CN117454315B - Man-machine terminal picture data interaction method and system - Google Patents
Man-machine terminal picture data interaction method and system Download PDFInfo
- Publication number
- CN117454315B CN117454315B CN202311770009.6A CN202311770009A CN117454315B CN 117454315 B CN117454315 B CN 117454315B CN 202311770009 A CN202311770009 A CN 202311770009A CN 117454315 B CN117454315 B CN 117454315B
- Authority
- CN
- China
- Prior art keywords
- node
- traveling wave
- fault
- features
- determining
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 39
- 230000003993 interaction Effects 0.000 title claims abstract description 19
- 239000013598 vector Substances 0.000 claims abstract description 56
- 238000007637 random forest analysis Methods 0.000 claims abstract description 53
- 238000013528 artificial neural network Methods 0.000 claims abstract description 43
- 230000004927 fusion Effects 0.000 claims abstract description 22
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 21
- 230000007246 mechanism Effects 0.000 claims abstract description 20
- 238000012549 training Methods 0.000 claims description 15
- 230000006870 function Effects 0.000 claims description 14
- 230000008569 process Effects 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 5
- 239000011159 matrix material Substances 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 4
- 230000009466 transformation Effects 0.000 claims description 4
- 238000001514 detection method Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 description 6
- 230000008859 change Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000001902 propagating effect Effects 0.000 description 2
- 238000001228 spectrum Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 230000003213 activating effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000019771 cognition Effects 0.000 description 1
- 230000001149 cognitive effect Effects 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000013399 early diagnosis Methods 0.000 description 1
- 230000005670 electromagnetic radiation Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000003012 network analysis Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 238000010183 spectrum analysis Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000010937 topological data analysis Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- 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
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/042—Knowledge-based neural networks; Logical representations of neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
-
- 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
- Y04S10/52—Outage or fault management, e.g. fault detection or location
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Business, Economics & Management (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Economics (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Public Health (AREA)
- Probability & Statistics with Applications (AREA)
- Water Supply & Treatment (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Testing And Monitoring For Control Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a man-machine terminal picture data interaction method and a system, which relate to the technical field of data interaction and comprise the following steps: installing a traveling wave sensor, acquiring fault electromagnetic waves, determining a time stamp and traveling wave propagation speed, and determining a fault area through a fault positioning algorithm; acquiring traveling wave signal data of fault electromagnetic waves according to a fault region, simultaneously acquiring power grid topology data, extracting traveling wave characteristics and topology characteristics, determining structural characteristics, identifying the shortest path of nodes in a power grid topology graph, and combining clustering coefficients to perform characteristic fusion to obtain a comprehensive topology vector; the method comprises the steps of encoding a comprehensive topological vector based on a graph neural network, carrying out node feature fusion on traveling wave features and the encoded comprehensive topological vector, generating initial node features, updating the initial node features through the graph neural network by combining a attention mechanism introduced in advance, generating input node features, combining the traveling wave features, carrying out fault location through a random forest module in the graph neural network, and determining a fault position.
Description
Technical Field
The invention relates to the technical field of data interaction, in particular to a method and a system for interaction of picture data of a man-machine terminal.
Background
Equipment faults in a power system can have a serious impact on the overall system, requiring quick and accurate fault diagnosis and prediction. Through man-machine terminal picture data interaction, operation and maintenance personnel can utilize machine learning and data analysis technology to realize early diagnosis and prediction of power equipment faults
In the prior art, CN113726856A discloses a micro-service-based light-weight interaction method and system for regulating and controlling picture comprehensive data, wherein the method comprises the following steps: classifying functions deployed under a single application program of a client into two types of necessary functions and unnecessary functions; dividing unnecessary functions under a single application program into a plurality of micro services with single functions according to the functional characteristics; deploying a plurality of micro services with single functions on a plurality of servers at a server side to form a server cluster; receiving a call request sent by a client, and selecting a server for operating a request service according to the operating state of a server cluster; and returning the result of the server running the service to the client.
In summary, although the interaction of the picture data of the man-machine terminal can be realized in the prior art, only the preset simple requirement can be realized, and the real-time interaction of the information in the power grid system can not be realized, so that a scheme is needed to solve the problems in the prior art.
Disclosure of Invention
The embodiment of the invention provides a man-machine terminal picture data interaction method and system, which are used for diagnosing power grid faults based on power grid data and realizing interaction through man-machine terminal picture data.
In a first aspect of the embodiment of the present invention, a method for interaction of picture data of a human-machine terminal is provided, including:
Installing a traveling wave sensor, acquiring fault electromagnetic waves according to the traveling wave sensor, determining a time stamp corresponding to the fault electromagnetic waves and traveling wave propagation speed of the fault electromagnetic waves, and determining a fault area through a fault positioning algorithm according to the time stamp and the traveling wave propagation speed;
Acquiring traveling wave signal data of the fault electromagnetic wave according to the fault region, acquiring power grid topology data, extracting traveling wave characteristics in the traveling wave signal data and topology characteristics in the power grid topology data, determining structural characteristics according to the topology characteristics, identifying a shortest path of a node in a power grid topology graph, and performing characteristic fusion according to the shortest path of the node and a cluster coefficient introduced in advance to obtain a comprehensive topology vector;
And encoding the comprehensive topological vector based on a preset graph neural network, carrying out node feature fusion on the traveling wave feature and the encoded comprehensive topological vector, generating initial node features, updating the initial node features through the graph neural network according to the initial node features and a attention mechanism introduced in advance, generating input node features, carrying out fault location through a random forest module in the graph neural network according to the input node features and the traveling wave features, and determining the fault position.
In an alternative embodiment of the present invention,
The step of installing the traveling wave sensor, the step of obtaining fault electromagnetic waves according to the traveling wave sensor, the step of determining a time stamp corresponding to the fault electromagnetic waves and traveling wave propagation speed of the fault electromagnetic waves, and the step of determining a fault area through a fault positioning algorithm according to the time stamp and the traveling wave propagation speed comprises the following steps:
installing a traveling wave sensor in a power line, acquiring fault electromagnetic waves caused by faults according to the traveling wave sensor, adding a time stamp to traveling wave detection data according to the time of acquiring the fault electromagnetic waves by the traveling wave sensor, determining the line length and the wire type in the power line, and determining the traveling wave propagation speed of the fault electromagnetic waves;
Judging the number of traveling wave sensors in a fault occurrence area according to the time stamp and the traveling wave propagation speed, if two or more traveling wave sensors exist, determining a fault area by comparing the arrival time of the waveform of the fault electromagnetic wave at different traveling wave sensors, and if only one traveling wave sensor exists, determining the fault area by reflecting the traveling wave information.
