CN113327024A - Power distribution network weak link identification method and system and storage medium - Google Patents

Power distribution network weak link identification method and system and storage medium Download PDF

Info

Publication number
CN113327024A
CN113327024A CN202110570760.6A CN202110570760A CN113327024A CN 113327024 A CN113327024 A CN 113327024A CN 202110570760 A CN202110570760 A CN 202110570760A CN 113327024 A CN113327024 A CN 113327024A
Authority
CN
China
Prior art keywords
evaluation index
neuron
training
power distribution
euclidean distance
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
CN202110570760.6A
Other languages
Chinese (zh)
Inventor
欧阳健娜
陈绍南
秦丽文
李珊
周杨珺
李欣桐
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.)
Electric Power Research Institute of Guangxi Power Grid Co Ltd
Original Assignee
Electric Power Research Institute of Guangxi Power Grid 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 Electric Power Research Institute of Guangxi Power Grid Co Ltd filed Critical Electric Power Research Institute of Guangxi Power Grid Co Ltd
Priority to CN202110570760.6A priority Critical patent/CN113327024A/en
Publication of CN113327024A publication Critical patent/CN113327024A/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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • 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
    • G06Q10/0635—Risk analysis of enterprise or organisation activities
    • 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)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Economics (AREA)
  • General Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Development Economics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Marketing (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Educational Administration (AREA)
  • Computational Linguistics (AREA)
  • Biomedical Technology (AREA)
  • Quality & Reliability (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • Computing Systems (AREA)
  • Molecular Biology (AREA)
  • Game Theory and Decision Science (AREA)
  • Biophysics (AREA)
  • Operations Research (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • Primary Health Care (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a method, a system and a storage medium for identifying weak links of a power distribution network, wherein the method comprises the following steps: acquiring a plurality of relevant factors influencing the risk state of the distribution transformer, and selecting a main evaluation index set from the relevant factors; performing vacancy filling and normalization processing on all data associated with each type of evaluation index in the main evaluation index set to obtain an input training set; and initializing the built LVQ neural network, importing the input training set into the LVQ neural network for training and analysis, and predicting the risk level of the distribution transformer. In the embodiment of the invention, the LVQ neural network can realize deep measurement and risk classification of various evaluation indexes of the power distribution network, thereby providing a valuable theoretical basis for generating a decision for the power distribution network.

