating power grid maintenance scheme
WO2020140560A1 (en) * 2019-05-31 2020-07-09 南京航空航天大学 Malfunction early-warning method for production logistics delivery equipment
CN111160719A (en) * 2019-12-10 2020-05-15 国网浙江省电力有限公司杭州供电公司 Two-network fused multi-index evaluation method

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
LI JJ, YUAN GJ AND FAN H: "Generating Trimap for Image Matting Using Color Co-Fusion", 《IEEE XPLORE》 *
李荣辉: "基于改进局部线性嵌入的滚动轴承高维数据特征约简方法研究", 《中国优秀硕士学位论文全文数据库 工程科技II辑》 *
芮立: "基于多因子分析的无线传感器网络可靠性评估模型设计", 《现代电子技术》 *
高恩芝，王士同，张如艳: "基于测地距离逼近的降维算法", 《计算机工程与应用》 *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113221435A (en) * 2021-02-08 2021-08-06 中国核电工程有限公司 Sensor screening method and device and sensor data reconstruction method and system
WO2022166466A1 (en) * 2021-02-08 2022-08-11 中国核电工程有限公司 Sensor screening method and apparatus and sensor data reconstruction method and system
CN115774845A (en) * 2022-12-08 2023-03-10 德州易泰数据湖信息技术有限公司 Internet data comparison system
CN116489666A (en) * 2023-05-29 2023-07-25 河南鑫安利安全科技股份有限公司 Wireless sensor network construction method and system for safety monitoring
CN116629252A (en) * 2023-07-21 2023-08-22 无锡小净共享网络科技有限公司 Remote parameter adaptive configuration method and system based on Internet of things
CN116629252B (en) * 2023-07-21 2023-09-26 无锡小净共享网络科技有限公司 Remote parameter adaptive configuration method and system based on Internet of Things

Similar Documents

Publication Publication Date Title
CN112183999A (en) Power distribution main equipment sensor reliability evaluation index feature extraction method
CN105117602B (en) A kind of metering device running status method for early warning
CN108038300A (en) Optical fiber state evaluating method based on improved membership function combination neutral net
CN105069537B (en) A kind of construction method of combination type air Quality Prediction
CN107734512A (en) A kind of network selecting method based on the analysis of gray scale relevance presenting levelses
CN101572623A (en) Method for comprehensively evaluating network performance based on subjective and objective combination evaluation
CN110110977A (en) It is a kind of based on improve principal component-step analysis power distribution network appraisal procedure
CN107295537A (en) A kind of method and system for wireless sensor network reliability of testing and assessing
CN112668943A (en) Distribution line health state assessment method and system
CN114629802B (en) Service awareness-based power communication backbone network quality assessment method
CN109934500A (en) A kind of low-voltage evaluation method, device, equipment and the storage medium of power distribution station
CN107633349A (en) Fault impact factor quantitative analysis method based on high-voltage switch gear
CN104991549A (en) Track circuit red-light strip default diagnosis method based on FTA and multilevel fuzzy-neural sub-networks
CN110705887A (en) Low-voltage transformer area operation state comprehensive evaluation method based on neural network model
CN112202597A (en) Method for evaluating importance of communication network node in low-voltage distribution area
CN105912857A (en) Selection and configuration method of distribution equipment state monitoring sensors
CN110571926A (en) intelligent power distribution network based on Internet of things technology and data model construction method thereof
CN109711687B (en) Insulator state fuzzy evaluation method based on improved entropy method
CN114757392A (en) Power grid system comprehensive evaluation index system optimization method based on game weight method
CN110533300B (en) Intelligent decision-making system for transformer based on game set pair cloud
CN113591402A (en) Digital power transformer health state fuzzy comprehensive evaluation method
CN116667881B (en) Electric power communication network key link evaluation method based on rapid density clustering
CN111612227A (en) Load prediction method based on K-means clustering and bat optimization neural network
CN116151799A (en) BP neural network-based distribution line multi-working-condition fault rate rapid assessment method
AU2021106280A4 (en) A Method of Water Environment Risk Assessment Based on Fuzzy Integral Model

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20210105