In an alternative embodiment of the present invention,
The step of obtaining traveling wave signal data of the fault electromagnetic wave according to the fault region, and simultaneously obtaining power grid topology data, wherein the step of extracting traveling wave characteristics in the traveling wave signal data and topology characteristics in the power grid topology data comprises the following steps:
Acquiring amplitude, frequency, energy distribution and phase angle of the fault electromagnetic wave according to the fault region, recording the amplitude, frequency, energy distribution and phase angle as traveling wave data signals, and acquiring power grid topology data comprising node information, connection relation and line parameters;
Preprocessing the traveling wave data signal, extracting amplitude, energy and duration in the traveling wave data signal, analyzing frequency components of the traveling wave data signal through Fourier transformation, and combining the extracted data into traveling wave characteristics;
and identifying each node in the power grid and the attribute corresponding to the node, determining a connection type and a connection parameter according to the connection relation of the nodes, and combining the connection relation, the connection type and the connection parameter into a topological feature.
In an alternative embodiment of the present invention,
Determining structural features according to the topological features, identifying shortest paths of nodes in a power grid topological graph, performing feature fusion according to the shortest paths of the nodes and a cluster coefficient introduced in advance, and obtaining a comprehensive topological vector comprises the following steps:
according to the topological characteristics, for each node in the power grid, calculating the number of nodes directly connected with the current node, recording the number as the divergence of the current node, extracting node information directly connected with the current node according to the divergence, and recording the nodes directly connected with the current node as neighbor nodes of the current node;
Determining the actual connection number between the neighbor node and the current node according to the neighbor node, calculating a clustering coefficient between the neighbor node and the current node, traversing the power grid, calculating the clustering coefficient of each node in the power grid and the average clustering coefficient of the power grid, calculating the local clustering coefficient of the current node through a preset local clustering algorithm, and fusing the characteristic information of the neighbor node and the path information between the nodes according to the local clustering coefficient to obtain a comprehensive topological vector.
In an alternative embodiment of the present invention,
The local clustering coefficient of the current node is calculated through a preset local clustering algorithm and is shown in the following formula:
；
Wherein C wi represents the local cluster coefficient of node i, Δijk represents the triangular relationship between nodes i, j and k, w ij represents the weight of edge ij, w ik represents the weight of edge ik, w jk represents the weight of edge jk, and f i represents the divergence of node i.
In an alternative embodiment of the present invention,
The method comprises the steps of encoding the comprehensive topological vector based on a preset graph neural network, carrying out node feature fusion on the traveling wave feature and the encoded comprehensive topological vector, generating initial node features, updating the initial node features through the graph neural network according to the node features and a attention mechanism introduced in advance, and generating input node features, wherein the step of generating the input node features comprises the following steps:
Acquiring the comprehensive topological vector, initializing node characteristics according to the comprehensive topological vector and the travelling wave characteristics, constructing an adjacent matrix of a current node, determining neighbor nodes of the current node, updating the current node according to the neighbor nodes, marking the current node as initial node characteristics, collecting node information of the neighbor nodes for each initial node characteristic, and calculating importance of each neighbor node to the current node, namely an attention coefficient, according to a multi-head attention mechanism introduced in advance;
And combining node information of each neighbor node, combining attention coefficients corresponding to each neighbor node, weighting and calculating comprehensive characteristics of the neighbor node through an activation function, and updating characteristic representation of the current node according to the comprehensive characteristics to generate input node characteristics.
In an alternative embodiment of the present invention,
And according to the input node characteristics, combining the traveling wave characteristics, performing fault positioning through a random forest module in the graph neural network, and determining the fault position comprises the following steps:
initializing a random forest module in the graph neural network according to the traveling wave characteristics, setting super parameters of the random forest module, namely the number and the depth of trees, and adding the input node characteristics into the random forest module as input information;
For each tree in the random forest module, randomly generating a training set, cross-verifying the performance of the random forest module under different numbers of trees according to the training set, selecting the number of the trees with the best performance, increasing the maximum depth for each tree, setting the minimum sample number of splitting each node in the tree according to the training set, simultaneously setting the minimum sample number of leaf nodes, observing the performance of the random forest module in the process of adjusting the super parameters, and adjusting the depth of the tree and the number of the trees in the random forest module according to the corresponding super parameters when the performance is not improved any more;
The random forest module sends the input node characteristics to each tree in the module, the tree generates a prediction fault probability corresponding to each node, the confidence coefficient of the prediction fault probability is judged by combining traveling wave information corresponding to the node, and if the confidence coefficient is larger than a preset confidence coefficient threshold value, the current node is considered to have a fault, and the fault position is acquired and returned.
In a second aspect of the embodiment of the present invention, a man-machine terminal picture data interaction system is provided, including:
The first unit is used for installing a traveling wave sensor, acquiring fault electromagnetic waves according to the traveling wave sensor, determining a time stamp corresponding to the fault electromagnetic waves and traveling wave propagation speed of the fault electromagnetic waves, and determining a fault area through a fault positioning algorithm according to the time stamp and the traveling wave propagation speed;
the second unit is used for acquiring traveling wave signal data of the fault electromagnetic wave according to the fault area, acquiring power grid topology data at the same time, extracting traveling wave characteristics in the traveling wave signal data and topology characteristics in the power grid topology data, determining structural characteristics according to the topology characteristics, identifying a shortest path of a node in a power grid topology graph, and carrying out characteristic fusion according to the shortest path of the node and a cluster coefficient introduced in advance to obtain a comprehensive topology vector;
And the third unit is used for encoding the comprehensive topological vector based on a preset graph neural network, carrying out node feature fusion on the traveling wave feature and the encoded comprehensive topological vector, generating initial node features, updating the initial node features through the graph neural network according to the initial node features and a attention mechanism introduced in advance, generating input node features, carrying out fault location through a random forest module in the graph neural network according to the input node features and the traveling wave features, and determining the fault position.
In a third aspect of an embodiment of the present invention,
There is provided an electronic device including:
A processor;
A memory for storing processor-executable instructions;
Wherein the processor is configured to invoke the instructions stored in the memory to perform the method described previously.
In a fourth aspect of an embodiment of the present invention,
There is provided a computer readable storage medium having stored thereon computer program instructions which, when executed by a processor, implement the method as described above.
According to the invention, the time stamp and traveling wave propagation speed of fault electromagnetic waves are obtained through the traveling wave sensor, the fault location algorithm is combined, the fault occurrence area can be precisely located, the fault location accuracy is improved, the structural characteristics are extracted by utilizing the power grid topology data, the shortest path and the clustering coefficient of the nodes in the power grid are beneficial to better capturing the structural information of the power system, more comprehensive system cognition is provided, the traveling wave characteristics and the coded comprehensive topological vector are fused by utilizing the graph neural network, the initial node characteristics are generated, the comprehensive grasp of the model on the node state is improved, more accurate input is provided for subsequent fault location, the attention mechanism is introduced to update the initial node characteristics, the attention degree of the model to the different nodes in the power grid is facilitated to be improved, and the fault location is more targeted.