Description

Power distribution network weak link identification method and system and storage medium
Technical Field
The invention relates to the technical field of power distribution network monitoring, in particular to a method and a system for identifying weak links of a power distribution network and a storage medium.
Background
With the large-scale development of modern power systems, the number and structure of system elements become very complex due to the implementation of power system intelligence, and the influence of each element on the reliability of the power system is different. Many current risk identification methods quantitatively reflect the reliability state of an electric power system by calculating the reliability evaluation index of the electric power system, and set a reliability index threshold corresponding to risk identification to realize identification of weak links of a power distribution network, but the result only reflects the risk state of the power distribution network, but cannot identify power distribution equipment or power distribution lines which particularly affect the operation state of the whole power distribution network. In the past, researchers also propose a reliability tracking method to realize power distribution network risk identification, appropriately distribute reliability indexes to each component based on a tracking concept, namely, give a distribution relation between each component and a power distribution network reliability index, calculate and analyze the contribution degree of each component to the power distribution network reliability index, and accordingly find out components which obviously affect the power distribution network reliability. In addition, other scholars also put forward that fault identification and risk indexes are added into a power distribution network operation state evaluation index system, and the risk types of the power distribution network are judged by calculating the risk indexes, but the method especially depends on data analysis, has extremely high requirement on the accuracy of data, and can accurately position the risks to the local parts only under the condition of ensuring that no or few bad data exist.
Disclosure of Invention
The invention aims to overcome the defects of the prior art, and provides a power distribution network weak link identification method, a power distribution network weak link identification system and a storage medium.
In order to solve the problems, the invention provides a power distribution network weak link identification method, which comprises the following steps:
acquiring a plurality of relevant factors influencing the risk state of the distribution transformer, and selecting a main evaluation index set from the relevant factors;
performing vacancy filling and normalization processing on all data associated with each type of evaluation index in the main evaluation index set to obtain an input training set;
and initializing the built LVQ neural network, importing the input training set into the LVQ neural network for training and analysis, and predicting the risk level of the distribution transformer.
Optionally, the main evaluation index set includes a health degree evaluation index set and an importance evaluation index set, where the health degree evaluation index set includes a primary side voltage value, a heavy overload frequency, a low voltage frequency, a historical defect state, and an equipment operation life, and the importance evaluation index set includes a rated capacity, a number of important users, a social influence coefficient, and a regional coefficient.
Optionally, the initializing the built LVQ neural network includes:
determining the number of competition layer neurons in the LVQ neural network as M1The method comprises the steps of obtaining a connection weight matrix between an input layer and a competition layer, obtaining an initial neighborhood of the competition layer and an initial learning rate, wherein the connection weight matrix comprises M1An individual neuron weight vector.
Optionally, the importing the input training set into the LVQ neural network for training and analysis, and predicting the risk level of the distribution transformer includes:
training and updating a connection weight matrix in the LVQ neural network by using the input training set, and updating the initial neighborhood of the competition layer and the initial learning rate based on the training times;
and performing predictive response on the input training set based on the updated LVQ neural network, and outputting the risk level of the distribution transformer.
Optionally, the training and updating the connection weight matrix in the LVQ neural network by using the input training set includes:
training a set of training sets to include M based on the input2A subset of evaluation indexes, calculating the M2The Euclidean distance between each evaluation index subset and each neuron weight vector is obtained, and M is obtained2(ii) a euclidean distance;
from the M2Screening the Euclidean distances to obtain a minimum Euclidean distance, and defining the neuron associated with the minimum Euclidean distance as an optimal neuron;
from the M2Screening the Euclidean distances to obtain a sub-small Euclidean distance, and defining the neuron associated with the sub-small Euclidean distance as a suboptimal neuron;
and adjusting the weight vectors of the optimal neuron and the suboptimal neuron based on that the relation between the minimum Euclidean distance and the sub-minimum Euclidean distance meets an error limiting condition.
Optionally, a relationship between the minimum euclidean distance and the second minimum euclidean distance satisfies an error limiting condition, and a corresponding expression is as follows:
Figure BDA0003082593610000031
wherein d isaIs a minimum Euclidean distance, dbZeta is the error accuracy for the next smallest Euclidean distance.
Optionally, the adjusting the weight vectors of the optimal neuron and the suboptimal neuron includes:
and acquiring a pre-designated classification of the input training set, and adjusting the weight vectors of the optimal neuron and the suboptimal neuron according to a matching result of the optimal neuron and the pre-designated classification and a matching result of the suboptimal neuron and the pre-designated classification.
Optionally, the updating the initial neighborhood of the competition layer and the initial learning rate based on the training times includes:
based on the training times, updating the initial neighborhood of the competition layer into:
Figure BDA0003082593610000032
based on the training times, updating the initial learning rate to:
Figure BDA0003082593610000033
wherein N isg(t) is the updated current competition layer neighborhood, Ng(0) For initial neighborhood of competition layer, T is training time, T is total training time, etatFor updated current learning rate, η0Is the initial learning rate.
In addition, the embodiment of the invention also provides a system for identifying weak links of a power distribution network, which comprises the following steps:
the system comprises a selection module, a calculation module and a calculation module, wherein the selection module is used for acquiring a plurality of related factors influencing the risk state of the distribution transformer and selecting a main evaluation index set from the related factors;
the processing module is used for carrying out vacancy filling and normalization processing on all data associated with each type of evaluation index in the main evaluation index set to obtain an input training set;
and the training module is used for initializing the built LVQ neural network, importing the input training set into the LVQ neural network for training and analysis, and predicting the risk level of the distribution transformer.
In addition, the embodiment of the invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method for identifying the weak link of the power distribution network is implemented.
In the embodiment of the invention, relevant evaluation index information is extracted from the aspects of the health degree and the importance degree of the power distribution network, and then the relation between various evaluation indexes and risk levels is deeply measured by using the LVQ neural network in a machine learning mode, so that the risk prediction and classification of various evaluation indexes of the power distribution network are realized, and a valuable theoretical basis is provided for the generation decision of the power distribution network.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart of a power distribution network weak link identification method in an embodiment of the present invention;
fig. 2 is a schematic structural composition diagram of a power distribution network weak link identification system in the embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Examples
Referring to fig. 1, fig. 1 is a schematic flow chart illustrating a method for identifying weak links of a power distribution network according to an embodiment of the present invention.
As shown in fig. 1, a method for identifying weak links of a power distribution network includes the following steps:
s101, obtaining a plurality of relevant factors influencing the risk state of the distribution transformer, and selecting a main evaluation index set from the relevant factors;
in the embodiment of the present invention, the main evaluation index set includes a health degree evaluation index set and an importance evaluation index set, where the health degree evaluation index set includes a primary side voltage value, a heavy overload number, a low voltage number, a historical defect state, and an equipment operation age, and the importance evaluation index set includes a rated capacity, a number of important users, a social influence coefficient, and an area coefficient, that is, the main evaluation index set actually includes nine types of evaluation indexes.
S102, performing vacancy filling and normalization processing on all data associated with each type of evaluation index in the main evaluation index set to obtain an input training set;
the implementation process of the invention is as follows: and filling up the vacancy of all data associated with each type of evaluation index by using a traditional Lagrange interpolation method, and normalizing each data in each type of processed evaluation index into:
Figure BDA0003082593610000051
and then obtaining the input vector corresponding to each type of evaluation index as follows:
Figure BDA0003082593610000052
the final input training set is:
Figure BDA0003082593610000053
wherein,
Figure BDA0003082593610000054
is the normalized value of the ith data in the jth type evaluation index,
Figure BDA0003082593610000055
as in the j-th evaluation indexThe (i) th data is (are) stored in the storage unit,
Figure BDA0003082593610000056
is the minimum data in the j-th type evaluation index,
Figure BDA0003082593610000057
is the maximum data in the j-th evaluation index, n is the total number of data in the j-th evaluation index, XjIs an input vector corresponding to the j-th evaluation index, M2The total number of types of evaluation indexes included in the main evaluation index set.
S103, initializing the built LVQ neural network, importing the input training set into the LVQ neural network for training and analysis, and predicting the risk level of the distribution transformer.
The implementation process of the invention comprises the following steps:
(1) determining the number of neurons in the competition layer in the LVQ (Learning Vector Quantization) neural network as M1The method comprises the steps of obtaining a connection weight matrix between an input layer and a competition layer, obtaining an initial neighborhood of the competition layer and an initial learning rate, wherein the connection weight matrix comprises M1A neuron weight vector;
(2) training and updating a connection weight matrix in the LVQ neural network by using the input training set, and updating the initial neighborhood of the competition layer and the initial learning rate based on the training times, which is specifically represented as follows:
step A, based on the fact that the input training set X actually contains M2A subset of evaluation indexes, calculating the M2The Euclidean distance between each evaluation index subset and each neuron weight vector is obtained, and M is obtained2Euclidean distance, wherein M2The calculation formula of the Euclidean distance between each evaluation index subset and each neuron weight vector is as follows:
Figure BDA0003082593610000061
in the formula: dkIs said M2Euclidean distance, W, between each evaluation index subset and the kth neuron weight vectorkIs the weight vector of the kth neuron;
step B. from said M2Screening the Euclidean distances to obtain the minimum Euclidean distance daThe minimum Euclidean distance d is setaDefining the associated neuron as an optimal neuron a;
step C. from said M2The next small Euclidean distance d is obtained by screening among the Euclidean distancesbThe sub-minimum Euclidean distance dbDefining the associated neuron as a suboptimal neuron b;
based on the minimum Euclidean distance daAnd said sub-minimum Euclidean distance dbSatisfies the error limiting condition, i.e. satisfies the relation min (d)a/db，db/da) > (1-zeta)/(1 + zeta), adjusting the weight vectors of the optimal neuron a and the suboptimal neuron b;
further, obtaining a pre-specified classification of the input training set X, and adjusting the weight vectors of the optimal neuron a and the suboptimal neuron b according to a matching result of the optimal neuron a and the pre-specified classification and a matching result of the suboptimal neuron b and the pre-specified classification, as follows:
when the optimal neuron a is consistent with the pre-specified classification and the suboptimal neuron b is inconsistent with the pre-specified classification, adjusting the weight vectors of the optimal neuron a and the suboptimal neuron b to be as follows:
Figure BDA0003082593610000062
when the suboptimal neuron b is consistent with the pre-specified classification and the optimal neuron a is inconsistent with the pre-specified classification, adjusting the weight vectors of the optimal neuron a and the suboptimal neuron b to be as follows:
Figure BDA0003082593610000071
in the formula: ζ is the error accuracy, wa(t) is the weight vector of the optimal neuron a utilized in the t-th network learning process, wa(t +1) is the weight vector of the optimal neuron a utilized in the (t +1) th network learning process, wb(t) is the weight vector of suboptimal neuron b utilized in the tth network learning process, wb(t +1) is the weight vector of the suboptimal neuron b utilized in the (t +1) th network learning process, etatThe learning rate utilized in the network learning process of the t time.
Step E, updating the initial neighborhood of the competition layer and the initial learning rate based on the training times as follows:
Figure BDA0003082593610000072
in the formula: n is a radical ofg(t) is the updated current competition layer neighborhood, Ng(0) For initial neighborhood of competition layer, T is training time, T is total training time, etatFor updated current learning rate, η0Is the initial learning rate.
(3) And performing predictive response on the input training set based on the updated LVQ neural network, and outputting the risk level of the distribution transformer.
In the embodiment of the invention, relevant evaluation index information is extracted from the aspects of the health degree and the importance degree of the power distribution network, and then the relation between various evaluation indexes and risk levels is deeply measured by using the LVQ neural network in a machine learning mode, so that the risk prediction and classification of various evaluation indexes of the power distribution network are realized, and a valuable theoretical basis is provided for the generation decision of the power distribution network.
Examples
Referring to fig. 2, fig. 2 is a schematic structural composition diagram of a power distribution network weak link identification system according to an embodiment of the present invention.
As shown in fig. 2, a system for identifying weak links of a power distribution network includes the following:
the system comprises a selection module 201, a calculation module and a calculation module, wherein the selection module 201 is used for acquiring a plurality of relevant factors influencing the risk state of the distribution transformer and selecting a main evaluation index set from the relevant factors;
in the embodiment of the present invention, the main evaluation index set includes a health degree evaluation index set and an importance evaluation index set, where the health degree evaluation index set includes a primary side voltage value, a heavy overload number, a low voltage number, a historical defect state, and an equipment operation age, and the importance evaluation index set includes a rated capacity, a number of important users, a social influence coefficient, and an area coefficient, that is, the main evaluation index set actually includes nine types of evaluation indexes.
A processing module 202, configured to perform gap filling and normalization processing on all data associated with each type of evaluation index in the main evaluation index set, to obtain an input training set;
the implementation process of the invention is as follows: and filling up the vacancy of all data associated with each type of evaluation index by using a traditional Lagrange interpolation method, and normalizing each data in each type of processed evaluation index into:
Figure BDA0003082593610000081
and then obtaining the input vector corresponding to each type of evaluation index as follows:
Figure BDA0003082593610000082
the final input training set is:
Figure BDA0003082593610000083
wherein,
Figure BDA0003082593610000084
is the normalized value of the ith data in the jth type evaluation index,
Figure BDA0003082593610000085
for the ith data in the jth type evaluation index,
Figure BDA0003082593610000086
is the minimum data in the j-th type evaluation index,
Figure BDA0003082593610000087
is the maximum data in the j-th evaluation index, n is the total number of data in the j-th evaluation index, XjIs an input vector corresponding to the j-th evaluation index, M2The total number of types of evaluation indexes included in the main evaluation index set.