Drawings
FIG. 1 is a flow chart of a method for interacting picture data of a man-machine terminal according to an embodiment of the invention;
Fig. 2 is a schematic structural diagram of a man-machine terminal picture data interaction system according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The technical scheme of the invention is described in detail below by specific examples. The following embodiments may be combined with each other, and some embodiments may not be repeated for the same or similar concepts or processes.
Fig. 1 is a flow chart of a method for interaction of picture data of a man-machine terminal according to an embodiment of the present invention, as shown in fig. 1, the method includes:
s1, installing a traveling wave sensor, acquiring fault electromagnetic waves according to the traveling wave sensor, determining a time stamp corresponding to the fault electromagnetic waves and traveling wave propagation speed of the fault electromagnetic waves, and determining a fault area through a fault positioning algorithm according to the time stamp and the traveling wave propagation speed;
The traveling wave sensor is a sensor for detecting faults in a power system, electromagnetic waves are generated at a fault point when faults occur in the power system, the traveling wave sensor can detect the electromagnetic waves and provide information about the faults, the fault electromagnetic waves are electromagnetic radiation generated when faults occur in the power system, the traveling wave propagation speed is the speed of the fault electromagnetic waves propagating in the power system, and the fault location algorithm is used for determining the fault position in the power system according to the electromagnetic wave information measured by the traveling wave sensor.
In an alternative embodiment of the present invention,
The step of installing the traveling wave sensor, the step of obtaining fault electromagnetic waves according to the traveling wave sensor, the step of determining a time stamp corresponding to the fault electromagnetic waves and traveling wave propagation speed of the fault electromagnetic waves, and the step of determining a fault area through a fault positioning algorithm according to the time stamp and the traveling wave propagation speed comprises the following steps:
installing a traveling wave sensor in a power line, acquiring fault electromagnetic waves caused by faults according to the traveling wave sensor, adding a time stamp to traveling wave detection data according to the time of acquiring the fault electromagnetic waves by the traveling wave sensor, determining the line length and the wire type in the power line, and determining the traveling wave propagation speed of the fault electromagnetic waves;
Judging the number of traveling wave sensors in a fault occurrence area according to the time stamp and the traveling wave propagation speed, if two or more traveling wave sensors exist, determining a fault area by comparing the arrival time of the waveform of the fault electromagnetic wave at different traveling wave sensors, and if only one traveling wave sensor exists, determining the fault area by reflecting the traveling wave information.
Installing a traveling wave sensor and ensuring that the traveling wave sensor can accurately capture fault electromagnetic waves, acquiring the fault electromagnetic waves through the traveling wave sensor, adding a time stamp for each data point according to the time of the fault electromagnetic waves reaching the traveling wave sensor, measuring the length of a line and determining the type of a wire according to the actual condition of a power line, and calculating the traveling wave propagation speed of the fault electromagnetic waves in the power line by using the known line length and the wire type;
and judging the number and positions of the traveling wave sensors which are required to be installed or are already installed in the fault occurrence area according to the layout of the sensors and the characteristics of the power line, if two or more traveling wave sensors exist, calculating the fault occurrence position by comparing the arrival time of fault electromagnetic waves on different sensors and using a time difference positioning method, and if only one traveling wave sensor exists, determining the fault area by analyzing and processing the reflected wave.
In this embodiment, fault electromagnetic wave data captured by each traveling wave sensor is monitored in real time, so that the system can quickly respond to a fault event, the reliability of the power system is improved, the graphical interface of the man-machine terminal is utilized to display the layout of a power line, the positions of sensors, the fault positioning result and the like in a graphical mode, the user can intuitively understand the position where the fault occurs, the cognitive load of the user is reduced, the user can interactively analyze the system, such as clicking a certain sensor point to check the detailed information of the point, including the fault electromagnetic wave waveform and the time stamp, so that the user can know the fault positioning process more deeply.
S2, acquiring traveling wave signal data of the fault electromagnetic wave according to the fault region, acquiring power grid topology data, extracting traveling wave characteristics in the traveling wave signal data and topology characteristics in the power grid topology data, determining structural characteristics according to the topology characteristics, identifying a shortest path of a node in a power grid topology graph, and performing characteristic fusion according to the shortest path of the node and a cluster coefficient introduced in advance to obtain a comprehensive topology vector;
The network topology data refers to information describing connection relations among elements (such as generators, substations, transmission lines and the like) in a power system, including connection modes and topological structures of nodes, branches and lines, and electrical connection relations among the nodes, the traveling wave features refer to properties and characteristics of electromagnetic waves propagating in the power system, the topological features describe structural connection relations of the power system, and generally refer to connection relations among the nodes and the branches, the structural features generally refer to characteristics of physical structures, connection modes, electrical parameters and the like among the description elements (such as substations, generators and the like) in the power system, the clustering coefficients are measures in network analysis and are used for describing the tightness degree among the nodes in the graph, the clustering coefficients of the nodes can represent the connection density between the nodes and the adjacent nodes in the power system, and the comprehensive topological vectors are feature vectors comprehensively considering the topological structures of the power system, and generally comprise combinations of a plurality of topological features such as node degree, the connection modes of the branches, the clustering coefficients and the like.
In an alternative embodiment of the present invention,
The step of obtaining traveling wave signal data of the fault electromagnetic wave according to the fault region, and simultaneously obtaining power grid topology data, wherein the step of extracting traveling wave characteristics in the traveling wave signal data and topology characteristics in the power grid topology data comprises the following steps:
Acquiring amplitude, frequency, energy distribution and phase angle of the fault electromagnetic wave according to the fault region, recording the amplitude, frequency, energy distribution and phase angle as traveling wave data signals, and acquiring power grid topology data comprising node information, connection relation and line parameters;
Preprocessing the traveling wave data signal, extracting amplitude, energy and duration in the traveling wave data signal, analyzing frequency components of the traveling wave data signal through Fourier transformation, and combining the extracted data into traveling wave characteristics;
and identifying each node in the power grid and the attribute corresponding to the node, determining a connection type and a connection parameter according to the connection relation of the nodes, and combining the connection relation, the connection type and the connection parameter into a topological feature.
Acquiring signals of fault electromagnetic waves by using a sensor according to the determined fault area, wherein the signals comprise information such as waveforms, voltages, currents and the like, and collecting power grid topology data, wherein the power grid topology data comprise node information (node numbers, types, coordinates and the like), connection relations (branch connection, node connection relations) and line parameters (impedance, length and the like);
preprocessing traveling wave data signals of fault electromagnetic waves, including filtering, noise reduction and trend removal, extracting characteristics such as amplitude, energy, duration and the like from the traveling wave data signals, analyzing frequency components by using Fourier transformation to obtain information such as frequency, frequency spectrum distribution and the like, and combining various extracted characteristics into a traveling wave characteristic vector, wherein the characteristic vector comprises a plurality of characteristic parameters such as amplitude, frequency, energy distribution, phase angle and the like;
And identifying each node and corresponding attribute information thereof according to the power grid topology data, determining the type (branch, transformer, switch and the like) and the connection parameters (impedance, admittance and the like) of connection according to the connection relation among the nodes, and combining the obtained connection relation, connection type and connection parameters into a topology feature vector, wherein the topology feature vector describes the topology structure of the power system.
In the embodiment, the information such as the connection relation between the nodes, the branch, the line parameters and the like is presented through the graphical interface, so that a user can clearly know the topological structure of the power system, the frequency components are visually presented on the man-machine terminal through the frequency spectrum analysis of the traveling wave data signals by the Fourier transform, the user can intuitively know the frequency spectrum distribution condition of the electromagnetic wave signals, the fault characteristics such as amplitude, frequency, energy distribution and phase angle are graphically presented on the man-machine terminal picture, the user can clearly see the change trend of the characteristics along with time, the rapid identification and the positioning of the fault are facilitated, and in conclusion, the operability and the user experience of the system are improved, so that the user can conveniently acquire, analyze and understand the state and the fault information of the power system on the graphical interface.
In an alternative embodiment of the present invention,
Determining structural features according to the topological features, identifying shortest paths of nodes in a power grid topological graph, performing feature fusion according to the shortest paths of the nodes and a cluster coefficient introduced in advance, and obtaining a comprehensive topological vector comprises the following steps:
according to the topological characteristics, for each node in the power grid, calculating the number of nodes directly connected with the current node, recording the number as the divergence of the current node, extracting node information directly connected with the current node according to the divergence, and recording the nodes directly connected with the current node as neighbor nodes of the current node;
Determining the actual connection number between the neighbor node and the current node according to the neighbor node, calculating a clustering coefficient between the neighbor node and the current node, traversing the power grid, calculating the clustering coefficient of each node in the power grid and the average clustering coefficient of the power grid, calculating the local clustering coefficient of the current node through a preset local clustering algorithm, and fusing the characteristic information of the neighbor node and the path information between the nodes according to the local clustering coefficient to obtain a comprehensive topological vector.
The divergence refers to the number of edges connected with the node, represents the connection degree of the node, is a basic topological feature of the node, the neighboring node refers to other nodes directly connected with the node, the average clustering coefficient is an average value of local clustering coefficients of all nodes in the graph, the degree of interconnection between neighbors of one node is measured, the local clustering coefficient refers to the ratio of the actual connection number between the neighbors of the node to the possible maximum connection number, and the tightness degree of sub-graphs around the node is reflected.
For each node in the power grid, calculating the number of nodes directly connected with the current node, marking the number as the divergence of the current node, extracting node information directly connected with the current node according to the divergence, and marking the node information as the neighbor node of the current node;
Traversing neighbor nodes of each node, determining the actual connection number between the neighbor nodes and the current node, determining the clustering coefficient between the neighbor nodes and the current node by calculating the ratio of the number of edges connected with each other and the possible maximum connection number of the neighbor nodes, traversing the whole power grid, calculating the clustering coefficient of each node, averaging the clustering coefficients to obtain the average clustering coefficient of the power grid, and calculating the local clustering coefficient of each node by using a preset local clustering algorithm, wherein the local clustering coefficient is an index related to the connection density of the neighbor nodes directly connected with the node, and fusing the divergence degree, the characteristic information of the neighbor nodes and the path information between the nodes according to the local clustering coefficient to obtain the comprehensive topological vector of each node.
In this embodiment, for each node, the man-machine terminal displays the neighbor node information directly connected with the man-machine terminal, so that the user can quickly understand the surrounding connection condition of the node, walk the neighbor nodes of each node, calculate the actual connection number and the clustering coefficient, display the calculation result on the man-machine terminal in a graphical manner, enable the user to see the clustering condition of each node in the power grid, fuse the divergence, the feature information of the neighbor nodes and the path information between the nodes to form a comprehensive topology vector, and display the topology feature of the node on the man-machine terminal in a graphical manner, so that the user can comprehensively understand the topology feature of the node.
In an alternative embodiment of the present invention,
The local clustering coefficient of the current node is calculated through a preset local clustering algorithm and is shown in the following formula:
；
Wherein C wi represents the local cluster coefficient of node i, Δijk represents the triangular relationship between nodes i, j and k, w ij represents the weight of edge ij, w ik represents the weight of edge ik, w jk represents the weight of edge jk, and f i represents the divergence of node i.
In the function, the local clustering coefficient calculation formula is applied to the man-machine terminal in real time, the system can dynamically calculate the local clustering coefficient of each node, real-time monitoring of the power grid topological dynamic characteristic is achieved, the local clustering coefficient is calculated more accurately by considering the weight of the edge, the actual connection strength between the node and the neighbor node can be better reflected, the local connection compactness of the node can be estimated more comprehensively by comprehensively considering the triangular relationship, the role of the node in the network can be understood, the divergence is introduced into the calculation of the local clustering coefficient, the calculation result not only considers the connection compactness, but also considers the diversity of the node in the whole network, and in the whole, the function enables a user to deeply understand the local connection characteristic of each node in the power grid, so that network topological analysis and fault diagnosis can be better carried out.
S3, coding the comprehensive topological vector based on a preset graph neural network, carrying out node feature fusion on the traveling wave feature and the coded comprehensive topological vector, generating initial node features, updating the initial node features through the graph neural network according to the initial node features and a attention mechanism introduced in advance, generating input node features, carrying out fault location through a random forest module in the graph neural network according to the input node features and the traveling wave features, and determining a fault position.
The graph neural network is a kind of neural network for processing graph data, the aim is to learn complex relations among nodes in a graph structure, so that effective representation learning can be carried out on characteristics of the nodes in the graph, the initial node characteristics are characteristics of each node in the graph, the attention mechanism is a method for learning attention degree among different parts by a model, in the graph neural network, the attention mechanism can be used for adjusting weight of information transmission among the nodes so that the network focuses on important nodes more, the input node characteristics are characteristics on which the nodes are updated during each round of iteration in the graph neural network, and the random forest is an integrated learning method and is usually composed of a plurality of decision trees.
In an alternative embodiment of the present invention,
The method comprises the steps of encoding the comprehensive topological vector based on a preset graph neural network, carrying out node feature fusion on the traveling wave feature and the encoded comprehensive topological vector, generating initial node features, updating the initial node features through the graph neural network according to the initial node features and a attention mechanism introduced in advance, and generating input node features, wherein the steps comprise:
Acquiring the comprehensive topological vector, initializing node characteristics according to the comprehensive topological vector and the travelling wave characteristics, constructing an adjacent matrix of a current node, determining neighbor nodes of the current node, updating the current node according to the neighbor nodes, marking the current node as initial node characteristics, collecting node information of the neighbor nodes for each initial node characteristic, and calculating importance of each neighbor node to the current node, namely an attention coefficient, according to a multi-head attention mechanism introduced in advance;
And combining node information of each neighbor node, combining attention coefficients corresponding to each neighbor node, weighting and calculating comprehensive characteristics of the neighbor node through an activation function, and updating characteristic representation of the current node according to the comprehensive characteristics to generate input node characteristics.