And the training module 203 is used for initializing the established LVQ neural network, importing the input training set into the LVQ neural network for training and analysis, and predicting the risk level of the distribution transformer.
The implementation process of the invention comprises the following steps:
(1) determining the number of neurons in the competition layer in the LVQ (Learning Vector Quantization) neural network as M1The method comprises the steps of obtaining a connection weight matrix between an input layer and a competition layer, obtaining an initial neighborhood of the competition layer and an initial learning rate, wherein the connection weight matrix comprises M1A neuron weight vector;
(2) training and updating a connection weight matrix in the LVQ neural network by using the input training set, and updating the initial neighborhood of the competition layer and the initial learning rate based on the training times, which is specifically represented as follows:
step A, based on the fact that the input training set X actually contains M2A subset of evaluation indexes, calculating the M2The Euclidean distance between each evaluation index subset and each neuron weight vector is obtained, and M is obtained2Euclidean distance, wherein M2The subset of evaluation indexes and the weight direction of each neuronThe formula for calculating the euclidean distance between quantities is:
Figure BDA0003082593610000091
in the formula: dkIs said M2Euclidean distance, W, between each evaluation index subset and the kth neuron weight vectorkIs the weight vector of the kth neuron;
step B. from said M2Screening the Euclidean distances to obtain the minimum Euclidean distance daThe minimum Euclidean distance d is setaDefining the associated neuron as an optimal neuron a;
step C. from said M2The next small Euclidean distance d is obtained by screening among the Euclidean distancesbThe sub-minimum Euclidean distance dbDefining the associated neuron as a suboptimal neuron b;
based on the minimum Euclidean distance daAnd said sub-minimum Euclidean distance dbSatisfies the error limiting condition, i.e. satisfies the relation min (d)a/db，db/da) > (1-zeta)/(1 + zeta), adjusting the weight vectors of the optimal neuron a and the suboptimal neuron b;
further, obtaining a pre-specified classification of the input training set X, and adjusting the weight vectors of the optimal neuron a and the suboptimal neuron b according to a matching result of the optimal neuron a and the pre-specified classification and a matching result of the suboptimal neuron b and the pre-specified classification, as follows:
when the optimal neuron a is consistent with the pre-specified classification and the suboptimal neuron b is inconsistent with the pre-specified classification, adjusting the weight vectors of the optimal neuron a and the suboptimal neuron b to be as follows:
Figure BDA0003082593610000092
when the suboptimal neuron b is consistent with the pre-specified classification and the optimal neuron a is inconsistent with the pre-specified classification, adjusting the weight vectors of the optimal neuron a and the suboptimal neuron b to be as follows:
Figure BDA0003082593610000101
in the formula: ζ is the error accuracy, wa(t) is the weight vector of the optimal neuron a utilized in the t-th network learning process, wa(t +1) is the weight vector of the optimal neuron a utilized in the (t +1) th network learning process, wb(t) is the weight vector of suboptimal neuron b utilized in the tth network learning process, wb(t +1) is the weight vector of the suboptimal neuron b utilized in the (t +1) th network learning process, etatThe learning rate utilized in the network learning process of the t time.
Step E, updating the initial neighborhood of the competition layer and the initial learning rate based on the training times as follows:
Figure BDA0003082593610000102
in the formula: n is a radical ofg(t) is the updated current competition layer neighborhood, Ng(0) For initial neighborhood of competition layer, T is training time, T is total training time, etatFor updated current learning rate, η0Is the initial learning rate.
(3) And performing predictive response on the input training set based on the updated LVQ neural network, and outputting the risk level of the distribution transformer.
In the embodiment of the invention, relevant evaluation index information is extracted from the aspects of the health degree and the importance degree of the power distribution network, and then the relation between various evaluation indexes and risk levels is deeply measured by using the LVQ neural network in a machine learning mode, so that the risk prediction and classification of various evaluation indexes of the power distribution network are realized, and a valuable theoretical basis is provided for the generation decision of the power distribution network.
In the computer-readable storage medium provided by the embodiment of the present invention, a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the method for identifying a weak link of a power distribution network in the above embodiments is implemented. The computer-readable storage medium includes, but is not limited to, any type of disk including floppy disks, hard disks, optical disks, CD-ROMs, and magnetic-optical disks, ROMs (Read-Only memories), RAMs (Random AcceSS memories), EPROMs (EraSable Programmable Read-Only memories), EEPROMs (Electrically EraSable Programmable Read-Only memories), flash memories, magnetic cards, or optical cards. That is, a storage device includes any medium that stores or transmits information in a form readable by a device (e.g., a computer, a cellular phone), and may be a read-only memory, a magnetic or optical disk, or the like.
The method, the system and the storage medium for identifying the weak link of the power distribution network provided by the embodiment of the invention are described in detail, a specific embodiment is adopted in the method for explaining the principle and the implementation mode of the invention, and the description of the embodiment is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.