Acquiring the comprehensive topological vector and the traveling wave feature obtained by calculation from the previous steps, initializing the feature representation of each node by using the comprehensive topological vector and the traveling wave feature, constructing an adjacent matrix of the current node, determining the neighbor node of the current node, collecting the node information of the neighbor node of each initial node feature, including the current feature representation of the neighbor node and the traveling wave feature, introducing a predefined multi-head attention mechanism, and determining the importance of each neighbor node to the current node, namely the attention coefficient, by calculating the attention score and applying a softmax function;
Combining node information of each neighbor node, combining attention coefficients corresponding to each neighbor node, calculating comprehensive characteristics of the neighbor nodes through activating function weighting, and generating input node characteristics by fusing the comprehensive characteristics with original characteristic representations of the current nodes according to the comprehensive characteristics, for example, updating the characteristic representations of the current nodes by residual connection.
In this embodiment, the operation of the multi-head attention mechanism is graphically displayed on the man-machine terminal, including the process of calculating the attention coefficient, the user can know the importance of each neighboring node to the current node and the change condition of the attention coefficient through an interactive mode, the collection of the neighboring node information and the calculation process of the comprehensive characteristics are presented through the graphical interface, the user can intuitively know how the neighboring node information is integrated into the characteristic representation of the current node, and the contribution degree of each neighboring node to the current node.
In an alternative embodiment of the present invention,
And according to the input node characteristics, combining the traveling wave characteristics, performing fault positioning through a random forest module in the graph neural network, and determining the fault position comprises the following steps:
initializing a random forest module in the graph neural network according to the traveling wave characteristics, setting super parameters of the random forest module, namely the number and the depth of trees, and adding the input node characteristics into the random forest module as input information;
For each tree in the random forest module, randomly generating a training set, cross-verifying the performance of the random forest module under different numbers of trees according to the training set, selecting the number of the trees with the best performance, increasing the maximum depth for each tree, setting the minimum sample number of splitting each node in the tree according to the training set, simultaneously setting the minimum sample number of leaf nodes, observing the performance of the random forest module in the process of adjusting the super parameters, and adjusting the depth of the tree and the number of the trees in the random forest module according to the corresponding super parameters when the performance is not improved any more;
The random forest module sends the input node characteristics to each tree in the module, the tree generates a prediction fault probability corresponding to each node, the confidence coefficient of the prediction fault probability is judged by combining traveling wave information corresponding to the node, and if the confidence coefficient is larger than a preset confidence coefficient threshold value, the current node is considered to have a fault, and the fault position is acquired and returned.
The super-parameters are parameters set before model training, the confidence degree refers to the confidence degree or confidence level of a model for a certain predicted result, the confidence degree is expressed in a probability form, the confidence threshold is a threshold in a two-class or multi-class problem and is used for determining whether the predicted result output by the model is accepted, and if the confidence degree of the model output is greater than or equal to the threshold, the model output is judged to be a positive class; otherwise, the negative category is determined.
And initializing a random forest module in the graph neural network by using the travelling wave characteristics as input. Setting super parameters of the random forest module, including the number and depth of trees, and adding the input node characteristics into the random forest module as information;
For each tree in the random forest module, randomly generating a training set, carrying out cross verification under the number of different trees to select the number of the trees with optimal performance, gradually increasing the maximum depth for each tree, setting the minimum sample number of each node in the tree for splitting according to the training set, simultaneously setting the minimum sample number of leaf nodes, observing the performance of the random forest module in the process of adjusting the super parameters, and adjusting the depth of the tree and the number of the trees in the random forest module according to the corresponding super parameters when the performance is not improved any more;
the random forest module sends the input node characteristics to each tree, each tree generates a predicted fault probability corresponding to each node, and the confidence level of the predicted fault probability is judged by combining traveling wave information corresponding to the nodes. If the confidence coefficient is larger than a preset confidence coefficient threshold value, the current node is considered to have a fault, and when the fault is judged to exist, the fault position is obtained according to the prediction result of the tree in the random forest module, and the result is returned.
In summary, the embodiment displays the performance index of the random forest module on the interface in real time, such as accuracy, precision, recall rate and the like, along with the adjustment of the super parameters, the user can observe the change of the performance in real time, help the user to make a more intelligent parameter adjustment decision, display the variation trend of the model performance under different numbers and depths of the trees through a chart or curve display super parameter adjustment process, help the user to better understand the influence of super parameter selection on the model, display the prediction fault probability and corresponding confidence coefficient of each node on the interface in real time, and help the user to intuitively know which nodes possibly have faults and the confidence coefficient through a thermal diagram, a chart and the like.
Fig. 2 is a schematic structural diagram of a man-machine terminal picture data interaction system according to an embodiment of the present invention, as shown in fig. 2, the system includes:
The first unit is used for installing a traveling wave sensor, acquiring fault electromagnetic waves according to the traveling wave sensor, determining a time stamp corresponding to the fault electromagnetic waves and traveling wave propagation speed of the fault electromagnetic waves, and determining a fault area through a fault positioning algorithm according to the time stamp and the traveling wave propagation speed;
the second unit is used for acquiring traveling wave signal data of the fault electromagnetic wave according to the fault area, acquiring power grid topology data at the same time, extracting traveling wave characteristics in the traveling wave signal data and topology characteristics in the power grid topology data, determining structural characteristics according to the topology characteristics, identifying a shortest path of a node in a power grid topology graph, and carrying out characteristic fusion according to the shortest path of the node and a cluster coefficient introduced in advance to obtain a comprehensive topology vector;
And the third unit is used for encoding the comprehensive topological vector based on a preset graph neural network, carrying out node feature fusion on the traveling wave feature and the encoded comprehensive topological vector, generating initial node features, updating the initial node features through the graph neural network according to the initial node features and a attention mechanism introduced in advance, generating input node features, carrying out fault location through a random forest module in the graph neural network according to the input node features and the traveling wave features, and determining the fault position.
In a third aspect of an embodiment of the present invention,
There is provided an electronic device including:
A processor;
A memory for storing processor-executable instructions;
Wherein the processor is configured to invoke the instructions stored in the memory to perform the method described previously.
In a fourth aspect of an embodiment of the present invention,
There is provided a computer readable storage medium having stored thereon computer program instructions which, when executed by a processor, implement the method as described above.
The present invention may be a method, apparatus, system, and/or computer program product. The computer program product may include a computer readable storage medium having computer readable program instructions embodied thereon for performing various aspects of the present invention.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention.