Claims (10)

1. A power distribution network weak link identification method is characterized by comprising the following steps:
acquiring a plurality of relevant factors influencing the risk state of the distribution transformer, and selecting a main evaluation index set from the relevant factors;
performing vacancy filling and normalization processing on all data associated with each type of evaluation index in the main evaluation index set to obtain an input training set;
and initializing the built LVQ neural network, importing the input training set into the LVQ neural network for training and analysis, and predicting the risk level of the distribution transformer.
2. The power distribution network weak link identification method according to claim 1, wherein the main evaluation index set comprises a health evaluation index set and an importance evaluation index set, wherein the health evaluation index set comprises a primary side voltage value, a heavy overload frequency, a low voltage frequency, a historical defect state and an equipment operation life, and the importance evaluation index set comprises a rated capacity, a number of important users, a social influence coefficient and a regional coefficient.
3. The method for identifying weak links of a power distribution network according to claim 1, wherein the initializing the constructed LVQ neural network comprises:
determining the number of competition layer neurons in the LVQ neural network as M1The method comprises the steps of obtaining a connection weight matrix between an input layer and a competition layer, obtaining an initial neighborhood of the competition layer and an initial learning rate, wherein the connection weight matrix comprises M1An individual neuron weight vector.
4. The method for identifying weak links of a power distribution network according to claim 3, wherein the importing the input training set into the LVQ neural network for training analysis, and the predicting the risk level of the distribution transformer comprises:
training and updating a connection weight matrix in the LVQ neural network by using the input training set, and updating the initial neighborhood of the competition layer and the initial learning rate based on the training times;
and performing predictive response on the input training set based on the updated LVQ neural network, and outputting the risk level of the distribution transformer.
5. The method for identifying weak links of a power distribution network according to claim 4, wherein the training and updating of the connection weight matrix in the LVQ neural network by using the input training set comprises:
training a set of training sets to include M based on the input2A subset of evaluation indexes, calculating the M2The Euclidean distance between each evaluation index subset and each neuron weight vector is obtained, and M is obtained2(ii) a euclidean distance;
from the M2Screening the Euclidean distances to obtain a minimum Euclidean distance, and defining the neuron associated with the minimum Euclidean distance as an optimal neuron;
from the M2Screening the Euclidean distances to obtain a sub-small Euclidean distance, and defining the neuron associated with the sub-small Euclidean distance as a suboptimal neuron;
and adjusting the weight vectors of the optimal neuron and the suboptimal neuron based on that the relation between the minimum Euclidean distance and the sub-minimum Euclidean distance meets an error limiting condition.
6. The method for identifying the weak link of the power distribution network according to claim 5, wherein a relationship between the minimum Euclidean distance and the second minimum Euclidean distance satisfies an error limiting condition, and a corresponding expression is as follows:
Figure FDA0003082593600000021
wherein d isaIs a minimum Euclidean distance, dbZeta is the error accuracy for the next smallest Euclidean distance.
7. The method for identifying weak links of a power distribution network according to claim 5, wherein the adjusting the weight vectors of the optimal neurons and the suboptimal neurons comprises:
and acquiring a pre-designated classification of the input training set, and adjusting the weight vectors of the optimal neuron and the suboptimal neuron according to a matching result of the optimal neuron and the pre-designated classification and a matching result of the suboptimal neuron and the pre-designated classification.
8. The power distribution network weak link identification method according to claim 4, wherein the updating the initial neighborhood of the competition layer and the initial learning rate based on the training times comprises:
based on the training times, updating the initial neighborhood of the competition layer into:
Figure FDA0003082593600000031
based on the training times, updating the initial learning rate to:
Figure FDA0003082593600000032
wherein N isg(t) is the updated current competition layer neighborhood, Ng(0) For initial neighborhood of competition layer, T is training time, T is total training time, etatFor updated current learning rate, η0Is the initial learning rate.
9. A system for identifying weak links in a power distribution network, the system comprising:
the system comprises a selection module, a calculation module and a calculation module, wherein the selection module is used for acquiring a plurality of related factors influencing the risk state of the distribution transformer and selecting a main evaluation index set from the related factors;
the processing module is used for carrying out vacancy filling and normalization processing on all data associated with each type of evaluation index in the main evaluation index set to obtain an input training set;
and the training module is used for initializing the built LVQ neural network, importing the input training set into the LVQ neural network for training and analysis, and predicting the risk level of the distribution transformer.
10. A computer-readable storage medium, on which a computer program is stored, wherein the computer program, when being executed by a processor, implements the method for identifying weak links in a power distribution network according to any one of claims 1 to 8.
CN202110570760.6A 2021-05-25 2021-05-25 Power distribution network weak link identification method and system and storage medium Pending CN113327024A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110570760.6A CN113327024A (en) 2021-05-25 2021-05-25 Power distribution network weak link identification method and system and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110570760.6A CN113327024A (en) 2021-05-25 2021-05-25 Power distribution network weak link identification method and system and storage medium

Publications (1)

Publication Number Publication Date
CN113327024A true CN113327024A (en) 2021-08-31

Family

ID=77416608

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110570760.6A Pending CN113327024A (en) 2021-05-25 2021-05-25 Power distribution network weak link identification method and system and storage medium

Country Status (1)

Country Link
CN (1) CN113327024A (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104616091A (en) * 2014-11-19 2015-05-13 南昌大学 Analytic hierarchy process based comprehensive analysis method for urban distribution network
CN110378610A (en) * 2019-07-25 2019-10-25 广西电网有限责任公司电力科学研究院 Distribution weak link identification method based on user's different degree and equipment running status

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104616091A (en) * 2014-11-19 2015-05-13 南昌大学 Analytic hierarchy process based comprehensive analysis method for urban distribution network
CN110378610A (en) * 2019-07-25 2019-10-25 广西电网有限责任公司电力科学研究院 Distribution weak link identification method based on user's different degree and equipment running status

Similar Documents

Publication Publication Date Title
CN107506868B (en) Method and device for predicting short-time power load
WO2023116111A1 (en) Disk fault prediction method and apparatus
CN112508442B (en) Transient stability assessment method and system based on automatic and interpretable machine learning
CN113887916A (en) Dynamic quantitative evaluation method and system for line loss of power distribution network
CN117851921B (en) Equipment life prediction method and device based on transfer learning
CN110837915A (en) Low-voltage load point prediction and probability prediction method for power system based on hybrid integrated deep learning
CN111582315B (en) Sample data processing method and device and electronic equipment
CN111311001B (en) Bi-LSTM network short-term load prediction method based on DBSCAN algorithm and feature selection
CN112364098A (en) Hadoop-based distributed power system abnormal data identification method and system
CN118032327B (en) Equipment intelligent lubrication monitoring method and device based on artificial intelligence
CN117674119A (en) Power grid operation risk assessment method, device, computer equipment and storage medium
CN115204698A (en) Real-time analysis method for power supply stability of low-voltage transformer area
CN113033898A (en) Electrical load prediction method and system based on K-means clustering and BI-LSTM neural network
CN111723010B (en) Software BUG classification method based on sparse cost matrix
CN113988558A (en) Power grid dynamic safety assessment method based on blind area identification and electric coordinate system expansion
CN117368789A (en) Power failure co-location method and system based on multi-source heterogeneous data
CN113298296A (en) Method for predicting day-ahead load probability of power transmission substation from bottom to top
CN117290673A (en) Ship energy consumption high-precision prediction system based on multi-model fusion
CN110717577A (en) Time series prediction model construction method for noting regional information similarity
CN113327024A (en) Power distribution network weak link identification method and system and storage medium
CN116561569A (en) Industrial power load identification method based on EO feature selection and AdaBoost algorithm
CN111026661B (en) Comprehensive testing method and system for software usability
Xiao-Xu et al. An intelligent inspection robot of power distribution network based on image automatic recognition system
CN111160419B (en) Deep learning-based electronic transformer data classification prediction method and device
CN116979532B (en) Power system dynamic state estimation observability evaluation method and system

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: 20210831