Claims (6)
1. The man-machine terminal picture data interaction method is characterized by comprising the following steps of:
Installing a traveling wave sensor, acquiring fault electromagnetic waves through the traveling wave sensor, determining a time stamp corresponding to the fault electromagnetic waves and traveling wave propagation speed of the fault electromagnetic waves, and determining a fault area through a fault positioning algorithm according to the time stamp and the traveling wave propagation speed;
Acquiring traveling wave signal data of the fault electromagnetic wave according to the fault region, acquiring power grid topology data, extracting traveling wave characteristics in the traveling wave signal data and topology characteristics in the power grid topology data, determining structural characteristics according to the topology characteristics, identifying a shortest path of a node in a power grid topology graph, and performing characteristic fusion according to the shortest path of the node and a cluster coefficient introduced in advance to obtain a comprehensive topology vector;
Coding the comprehensive topological vector based on a preset graph neural network, carrying out node feature fusion on the traveling wave feature and the coded comprehensive topological vector, generating initial node features, updating the initial node features through the graph neural network according to the initial node features and a attention mechanism introduced in advance, generating input node features, carrying out fault location through a random forest module in the graph neural network according to the input node features and the traveling wave features, and determining a fault position;
Determining structural features according to the topological features, identifying shortest paths of nodes in a power grid topological graph, performing feature fusion according to the shortest paths of the nodes and a cluster coefficient introduced in advance, and obtaining a comprehensive topological vector comprises the following steps:
according to the topological characteristics, for each node in the power grid, calculating the number of nodes directly connected with the current node, recording the number as the divergence of the current node, extracting node information directly connected with the current node according to the divergence, and recording the nodes directly connected with the current node as neighbor nodes of the current node;
Determining the actual connection number between the neighbor node and the current node according to the neighbor node, calculating a clustering coefficient between the neighbor node and the current node, traversing the power grid, calculating a clustering coefficient of each node in the power grid and an average clustering coefficient of the power grid, calculating a local clustering coefficient of the current node through a preset local clustering algorithm, and fusing the characteristic information of the neighbor node and path information between the nodes according to the local clustering coefficient to obtain a comprehensive topological vector;
The local clustering coefficient of the current node is calculated through a preset local clustering algorithm and is shown in the following formula:
；
Wherein C wi represents the local cluster coefficient of node i, Δijk represents the triangular relationship between nodes i, j and k, w ij represents the weight of edge ij, w ik represents the weight of edge ik, w jk represents the weight of edge jk, and f i represents the divergence of node i;
The method comprises the steps of encoding the comprehensive topological vector based on a preset graph neural network, carrying out node feature fusion on the traveling wave feature and the encoded comprehensive topological vector, generating initial node features, updating the initial node features through the graph neural network according to the initial node features and a attention mechanism introduced in advance, and generating input node features, wherein the steps comprise:
Acquiring the comprehensive topological vector, initializing node characteristics according to the comprehensive topological vector and the travelling wave characteristics, constructing an adjacent matrix of a current node, determining neighbor nodes of the current node, updating the current node according to the neighbor nodes, marking the current node as initial node characteristics, collecting node information of the neighbor nodes for each initial node characteristic, and calculating importance of each neighbor node to the current node, namely an attention coefficient, according to a multi-head attention mechanism introduced in advance;
combining node information of each neighbor node, weighting and calculating comprehensive characteristics of the neighbor node through an activation function by combining attention coefficients corresponding to each neighbor node, and updating characteristic representation of the current node according to the comprehensive characteristics to generate input node characteristics;
And according to the input node characteristics, combining the traveling wave characteristics, performing fault positioning through a random forest module in the graph neural network, and determining the fault position comprises the following steps:
initializing a random forest module in the graph neural network according to the traveling wave characteristics, setting super parameters of the random forest module, namely the number and the depth of trees, and adding the input node characteristics into the random forest module as input information;
For each tree in the random forest module, randomly generating a training set, cross-verifying the performance of the random forest module under different numbers of trees according to the training set, selecting the number of the trees with the best performance, increasing the maximum depth for each tree, setting the minimum sample number of splitting each node in the tree according to the training set, simultaneously setting the minimum sample number of leaf nodes, observing the performance of the random forest module in the process of adjusting the super parameters, and adjusting the depth of the tree and the number of the trees in the random forest module according to the corresponding super parameters when the performance is not improved any more;
The random forest module sends the input node characteristics to each tree in the module, the tree generates a prediction fault probability corresponding to each node, the confidence coefficient of the prediction fault probability is judged by combining traveling wave information corresponding to the node, and if the confidence coefficient is larger than a preset confidence coefficient threshold value, the current node is considered to have a fault, and the fault position is acquired and returned.
2. The method according to claim 1, wherein the obtaining the fault electromagnetic wave by the traveling wave sensor, determining a time stamp corresponding to the fault electromagnetic wave and a traveling wave propagation speed of the fault electromagnetic wave, and determining the fault area by the fault location algorithm according to the time stamp and the traveling wave propagation speed comprises:
Acquiring fault electromagnetic waves caused by faults according to the traveling wave sensor installed in a power line, adding a time stamp to traveling wave detection data according to the time of acquiring the fault electromagnetic waves by the traveling wave sensor, determining the line length and the wire type in the power line, and determining the traveling wave propagation speed of the fault electromagnetic waves;
judging the number of the traveling wave sensors in the fault occurrence area according to the time stamp and the traveling wave propagation speed, if two or more traveling wave sensors exist, determining the fault area by comparing the arrival time of the waveform of the fault electromagnetic wave at different traveling wave sensors, and if only one traveling wave sensor exists, determining the fault area by the information of the reflected traveling wave.
3. The method according to claim 1, wherein the acquiring traveling wave signal data of the fault electromagnetic wave according to the fault region, and simultaneously acquiring power grid topology data, and extracting traveling wave features in the traveling wave signal data and topology features in the power grid topology data includes:
Acquiring amplitude, frequency, energy distribution and phase angle of the fault electromagnetic wave according to the fault region, recording the amplitude, frequency, energy distribution and phase angle as traveling wave data signals, and acquiring power grid topology data comprising node information, connection relation and line parameters;
Preprocessing the traveling wave data signal, extracting amplitude, energy and duration in the traveling wave data signal, analyzing frequency components of the traveling wave data signal through Fourier transformation, and combining the extracted data into traveling wave characteristics;
and identifying each node in the power grid and the attribute corresponding to the node, determining a connection type and a connection parameter according to the connection relation of the nodes, and combining the connection relation, the connection type and the connection parameter into a topological feature.
4. The man-machine terminal picture data interaction system is characterized by comprising:
The first unit is used for installing a traveling wave sensor, acquiring fault electromagnetic waves through the traveling wave sensor, determining a time stamp corresponding to the fault electromagnetic waves and traveling wave propagation speed of the fault electromagnetic waves, and determining a fault area through a fault positioning algorithm according to the time stamp and the traveling wave propagation speed;
the second unit is used for acquiring traveling wave signal data of the fault electromagnetic wave according to the fault area, acquiring power grid topology data at the same time, extracting traveling wave characteristics in the traveling wave signal data and topology characteristics in the power grid topology data, determining structural characteristics according to the topology characteristics, identifying a shortest path of a node in a power grid topology graph, and carrying out characteristic fusion according to the shortest path of the node and a cluster coefficient introduced in advance to obtain a comprehensive topology vector;
The third unit is used for encoding the comprehensive topological vector based on a preset graph neural network, carrying out node feature fusion on the traveling wave feature and the encoded comprehensive topological vector, generating initial node features, updating the initial node features through the graph neural network according to the initial node features and a attention mechanism introduced in advance, generating input node features, carrying out fault location through a random forest module in the graph neural network according to the input node features and the traveling wave features, and determining a fault position;
Determining structural features according to the topological features, identifying shortest paths of nodes in a power grid topological graph, performing feature fusion according to the shortest paths of the nodes and a cluster coefficient introduced in advance, and obtaining a comprehensive topological vector comprises the following steps:
According to the topological characteristics, constructing a power grid topological graph, calculating the number of nodes directly connected with the current node for each node in the power grid, recording the number as the divergence of the current node, extracting node information directly connected with the current node according to the divergence, and recording the nodes directly connected with the current node as neighbor nodes of the current node;
Determining the actual connection number between the neighbor node and the current node according to the neighbor node, calculating a clustering coefficient between the neighbor node and the current node, traversing the power grid, calculating a clustering coefficient of each node in the power grid and an average clustering coefficient of the power grid, calculating a local clustering coefficient of the current node through a preset local clustering algorithm, and fusing the characteristic information of the neighbor node and path information between the nodes according to the local clustering coefficient to obtain a comprehensive topological vector;
The local clustering coefficient of the current node is calculated through a preset local clustering algorithm and is shown in the following formula:
；
Wherein C wi represents the local cluster coefficient of node i, Δijk represents the triangular relationship between nodes i, j and k, w ij represents the weight of edge ij, w ik represents the weight of edge ik, w jk represents the weight of edge jk, and f i represents the divergence of node i;
The method comprises the steps of encoding the comprehensive topological vector based on a preset graph neural network, carrying out node feature fusion on the traveling wave feature and the encoded comprehensive topological vector, generating initial node features, updating the initial node features through the graph neural network according to the initial node features and a attention mechanism introduced in advance, and generating input node features, wherein the steps comprise:
Acquiring the comprehensive topological vector, initializing node characteristics according to the comprehensive topological vector and the travelling wave characteristics, constructing an adjacent matrix of a current node, determining neighbor nodes of the current node, updating the current node according to the neighbor nodes, marking the current node as initial node characteristics, collecting node information of the neighbor nodes for each initial node characteristic, and calculating importance of each neighbor node to the current node, namely an attention coefficient, according to a multi-head attention mechanism introduced in advance;
combining node information of each neighbor node, weighting and calculating comprehensive characteristics of the neighbor node through an activation function by combining attention coefficients corresponding to each neighbor node, and updating characteristic representation of the current node according to the comprehensive characteristics to generate input node characteristics;
And according to the input node characteristics, combining the traveling wave characteristics, performing fault positioning through a random forest module in the graph neural network, and determining the fault position comprises the following steps:
initializing a random forest module in the graph neural network according to the traveling wave characteristics, setting super parameters of the random forest module, namely the number and the depth of trees, and adding the input node characteristics into the random forest module as input information;
For each tree in the random forest module, randomly generating a training set, cross-verifying the performance of the random forest module under different numbers of trees according to the training set, selecting the number of the trees with the best performance, increasing the maximum depth for each tree, setting the minimum sample number of splitting each node in the tree according to the training set, simultaneously setting the minimum sample number of leaf nodes, observing the performance of the random forest module in the process of adjusting the super parameters, and adjusting the depth of the tree and the number of the trees in the random forest module according to the corresponding super parameters when the performance is not improved any more;
The random forest module sends the input node characteristics to each tree in the module, the tree generates a prediction fault probability corresponding to each node, the confidence coefficient of the prediction fault probability is judged by combining traveling wave information corresponding to the node, and if the confidence coefficient is larger than a preset confidence coefficient threshold value, the current node is considered to have a fault, and the fault position is acquired and returned.
5. An electronic device, comprising:
A processor;
A memory for storing processor-executable instructions;
wherein the processor is configured to invoke the instructions stored in the memory to perform the method of any of claims 1 to 3.
6. A computer readable storage medium having stored thereon computer program instructions, which when executed by a processor, implement the method of any of claims 1 to 3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311770009.6A CN117454315B (en) | 2023-12-21 | 2023-12-21 | Man-machine terminal picture data interaction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311770009.6A CN117454315B (en) | 2023-12-21 | 2023-12-21 | Man-machine terminal picture data interaction method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117454315A CN117454315A (en) | 2024-01-26 |
| CN117454315B true CN117454315B (en) | 2024-05-28 |
Family
ID=89585872
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311770009.6A Active CN117454315B (en) | 2023-12-21 | 2023-12-21 | Man-machine terminal picture data interaction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117454315B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118473910B (en) * | 2024-07-08 | 2024-09-10 | 鄂尔多斯市泛胜数据技术有限公司 | Electric power Internet of things fault detection method and system based on edge cloud cooperation |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6430150B1 (en) * | 1996-02-14 | 2002-08-06 | Fujitsu Limited | Communication node, restoration method and communication network |
| WO2017196821A1 (en) * | 2016-05-09 | 2017-11-16 | Strong Force Iot Portfolio 2016, Llc | Methods and systems for the industrial internet of things |
| WO2019232595A1 (en) * | 2018-06-07 | 2019-12-12 | Federation University Australia | A method of estimating the location of a fault on an electrical distribution network and an associated system |
| CN113835000A (en) * | 2021-09-23 | 2021-12-24 | 南方电网科学研究院有限责任公司 | Power distribution network fault positioning method and device, terminal and storage medium |
| CN113945799A (en) * | 2021-10-15 | 2022-01-18 | 广东电网有限责任公司惠州供电局 | Power line network fault positioning method and device, electronic equipment and storage medium |
| CN114636894A (en) * | 2022-03-09 | 2022-06-17 | 四川思极科技有限公司 | Power distribution network topology change identification method for optimizing traveling wave positioning |
| CN114675134A (en) * | 2022-04-18 | 2022-06-28 | 四川思极科技有限公司 | Power distribution network fault positioning method and system based on traveling wave space-time matrix |
| CN114839473A (en) * | 2022-03-31 | 2022-08-02 | 广西电网有限责任公司电力科学研究院 | Active power distribution network fault positioning method based on model migration diagram self-encoder network |
| CN115730261A (en) * | 2022-11-21 | 2023-03-03 | 山东博鸿电气股份有限公司 | Power distribution network line fault detection method and system based on traveling waves |
| CN115954865A (en) * | 2022-12-09 | 2023-04-11 | 广西电网有限责任公司 | Power distribution network low-voltage topology analysis method and system based on artificial intelligence |
| CN116243110A (en) * | 2023-04-14 | 2023-06-09 | 长沙理工大学 | Primary power distribution network fault identification and positioning method and related equipment |
| CN116822334A (en) * | 2023-05-30 | 2023-09-29 | 贵州黔驰信息股份有限公司 | Visual power grid model fault response method and system |
| CN116973694A (en) * | 2023-09-22 | 2023-10-31 | 国网浙江宁波市鄞州区供电有限公司 | Power distribution network fault diagnosis optimization method and system |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110927520A (en) * | 2019-11-25 | 2020-03-27 | 山东理工大学 | Direct-current distribution line multi-end traveling wave fault positioning method and positioning device |
| EP3955012B1 (en) * | 2020-08-13 | 2024-09-25 | Siemens Aktiengesellschaft | Method and device for determining the location of a fault on a line of an electrical energy supply network |
-
2023
- 2023-12-21 CN CN202311770009.6A patent/CN117454315B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6430150B1 (en) * | 1996-02-14 | 2002-08-06 | Fujitsu Limited | Communication node, restoration method and communication network |
| WO2017196821A1 (en) * | 2016-05-09 | 2017-11-16 | Strong Force Iot Portfolio 2016, Llc | Methods and systems for the industrial internet of things |
| WO2019232595A1 (en) * | 2018-06-07 | 2019-12-12 | Federation University Australia | A method of estimating the location of a fault on an electrical distribution network and an associated system |
| CN113835000A (en) * | 2021-09-23 | 2021-12-24 | 南方电网科学研究院有限责任公司 | Power distribution network fault positioning method and device, terminal and storage medium |
| CN113945799A (en) * | 2021-10-15 | 2022-01-18 | 广东电网有限责任公司惠州供电局 | Power line network fault positioning method and device, electronic equipment and storage medium |
| CN114636894A (en) * | 2022-03-09 | 2022-06-17 | 四川思极科技有限公司 | Power distribution network topology change identification method for optimizing traveling wave positioning |
| CN114839473A (en) * | 2022-03-31 | 2022-08-02 | 广西电网有限责任公司电力科学研究院 | Active power distribution network fault positioning method based on model migration diagram self-encoder network |
| CN114675134A (en) * | 2022-04-18 | 2022-06-28 | 四川思极科技有限公司 | Power distribution network fault positioning method and system based on traveling wave space-time matrix |
| CN115730261A (en) * | 2022-11-21 | 2023-03-03 | 山东博鸿电气股份有限公司 | Power distribution network line fault detection method and system based on traveling waves |
| CN115954865A (en) * | 2022-12-09 | 2023-04-11 | 广西电网有限责任公司 | Power distribution network low-voltage topology analysis method and system based on artificial intelligence |
| CN116243110A (en) * | 2023-04-14 | 2023-06-09 | 长沙理工大学 | Primary power distribution network fault identification and positioning method and related equipment |
| CN116822334A (en) * | 2023-05-30 | 2023-09-29 | 贵州黔驰信息股份有限公司 | Visual power grid model fault response method and system |
| CN116973694A (en) * | 2023-09-22 | 2023-10-31 | 国网浙江宁波市鄞州区供电有限公司 | Power distribution network fault diagnosis optimization method and system |
Non-Patent Citations (5)
| Title |
|---|
| Application of Transient Traveling Wave Technology in Fault Location and Condition Monitoring of Collecting Line in Wind Farm;Shihao Zhang 等;《2023 IEEE 3rd International Conference on Power, Electronics and Computer Applications (ICPECA)》;20230329;全文 * |
| Fault location method for multi-segment hybrid line based on traveling wave time difference;Weihua Sun 等;《2022 4th International Conference on Electrical Engineering and Control Technologies》;20230207;全文 * |
| 广域行波信息与图注意力网络相结合的 输电网故障定位;张翼等;《仪器仪表学报》;20220630;全文 * |
| 考虑电网拓扑结构的行波故障定位方法;曲广龙;杨洪耕;吴晓清;周辉;;电力系统及其自动化学报;20131215(06);全文 * |
| 面向复杂信息系统的多源异构数据融合技术;陈日成;金涛;;中国测试;20200731(07);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117454315A (en) | 2024-01-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106443310B (en) | A kind of transformer fault detection method based on SOM neural network | |
| CN117454315B (en) | Man-machine terminal picture data interaction method and system | |
| EP3460496B1 (en) | A method and apparatus for automatic localization of a fault | |
| CN113049084B (en) | Attention mechanism-based Resnet distributed optical fiber sensing signal identification method | |
| JP6738135B2 (en) | How to perform electrical cable fault detection on a computer | |
| CN111652496A (en) | Operation risk assessment method and device based on network security situation awareness system | |
| CA2285239A1 (en) | System and method for telecommunications system fault diagnostics | |
| CN113486337B (en) | Network security situation element identification system and method based on particle swarm optimization | |
| CN117668751B (en) | High-low voltage power system fault diagnosis method and device | |
| CN111586728B (en) | Small sample characteristic-oriented heterogeneous wireless network fault detection and diagnosis method | |
| CN110702966A (en) | Fault arc detection method, device and system based on probabilistic neural network | |
| CN113111731A (en) | Deep neural network black box countermeasure sample generation method and system based on channel measurement information | |
| CN118214502A (en) | Digital broadcast signal quality real-time monitoring method and system | |
| CN111951505B (en) | Fence vibration intrusion positioning and mode identification method based on distributed optical fiber system | |
| CN116699400A (en) | Generator rotor short-circuit fault monitoring system, method and readable storage medium | |
| CN116484746A (en) | Digital twin system, method and medium for identifying state of operating mechanism of circuit breaker | |
| Guo et al. | Application of machine learning in wire damage detection for safety procedure | |
| Ishak et al. | Detection of Power Distribution Fault in Thermal Images Using CNN | |
| CN118316797B (en) | Automatic networking method and system for remote Internet of things equipment | |
| CN118381683B (en) | Distributed monitoring method and device for industrial control network attack event | |
| CN107578170A (en) | A kind of fire-fighting system safety evaluation method based on data characteristics selection | |
| CN114169545B (en) | Method, device, medium and equipment for intelligent fault diagnosis and operation and maintenance guide of thermal power plant equipment | |
| CN117648632B (en) | Method, device, equipment and computer program product for identifying optical fiber vibration abnormality | |
| CN118566637A (en) | Fault positioning system based on power distribution network topology | |
| Al-Kasassbeh | Change detection methods for computer network problems |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |