CN116523001A - Method, device and computer equipment for constructing weak line identification model of power grid - Google Patents
Method, device and computer equipment for constructing weak line identification model of power grid Download PDFInfo
- Publication number
- CN116523001A CN116523001A CN202310504717.9A CN202310504717A CN116523001A CN 116523001 A CN116523001 A CN 116523001A CN 202310504717 A CN202310504717 A CN 202310504717A CN 116523001 A CN116523001 A CN 116523001A
- Authority
- CN
- China
- Prior art keywords
- power grid
- weak line
- data
- historical
- graph data
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 90
- 238000012549 training Methods 0.000 claims abstract description 49
- 230000002776 aggregation Effects 0.000 claims abstract description 40
- 238000004220 aggregation Methods 0.000 claims abstract description 40
- 238000004590 computer program Methods 0.000 claims abstract description 40
- 238000004458 analytical method Methods 0.000 claims abstract description 27
- 230000008569 process Effects 0.000 claims abstract description 25
- 230000006870 function Effects 0.000 claims description 42
- 238000011156 evaluation Methods 0.000 claims description 40
- 238000004088 simulation Methods 0.000 claims description 30
- 238000004422 calculation algorithm Methods 0.000 claims description 23
- 238000010276 construction Methods 0.000 claims description 17
- 238000005457 optimization Methods 0.000 claims description 17
- 238000012545 processing Methods 0.000 claims description 13
- 238000012216 screening Methods 0.000 claims description 9
- 238000012937 correction Methods 0.000 abstract description 11
- 239000003795 chemical substances by application Substances 0.000 description 33
- 238000010586 diagram Methods 0.000 description 9
- 239000013598 vector Substances 0.000 description 7
- 238000004364 calculation method Methods 0.000 description 6
- 230000007246 mechanism Effects 0.000 description 6
- 238000013500 data storage Methods 0.000 description 5
- 238000000605 extraction Methods 0.000 description 5
- 239000010410 layer Substances 0.000 description 3
- 238000011002 quantification Methods 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 230000010365 information processing Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000012502 risk assessment Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 241000282461 Canis lupus Species 0.000 description 1
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 230000004931 aggregating effect Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000001149 cognitive effect Effects 0.000 description 1
- 230000010485 coping Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000005251 gamma ray Effects 0.000 description 1
- 229910021389 graphene Inorganic materials 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 239000002356 single layer Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- 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
- G06N3/0475—Generative 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
- G06N3/094—Adversarial learning
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- Business, Economics & Management (AREA)
- Economics (AREA)
- Public Health (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)
- Supply And Distribution Of Alternating Current (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application relates to a method, a device, computer equipment, a storage medium and a computer program product for constructing a weak line identification model of a power grid, wherein the method establishes power grid graph data based on historical power grid operation state data and historical topological information through a graph representation method; carrying out multi-head attention information aggregation analysis on the power grid graph data to obtain multi-head attention information of the power grid graph data; and performing key operation feature identification to determine key operation features of the power grid; finally, training the initial intelligent agent model to obtain a weak line identification model of the power grid, identifying key features at an online operation stage to identify the weak line effectively, constructing a loss function of graph attention network training by using a cross entropy function with weights in the model training process, setting a weight correction coefficient for the line with historical risk or fault event, and effectively improving the identification efficiency of the weak line identification model of the power grid to the weak line on the premise of ensuring the identification accuracy.
Description
Technical Field
The present invention relates to the field of computer technologies, and in particular, to a method, an apparatus, a computer device, a storage medium, and a computer program product for constructing a weak line identification model of a power grid.
Background
With frequent occurrence of extreme weather events such as typhoons, cold tides and the like, the operation safety and the power supply stability of the power system are seriously threatened. Meanwhile, high-permeability new energy and massive flexible load access bring more uncertainty and instability to power grid operation and bring difficulty to power grid dispatching optimization solution. In the extreme weather disaster-causing process, accurate risk evolution sensing and evaluation are required to be carried out, and weak links of power grid operation are accurately excavated.
The current power grid risk assessment and weak link identification method is mainly based on power grid real-time power flow and risk index calculation, is long in calculation time and low in efficiency, and cannot meet the real-time requirement of disaster risk assessment.
Disclosure of Invention
In view of the foregoing, it is desirable to provide a method, an apparatus, a computer device, a computer readable storage medium, and a computer program product for constructing a weak line identification model of a power grid, which can ensure the weak line identification efficiency of the power grid.
In a first aspect, the application provides a method for constructing a weak line identification model of a power grid. The method comprises the following steps:
establishing power grid graph data based on historical power grid running state data and historical topological information through a graph representation method;
performing multi-head attention information aggregation analysis on the power grid graph data to obtain multi-head attention information of the power grid graph data;
performing key operation feature identification based on the power grid graph data and the multi-head attention information, and determining power grid operation key features;
training an initial intelligent agent model based on the power grid graph data and the power grid operation key characteristics to obtain a power grid weak line identification model, wherein a loss function of the initial intelligent agent model is constructed by a cross entropy function with a weight coefficient, and the weight coefficient is set based on a line with historical risk or fault event.
In one embodiment, the performing multi-head attention information aggregation analysis on the grid graph data to obtain multi-head attention information of the grid graph data includes:
determining the attention coefficients of the lines and adjacent information nodes in the power grid graph data based on the topology information and the attribute characteristic information of the power grid graph data;
And carrying out node information aggregation processing on the power grid graph data based on the attention coefficient to obtain multi-head attention information of the power grid graph data.
In one embodiment, the method further comprises:
and in the training process of the initial intelligent agent model, node characteristic screening and optimization are carried out on the weak line identification model of the power grid through a back propagation optimizer.
In one embodiment, the method further comprises:
acquiring historical operation section data;
performing simulation based on the historical operation section data to obtain a simulation result corresponding to the historical operation section data;
generating a power grid running state data sample through a generation type countermeasure network based on the simulation result;
and constructing a sample set of the historical power grid operation state data based on the power grid operation state data sample.
In one embodiment, the method further comprises:
acquiring real-time running state data of a power grid;
and inputting the real-time running state data of the power grid into the power grid weak line identification model to determine the power grid weak line.
In one embodiment, the method further comprises:
determining index values of the weak line of the power grid in each weak degree evaluation index;
Performing initial weighting treatment on the weakness evaluation indexes through a membership rule to obtain initial weights of the weakness evaluation indexes;
performing self-adaptive adjustment on the initial weight through a heuristic algorithm to obtain an index weight;
and determining the weakness parameters of the weak line of the power grid based on the index values of the weak line of the power grid in each weakness evaluation index and the index weights.
In a second aspect, the application also provides a weak line identification model construction device for the power grid. The device comprises:
the data construction module is used for establishing power grid graph data based on historical power grid running state data and historical topological information through a graph representation method;
the aggregation analysis module is used for carrying out multi-head attention information aggregation analysis on the power grid graph data to obtain multi-head attention information of the power grid graph data;
the feature recognition module is used for carrying out key operation feature recognition based on the power grid graph data and the multi-head attention information and determining power grid operation key features;
the model construction module is used for training an initial intelligent agent model based on the power grid graph data and the power grid operation key characteristics to obtain a power grid weak line identification model, a loss function of the initial intelligent agent model is constructed by a cross entropy function with a weight coefficient, and the weight coefficient is set based on a line with historical risk or fault event.
In a third aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor which when executing the computer program performs the steps of:
establishing power grid graph data based on historical power grid running state data and historical topological information through a graph representation method;
performing multi-head attention information aggregation analysis on the power grid graph data to obtain multi-head attention information of the power grid graph data;
performing key operation feature identification based on the power grid graph data and the multi-head attention information, and determining power grid operation key features;
training an initial intelligent agent model based on the power grid graph data and the power grid operation key characteristics to obtain a power grid weak line identification model, wherein a loss function of the initial intelligent agent model is constructed by a cross entropy function with a weight coefficient, and the weight coefficient is set based on a line with historical risk or fault event.
In a fourth aspect, the present application also provides a computer-readable storage medium. The computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of:
Establishing power grid graph data based on historical power grid running state data and historical topological information through a graph representation method;
performing multi-head attention information aggregation analysis on the power grid graph data to obtain multi-head attention information of the power grid graph data;
performing key operation feature identification based on the power grid graph data and the multi-head attention information, and determining power grid operation key features;
training an initial intelligent agent model based on the power grid graph data and the power grid operation key characteristics to obtain a power grid weak line identification model, wherein a loss function of the initial intelligent agent model is constructed by a cross entropy function with a weight coefficient, and the weight coefficient is set based on a line with historical risk or fault event.
In a fifth aspect, the present application also provides a computer program product. The computer program product comprises a computer program which, when executed by a processor, implements the steps of:
establishing power grid graph data based on historical power grid running state data and historical topological information through a graph representation method;
performing multi-head attention information aggregation analysis on the power grid graph data to obtain multi-head attention information of the power grid graph data;
Performing key operation feature identification based on the power grid graph data and the multi-head attention information, and determining power grid operation key features;
training an initial intelligent agent model based on the power grid graph data and the power grid operation key characteristics to obtain a power grid weak line identification model, wherein a loss function of the initial intelligent agent model is constructed by a cross entropy function with a weight coefficient, and the weight coefficient is set based on a line with historical risk or fault event.
The method for constructing the weak line identification model of the power grid comprises the steps of establishing power grid graph data based on historical power grid operation state data and historical topological information through a graph representation method; and performing multi-head attention information aggregation analysis on the power grid graph data to obtain multi-head attention information of the power grid graph data, and simultaneously performing key operation feature identification based on the power grid graph data and the multi-head attention information to determine key operation features of the power grid. Therefore, the identification capability for information aggregation and characteristic influence quantification can be effectively improved in the modeling process. And training an initial intelligent agent model based on the power grid graph data and the power grid operation key features to obtain a power grid weak line identification model, so that the obtained power grid weak line identification model can be used for identifying the weak line effectively and rapidly in an online operation stage, the application efficiency and feasibility of the online stage are improved, a weighted cross entropy function is used in a model training process to form a loss function of graph attention network training, and a weight correction coefficient is set for a line with historical risk or fault event, so that the identification efficiency of the power grid weak line identification model on the weak line can be effectively improved on the premise of ensuring the identification accuracy.
Drawings
FIG. 1 is an application environment diagram of a method for constructing a weak line identification model of a power grid in one embodiment;
FIG. 2 is a flow chart of a method for constructing a weak line identification model of a power grid in one embodiment;
FIG. 3 is a schematic diagram of an operation flow of the power grid weak line identification model construction in one embodiment;
FIG. 4 is a flowchart illustrating a weak line identification procedure in an embodiment;
FIG. 5 is a schematic diagram of a grid line of weakness identification system in one embodiment;
FIG. 6 is a block diagram of a power grid weak line identification model construction device in one embodiment;
fig. 7 is an internal structural diagram of a computer device in one embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application.
The method for constructing the weak line identification model of the power grid, which is provided by the embodiment of the application, can be applied to an application environment shown in fig. 1. Wherein the terminal 102 communicates with the server 104 via a network. The data storage system may store data required by the server 104 during the grid line of weakness identification process. The data storage system may be integrated on the server 104 or may be located on a cloud or other network server. The terminal 102 may send a request to the server 104 to build a power grid weak line identification model to build a corresponding power grid weak line identification model by the server 104. After receiving the request for constructing the weak line identification model of the power grid, the server 104 searches for historical power grid operation state data and historical topology information related to the weak line identification model of the power grid to be constructed, and then establishes power grid graph data based on the historical power grid operation state data and the historical topology information through a graph representation method; carrying out multi-head attention information aggregation analysis on the power grid graph data to obtain multi-head attention information of the power grid graph data; carrying out key operation feature identification based on the power grid graph data and the multi-head attention information, and determining power grid operation key features; training an initial intelligent agent model based on power grid graph data and power grid operation key characteristics to obtain a power grid weak line identification model, constructing a loss function of the initial intelligent agent model by a cross entropy function with a weight coefficient, and setting the weight coefficient based on a line with historical risk or fault event. The terminal 102 may be, but not limited to, various personal computers, notebook computers, smart phones, tablet computers, internet of things devices, and portable wearable devices, where the internet of things devices may be smart speakers, smart televisions, smart air conditioners, smart vehicle devices, and the like. The portable wearable device may be a smart watch, smart bracelet, headset, or the like. The server 104 may be implemented as a stand-alone server or as a server cluster of multiple servers.
In one embodiment, as shown in fig. 2, a method for constructing a weak line identification model of a power grid is provided, and the method is applied to the server 104 in fig. 1 for illustration, and includes the following steps:
step 201, building power grid graph data based on historical power grid operation state data and historical topological information through a graph representation method.
Wherein, the graph representation means that the grid information related to the historical grid operation state data and the historical topology information is represented in the form of graph data. Wherein the Graph is a discrete structure consisting of vertices and edges connecting the vertices. In computer science, a graph is one of the most flexible data structures, and many problems can be solved by modeling using a graph model, for example, weak line identification of a power grid in the application can also be realized by graph data modeling. The historical power grid operation state data refer to past operation data of a power grid which needs to be predicted at present, and specifically include power flow data, connection load information, line importance, line load and the like. And for historical topological data, the historical data refers to topological data. The topological structure of the power grid is an expression mode for abstracting various devices, components and circuits of the power grid into symbol connection patterns so as to perform various analyses and calculations.
Specifically, the scheme of the application realizes the identification of the weak line of the power grid in a mode of modeling the graph data. When modeling is performed, firstly, basic data for modeling needs to be obtained, so that historical power grid operation state data and historical topology information including data such as line connection relation, tide data, connection load information, line importance, line load rate and the like can be directly extracted from the historical data, then corresponding graph data are constructed based on the historical power grid operation state data and the historical topology information, when the graph data are constructed, the power grid operation state can be firstly sliced, then the graph data are converted into a graph network taking a power grid line as a graph node through a graph representation method, and line attribute features are taken as feature information X of the graph node and are specifically expressed as follows:
X＝[x 1 ,x 2 ,...x i ,...,x n ]，x i ＝[x 1i ,x 2i ,...x mi ] T
wherein x is i Is the characteristic information vector of the ith line, x mi And taking a graph data network of the running section of the power grid as input for the mth line attribute characteristic information of the ith line.
And 203, performing multi-head attention information aggregation analysis on the power grid graph data to obtain multi-head attention information of the power grid graph data.
Attention, among others, refers to the mechanism of attention, which stems from the study of human vision. In cognitive sciences, due to bottlenecks in information processing, humans may selectively focus on a portion of all information while ignoring other visible information. The mechanism of attention has two main aspects: deciding which part of the input needs to be focused on; the limited information processing resources are allocated to the important parts. And for multi-head attention (multi-head attention), it means that multiple queries are utilized to calculate in parallel the selection of multiple information from the input information. Each focusing on a different part of the input information.
Specifically, in the scheme of the application, the analysis processing process of the power grid graph data can be realized in a multi-head attention mode, and particularly multi-head attention information of the power grid graph data can be obtained by carrying out multi-head attention information aggregation analysis on the obtained power grid graph data. For the aggregation analysis process, attention coefficients of lines and adjacent information nodes in the graph data can be calculated based on topology information in the graph data and attribute characteristic information obtained based on running state data, then the graph node information is weighted and aggregated, and node characteristic mean value vectors of a multi-head attention network are obtained, so that multi-head attention information is obtained.
Step 205, performing key operation feature identification based on the power grid graph data and the multi-head attention information, and determining the key operation features of the power grid.
The key operation characteristics are operation characteristics with larger influence on the operation of the power grid line, so that the power grid weak line can be distinguished and identified based on the key operation characteristics.
Specifically, after the two data of the power grid graph data and the multi-head attention information are obtained, key operation feature identification can be directly performed based on the power grid graph data and the multi-head attention information, and power grid operation key features involved in the power grid operation process are determined. In a specific embodiment, extraction of key power grid operation features can be achieved by constructing a power grid operation model based on graph attention, and sampling and parallel training can be continuously performed to obtain key line operation features by setting a back propagation optimizer to conduct node feature screening and optimization in the training process of the power grid operation model. By extracting the characteristic information before the weak line of the power grid is identified, the intelligent body model can be effectively supported to carry out high-efficiency training and self-adaptive adjustment under the graph data and graph annotation force algorithm.
Step 207, training an initial intelligent agent model based on the power grid graph data and the power grid operation key characteristics to obtain a power grid weak line identification model, wherein a loss function of the initial intelligent agent model is constructed by a cross entropy function with a weight coefficient, and the weight coefficient is set based on a line with a historical risk or a fault event.
Specifically, for the construction process of the weak line identification model of the power grid, the initial agent model can be trained based on the power grid graph data and the determined key power grid operation characteristics to obtain the weak line identification model of the power grid. In the training process, the loss function of the intelligent agent model can be constructed through a cross entropy function with weight coefficients, and the weight coefficients are set on the basis of lines with historical risks or fault events. For the loss function referred to in this application, it may be specifically:
in the method, in the process of the invention,m is the number of lines; l (L) i If the line is a weak line, then l is i 1, otherwise 0; p is p 1i 、p 2i Judging the probability value of the line as weak or normal line; w (w) key 、And the duty ratio weights of the weak line and the normal line in the current training sample are respectively.
And for the probability value distinguished by the line type calculated by the historical risk or fault event sample, the probability value is specifically:
Wherein m is l1 、m l2 The weak line and normal line numbers in the pool are replayed for training experience. In a specific embodiment, the processing flow of the method comprises the processing of running data graph representation and multi-head attention information aggregation, then key running feature identification and extraction are carried out, and training of an intelligent body model for identifying the weak line of the power grid under multi-head attention is achieved, so that the identification of the weak line of the power grid is achieved through the trained intelligent body model. By setting a weight correction coefficient for a line with a historical risk or a fault event, an algorithm function of feedback correction is provided for the business of integrating data and experience knowledge of weak line identification, and self-optimizing adjustment and training learning of knowledge-embedded power grid weak line intelligent bodies are realized. In one embodiment, a correction link can be further set by combining the discrimination results of the key lines in the history approximation scene when the intelligent body model is constructed, so that intelligent body identification optimizing can be adaptively adjusted. Specifically, a correction link can be set by comparing the judging result of the key line in the history approximate scene, and the formula of the correction link is as follows:
Reward＝P acr -∑V pen,i
wherein, reward is a graph annotation meaning network agent model rewarding function; p (P) acr Sigma V is the recognition accuracy of the verification after the sample pen,i Penalty coefficients for violating constraint terms. The penalty term correction link is set to adjust the winning function of the intelligent body model, the feature similarity between extreme operation scenes of the power grid is identified, the feature distance calculation of the history similar scenes and the post-verification accuracy are utilized to guide the optimizing efficiency of model training and learning, so that the training convergence efficiency of the weak line identification intelligent body is accelerated, and the deployment and long-term updating operation of the application stage of the method are improved.
The method for constructing the weak line identification model of the power grid comprises the steps of establishing power grid graph data based on historical power grid running state data and historical topological information through a graph representation method; and performing multi-head attention information aggregation analysis on the power grid graph data to obtain multi-head attention information of the power grid graph data, and simultaneously performing key operation feature identification based on the power grid graph data and the multi-head attention information to determine key operation features of the power grid. Therefore, the identification capability for information aggregation and characteristic influence quantification can be effectively improved in the modeling process. And training an initial intelligent agent model based on the power grid graph data and the power grid operation key characteristics to obtain a power grid weak line identification model, so that the obtained power grid weak line identification model can be used for identifying the weak line effectively and rapidly in an online operation stage, the application efficiency and feasibility of the online stage are improved, a loss function of graph attention network training is formed by using a cross entropy function with weights in the model training process, and a weight correction coefficient is set for a line with a historical risk or a fault event, so that the identification efficiency of the power grid weak line identification model on the weak line can be effectively improved on the premise of ensuring the identification accuracy.
In one embodiment, step 203 comprises: determining the attention coefficients of the lines and adjacent information nodes in the power grid graph data based on the topology information and the attribute characteristic information of the power grid graph data; and carrying out node information aggregation processing on the power grid graph data based on the attention coefficient to obtain multi-head attention information of the power grid graph data.
Specifically, in the scheme of the application, the multi-head attention aggregation mechanism is used for adaptively obtaining information aggregation weights under different inputs through a multi-layer cross matrix, and the multi-head attention aggregation mechanism is used for coping with extraction of different key index amounts and dynamic aggregation of mass data in different power grid operation scenes, so that the characteristic information aggregation efficiency and data effectiveness under different scenes (inputs) are improved; as shown in fig. 3, in the multi-head attention information aggregation in the figure, a plurality of groups of (query/key/value) combined weight matrixes of dynamic attention under a multi-layer subspace are set, and are randomly initialized in training, different information aggregation effects and feedback weights are formed in different multi-head attention mechanisms, and high-dimensional data with different characteristic weight values are formed after the power grid data under different inputs are aggregated by the multi-head attention mechanisms, so that characteristic weight vectors are provided for operation characteristic extraction. For the course of the attention coefficient, reference may be made in particular to the formula:
Wherein: gamma ray ij The attention coefficient between the node i and the node j; x is x i An initial attribute information vector for node i; w epsilon R n→n' Mapping the node feature vector from n dimensions to n' dimensions; d (i) is a set of adjacent nodes of node i;is a single-layer neural network; the LeakyReLU function calculates the attention weight between node i and its neighbor node j. Then weighting and aggregating each graph node information in the graph data based on the attention coefficient to form an attribute vector of a new feature, weighting parameters of the multi-layer graph attention network, and taking a node feature mean value of the multi-head attention network as an output vector ∈>
Wherein K is the number of attention heads; sigma () is an activation function, typically a ReLU function. In this embodiment, the processing of node information aggregation is implemented by the attention coefficient between the line and the information node, so as to ensure the accuracy of multi-head attention extraction
In one embodiment, the method further comprises: in the training process of the initial intelligent agent model, node characteristic screening and optimization are carried out on the weak line identification model of the power grid through a back propagation optimizer.
Specifically, in the training process of the intelligent body model, after each training is completed, a training condition of the current intelligent body model can be obtained through a loss function, gradient descent is performed on the loss value for the parameters trained by the next model, and a proper value is found, wherein the step is back propagation, specifically, back propagation can be performed through a back wave () method of the loss function, and an optimizer is generally used for optimizing the model parameters. In one embodiment, a back propagation optimizer is arranged to perform node feature screening and optimization, and a random gradient optimization model is adopted as follows
Wherein N is a graph annotation meaning network parameter; gamma is the learning rate; c is the iteration round.
The optimizer performs parallel optimization, samples in batches and updates the optimizing direction, and an optimizing training model is as follows:
in one embodiment, the method further comprises: performing simulation based on the historical power grid running state data and the target scene to obtain a simulation result; generating a power grid running state data sample through a generation type countermeasure network based on a simulation result; and constructing a sample set of historical grid operation state data based on the grid operation state data sample and the historical grid operation state data.
The target scene refers to a weather scene set based on sample expansion requirements, such as extreme weather of typhoons, cold waves and the like, and the data set can be expanded by setting common sunny conditions or rainy conditions according to requirements.
Specifically, the sample set of the historical power grid operation state data mainly comprises power grid operation state data and topology information under different power grid operation scenes, and the data requirements of the graph attention network algorithm identification can be ensured by constructing the sample set because the data sample of the actual extreme power grid operation scene data is rarely generated in an offline mode. The sample set of the historical power grid running state data can be specifically formed by randomly sampling historical running section data and simulation results, and the historical running section data and the simulation results are used for carrying out data expansion by adopting a generation countermeasure network algorithm through updating and identifying process state parameters according to a running scene obtained by sampling in the training process. In addition, when extreme weather events such as typhoons, chills and the like occur frequently, the extreme weather will seriously jeopardize the operation safety and the power supply stability of the power system. Therefore, it is necessary to expand the scene of the weak line of the power grid for these extreme weather events, and the historical data of the extreme weather events such as typhoons, chills and the like in the historical data is lacking. Thus acquiring historical operating section data; performing simulation based on the historical operation section data to obtain a simulation result corresponding to the historical operation section data; generating a power grid running state data sample through a generation type countermeasure network based on a simulation result; and constructing a sample set of historical grid operation state data based on the grid operation state data samples. The data expansion of the polar weather operation scene is realized by adopting the generation countermeasure network to carry out the analog simulation, so that the coverage rate of the polar weather operation scene characteristics is improved. In this embodiment, the sample set of the historical power grid operation state data is expanded in a simulation mode, so that effective assurance can be ensured
In one embodiment, the method further comprises: acquiring real-time running state data of a power grid; and inputting the real-time running state data of the power grid into a power grid weak line identification model, and determining the power grid weak line.
Specifically, after the weak line identification model of the power grid is established, the weak line of the power grid can be identified based on the established weak line identification model of the power grid, and the weak line of the power grid is judged based on the current line running condition. Therefore, when the weak line is identified, the real-time running state data of the power grid can be input into a power grid weak line identification model, and the power grid weak line existing in the current running line is determined based on the power grid weak line identification model. In one embodiment, the complete flow of the present application may refer to fig. 4, which includes two flows of training a weak line identification model of a power grid in a scene and outputting a real-time weak line, and after outputting the weak line, a new sample may be further constructed based on the identified weak line to update and iterate the weak line identification model of the power grid, so as to ensure the identification accuracy of the weak line identification model of the power grid. In the embodiment, the weak line represented by the real-time running state data of the power grid is identified through the weak line identification model of the power grid, so that the accuracy of the identification of the weak line of the power grid can be effectively ensured.
In one embodiment, the method further comprises: determining index values of the weak line of the power grid in each weak degree evaluation index; performing initial weighting treatment on the weakness evaluation indexes through a membership rule to obtain initial weights of the weakness evaluation indexes; self-adaptive adjustment is carried out on the initial weight through a heuristic algorithm, so that index weight is obtained; and determining the weakness parameters of the weak line of the power grid based on the index values and the index weights of the weak line of the power grid in each weakness evaluation index.
The weakness evaluation index is index data for evaluating the weakness of the power grid, which are included in a power grid line weakness evaluation index system, and specifically comprises line fault probability, fault load loss, voltage out-of-limit and operation risk values. The membership degree principle is also called as a maximum membership degree principle, and refers to a criterion for judging which fuzzy subset of n fuzzy subsets of the domain X any element X belongs to when fuzzy clustering and fuzzy pattern recognition are performed. If x is assigned to the fuzzy subset. Heuristic algorithms (heuristic algorithm) are proposed with respect to optimization algorithms. An optimization algorithm for a problem finds an optimal solution for each instance of the problem. The heuristic algorithm may be defined as follows: an algorithm based on visual or empirical construction gives a viable solution to each instance of the combinatorial optimization problem to be solved at acceptable expense (referring to computation time and space), which generally cannot be predicted from the optimal solution. The heuristic algorithm specifically comprises a gray wolf optimization algorithm, an improved particle swarm algorithm and the like.
Specifically, in the scheme of the application, the comparison and identification of the weak line of the power grid can be realized by constructing a weak degree evaluation index system in advance. Firstly, determining all weak lines existing in a current power grid line, and then determining index values of the weak lines of the power grid under weak evaluation indexes such as line fault probability, fault load loss, voltage out-of-limit, operation risk value and the like. And simultaneously calculating weights of different indexes, and finally calculating the final weakness degree parameter by combining the weights and the index values. In a specific embodiment, firstly, carrying out initial weighting treatment on the weakness evaluation indexes through a membership rule to obtain initial weights of the weakness evaluation indexes; the initial weight is adaptively adjusted through a heuristic algorithm, the process of obtaining the index weight can be directly calculated when a weaknesses evaluation index system is constructed, and then the weaknesses parameters can be directly calculated only by the index values of the weaknesses evaluation indexes based on the weak lines of the power grid when the weaknesses parameters are calculated, so that the calculation efficiency of the weaknesses parameters can be effectively improved. In a specific embodiment, the line priority can be set based on the weakness parameters of different weak lines, so that the management sequence setting of the different weak lines is completed, and the operation effect of the power grid line is ensured. In the embodiment, the grid line weakness evaluation index system considers the evaluation system of dimensions such as line tide indexes, load characteristics, risk quantification indexes and the like when being established, and then the self-adaptive adjustment weight is corrected by combining the experience membership and the heuristic algorithm, so that the comprehensiveness and accuracy of the weak line evaluation can be effectively improved.
In one embodiment, the method for constructing the weak line identification model of the power grid in the embodiment can be implemented through a weak line identification system of the power grid, and a structure diagram of the weak line identification system of the power grid based on a graph attention network algorithm can be shown by referring to fig. 5, and the system comprises a data acquisition module, an offline training module, a data storage module and an online identification module. The data acquisition module is used for acquiring real-time running state data and topology information by associating the running monitoring system, the load characteristic information and the scheduling correction port with human-machine operation. The off-line training module performs line risk judging index calculation, preset fault simulation, running diagram data expansion and representation conversion, off-line training of the diagram attention network, and the associated data storage module data interface performs line identification sample data calling and identification result storage. The trained weak line identification model of the power grid can be on line in an online identification module, and in a real-time operation stage, functions of real-time monitoring, online identification and result inspection are provided, and the weak line identification result of the power grid is displayed. The data storage module of the system mainly provides a storage database, stores historical line faults, line key characteristic parameters and intelligent body model correction parameters, and stores corresponding identification link information with an off-line training simulation associated data interface;
It should be understood that, although the steps in the flowcharts related to the above embodiments are sequentially shown as indicated by arrows, these steps are not necessarily sequentially performed in the order indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in the flowcharts described in the above embodiments may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily performed sequentially, but may be performed alternately or alternately with at least some of the other steps or stages.
Based on the same inventive concept, the embodiment of the application also provides a power grid weak line identification model construction device for realizing the power grid weak line identification model construction method. The implementation scheme of the device for solving the problem is similar to the implementation scheme described in the method, so the specific limitation in the embodiment of the device for constructing the weak line identification model of the power grid provided below can be referred to the limitation of the method for constructing the weak line identification model of the power grid hereinabove, and will not be repeated here.
In one embodiment, as shown in fig. 6, there is provided a weak line identification model construction device for a power grid, including:
the data construction module 601 is configured to establish, by using a graph representation method, grid graph data based on historical grid operation state data and historical topology information;
the aggregation analysis module 603 is configured to perform multi-head attention information aggregation analysis on the grid graph data to obtain multi-head attention information of the grid graph data;
the feature recognition module 605 is configured to perform key operation feature recognition based on the grid graph data and the multi-head attention information, and determine a key operation feature of the grid;
the model construction module 607 is configured to train the initial agent model based on the grid graph data and the key feature of the grid operation to obtain a weak line identification model of the grid, where the loss function of the initial agent model is constructed by a cross entropy function with a weight coefficient, and the weight coefficient is set based on the line with a historical risk or a fault event.
In one embodiment, the aggregate analysis module 603 is specifically configured to: determining the attention coefficients of the lines and adjacent information nodes in the power grid graph data based on the topology information and the attribute characteristic information of the power grid graph data; and carrying out node information aggregation processing on the power grid graph data based on the attention coefficient to obtain multi-head attention information of the power grid graph data.
In one embodiment, the method further comprises a node optimization module for: in the training process of the initial intelligent agent model, node characteristic screening and optimization are carried out on the weak line identification model of the power grid through a back propagation optimizer.
In one embodiment, the system further comprises a sample expansion module for: performing simulation based on the historical power grid running state data and the target scene to obtain a simulation result; generating a power grid running state data sample through a generation type countermeasure network based on a simulation result; and constructing a sample set of historical grid operation state data based on the grid operation state data sample and the historical grid operation state data.
In one embodiment, the system further comprises a real-time identification module for: acquiring real-time running state data of a power grid; and inputting the real-time running state data of the power grid into a power grid weak line identification model, and determining the power grid weak line.
In one embodiment, the system further comprises a weak parameter identification module for: determining index values of the weak line of the power grid in each weak degree evaluation index; performing initial weighting treatment on the weakness evaluation indexes through a membership rule to obtain initial weights of the weakness evaluation indexes; self-adaptive adjustment is carried out on the initial weight through a heuristic algorithm, so that index weight is obtained; and determining the weakness parameters of the weak line of the power grid based on the index values and the index weights of the weak line of the power grid in each weakness evaluation index.
The modules in the power grid weak line identification model construction device can be realized in whole or in part through software, hardware and a combination thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.
In one embodiment, a computer device is provided, which may be a server, the internal structure of which may be as shown in fig. 7. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer device is for storing video analytics data. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method for constructing a weak line identification model of a power grid.
It will be appreciated by those skilled in the art that the structure shown in fig. 7 is merely a block diagram of some of the structures associated with the present application and is not limiting of the computer device to which the present application may be applied, and that a particular computer device may include more or fewer components than shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is provided comprising a memory and a processor, the memory having stored therein a computer program, the processor when executing the computer program performing the steps of:
establishing power grid graph data based on historical power grid running state data and historical topological information through a graph representation method;
carrying out multi-head attention information aggregation analysis on the power grid graph data to obtain multi-head attention information of the power grid graph data;
carrying out key operation feature identification based on the power grid graph data and the multi-head attention information, and determining power grid operation key features;
training an initial intelligent agent model based on power grid graph data and power grid operation key characteristics to obtain a power grid weak line identification model, constructing a loss function of the initial intelligent agent model by a cross entropy function with a weight coefficient, and setting the weight coefficient based on a line with historical risk or fault event.
In one embodiment, the processor when executing the computer program further performs the steps of: determining the attention coefficients of the lines and adjacent information nodes in the power grid graph data based on the topology information and the attribute characteristic information of the power grid graph data; and carrying out node information aggregation processing on the power grid graph data based on the attention coefficient to obtain multi-head attention information of the power grid graph data.
In one embodiment, the processor when executing the computer program further performs the steps of: in the training process of the initial intelligent agent model, node characteristic screening and optimization are carried out on the weak line identification model of the power grid through a back propagation optimizer.
In one embodiment, the processor when executing the computer program further performs the steps of: performing simulation based on the historical power grid running state data and the target scene to obtain a simulation result; generating a power grid running state data sample through a generation type countermeasure network based on a simulation result; and constructing a sample set of historical grid operation state data based on the grid operation state data sample and the historical grid operation state data.
In one embodiment, the processor when executing the computer program further performs the steps of: acquiring real-time running state data of a power grid; and inputting the real-time running state data of the power grid into a power grid weak line identification model, and determining the power grid weak line.
In one embodiment, the processor when executing the computer program further performs the steps of: determining index values of the weak line of the power grid in each weak degree evaluation index; performing initial weighting treatment on the weakness evaluation indexes through a membership rule to obtain initial weights of the weakness evaluation indexes; self-adaptive adjustment is carried out on the initial weight through a heuristic algorithm, so that index weight is obtained; and determining the weakness parameters of the weak line of the power grid based on the index values and the index weights of the weak line of the power grid in each weakness evaluation index.
In one embodiment, a computer readable storage medium is provided having a computer program stored thereon, which when executed by a processor, performs the steps of:
establishing power grid graph data based on historical power grid running state data and historical topological information through a graph representation method;
carrying out multi-head attention information aggregation analysis on the power grid graph data to obtain multi-head attention information of the power grid graph data;
carrying out key operation feature identification based on the power grid graph data and the multi-head attention information, and determining power grid operation key features;
training an initial intelligent agent model based on power grid graph data and power grid operation key characteristics to obtain a power grid weak line identification model, constructing a loss function of the initial intelligent agent model by a cross entropy function with a weight coefficient, and setting the weight coefficient based on a line with historical risk or fault event.
In one embodiment, the computer program when executed by the processor further performs the steps of: determining the attention coefficients of the lines and adjacent information nodes in the power grid graph data based on the topology information and the attribute characteristic information of the power grid graph data; and carrying out node information aggregation processing on the power grid graph data based on the attention coefficient to obtain multi-head attention information of the power grid graph data.
In one embodiment, the computer program when executed by the processor further performs the steps of: in the training process of the initial intelligent agent model, node characteristic screening and optimization are carried out on the weak line identification model of the power grid through a back propagation optimizer.
In one embodiment, the computer program when executed by the processor further performs the steps of: performing simulation based on the historical power grid running state data and the target scene to obtain a simulation result; generating a power grid running state data sample through a generation type countermeasure network based on a simulation result; and constructing a sample set of historical grid operation state data based on the grid operation state data sample and the historical grid operation state data.
In one embodiment, the computer program when executed by the processor further performs the steps of: acquiring real-time running state data of a power grid; and inputting the real-time running state data of the power grid into a power grid weak line identification model, and determining the power grid weak line.
In one embodiment, the computer program when executed by the processor further performs the steps of: determining index values of the weak line of the power grid in each weak degree evaluation index; performing initial weighting treatment on the weakness evaluation indexes through a membership rule to obtain initial weights of the weakness evaluation indexes; self-adaptive adjustment is carried out on the initial weight through a heuristic algorithm, so that index weight is obtained; and determining the weakness parameters of the weak line of the power grid based on the index values and the index weights of the weak line of the power grid in each weakness evaluation index.
In one embodiment, a computer program product is provided comprising a computer program which, when executed by a processor, performs the steps of:
establishing power grid graph data based on historical power grid running state data and historical topological information through a graph representation method;
carrying out multi-head attention information aggregation analysis on the power grid graph data to obtain multi-head attention information of the power grid graph data;
carrying out key operation feature identification based on the power grid graph data and the multi-head attention information, and determining power grid operation key features;
training an initial intelligent agent model based on power grid graph data and power grid operation key characteristics to obtain a power grid weak line identification model, constructing a loss function of the initial intelligent agent model by a cross entropy function with a weight coefficient, and setting the weight coefficient based on a line with historical risk or fault event.
In one embodiment, the computer program when executed by the processor further performs the steps of: determining the attention coefficients of the lines and adjacent information nodes in the power grid graph data based on the topology information and the attribute characteristic information of the power grid graph data; and carrying out node information aggregation processing on the power grid graph data based on the attention coefficient to obtain multi-head attention information of the power grid graph data.
In one embodiment, the computer program when executed by the processor further performs the steps of: in the training process of the initial intelligent agent model, node characteristic screening and optimization are carried out on the weak line identification model of the power grid through a back propagation optimizer.
In one embodiment, the computer program when executed by the processor further performs the steps of: performing simulation based on the historical power grid running state data and the target scene to obtain a simulation result; generating a power grid running state data sample through a generation type countermeasure network based on a simulation result; and constructing a sample set of historical grid operation state data based on the grid operation state data sample and the historical grid operation state data.
In one embodiment, the computer program when executed by the processor further performs the steps of: acquiring real-time running state data of a power grid; and inputting the real-time running state data of the power grid into a power grid weak line identification model, and determining the power grid weak line.
In one embodiment, the computer program when executed by the processor further performs the steps of: determining index values of the weak line of the power grid in each weak degree evaluation index; performing initial weighting treatment on the weakness evaluation indexes through a membership rule to obtain initial weights of the weakness evaluation indexes; self-adaptive adjustment is carried out on the initial weight through a heuristic algorithm, so that index weight is obtained; and determining the weakness parameters of the weak line of the power grid based on the index values and the index weights of the weak line of the power grid in each weakness evaluation index.
It should be noted that, user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, presented data, etc.) referred to in the present application are information and data authorized by the user or sufficiently authorized by each party.
Those skilled in the art will appreciate that implementing all or part of the above-described methods in accordance with the embodiments may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed may comprise the steps of the embodiments of the methods described above. Any reference to memory, database, or other medium used in the various embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile memory may include Read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high density embedded nonvolatile memory, resistive random access memory (ReRAM), magnetic random access memory (MagnetoresistiveRandomAccessMemory, MRAM), ferroelectric memory (FerroelectricRandomAccessMemory, FRAM), phase change memory (PhaseChangeMemory, PCM), graphene memory, and the like. Volatile memory may include random access memory (RandomAccessMemory, RAM) or external cache memory, and the like. By way of illustration, and not limitation, RAM can take many forms, such as static random access memory (StaticRandomAccessMemory, SRAM) or dynamic random access memory (DynamicRandomAccessMemory, DRAM), among others. The databases referred to in the various embodiments provided herein may include at least one of relational databases and non-relational databases. The non-relational database may include, but is not limited to, a blockchain-based distributed database, and the like. The processors referred to in the embodiments provided herein may be general purpose processors, central processing units, graphics processors, digital signal processors, programmable logic units, quantum computing-based data processing logic units, etc., without being limited thereto.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The foregoing examples represent only a few embodiments of the present application, which are described in more detail and are not thereby to be construed as limiting the scope of the present application. It should be noted that it would be apparent to those skilled in the art that various modifications and improvements could be made without departing from the spirit of the present application, which would be within the scope of the present application. Accordingly, the scope of protection of the present application shall be subject to the appended claims.
Claims (10)
1. The utility model provides a power grid weak line identification model construction method, which is characterized by comprising the following steps:
establishing power grid graph data based on historical power grid running state data and historical topological information through a graph representation method;
performing multi-head attention information aggregation analysis on the power grid graph data to obtain multi-head attention information of the power grid graph data;
performing key operation feature identification based on the power grid graph data and the multi-head attention information, and determining power grid operation key features;
Training an initial intelligent agent model based on the power grid graph data and the power grid operation key characteristics to obtain a power grid weak line identification model, wherein a loss function of the initial intelligent agent model is constructed by a cross entropy function with a weight coefficient, and the weight coefficient is set based on a line with historical risk or fault event.
2. The method of claim 1, wherein performing a multi-headed attention information aggregate analysis on the grid map data to obtain multi-headed attention information for the grid map data comprises:
determining the attention coefficients of the lines and adjacent information nodes in the power grid graph data based on the topology information and the attribute characteristic information of the power grid graph data;
and carrying out node information aggregation processing on the power grid graph data based on the attention coefficient to obtain multi-head attention information of the power grid graph data.
3. The method according to claim 1, wherein the method further comprises:
and in the training process of the initial intelligent agent model, node characteristic screening and optimization are carried out on the weak line identification model of the power grid through a back propagation optimizer.
4. The method according to claim 1, wherein the method further comprises:
Performing simulation based on the historical power grid running state data and the target scene to obtain a simulation result;
generating a power grid running state data sample through a generation type countermeasure network based on the simulation result;
and constructing a sample set of the historical power grid operation state data based on the power grid operation state data sample and the historical power grid operation state data.
5. The method according to claim 1, wherein the method further comprises:
acquiring real-time running state data of a power grid;
and inputting the real-time running state data of the power grid into the power grid weak line identification model to determine the power grid weak line.
6. The method of claim 5, wherein the method further comprises:
determining index values of the weak line of the power grid in each weak degree evaluation index;
performing initial weighting treatment on the weakness evaluation indexes through a membership rule to obtain initial weights of the weakness evaluation indexes;
performing self-adaptive adjustment on the initial weight through a heuristic algorithm to obtain an index weight;
and determining the weakness parameters of the weak line of the power grid based on the index values of the weak line of the power grid in each weakness evaluation index and the index weights.
7. A power grid weak line identification model construction device, characterized in that the device comprises:
the data construction module is used for establishing power grid graph data based on historical power grid running state data and historical topological information through a graph representation method;
the aggregation analysis module is used for carrying out multi-head attention information aggregation analysis on the power grid graph data to obtain multi-head attention information of the power grid graph data;
the feature recognition module is used for carrying out key operation feature recognition based on the power grid graph data and the multi-head attention information and determining power grid operation key features;
the model construction module is used for training an initial intelligent agent model based on the power grid graph data and the power grid operation key characteristics to obtain a power grid weak line identification model, a loss function of the initial intelligent agent model is constructed by a cross entropy function with a weight coefficient, and the weight coefficient is set based on a line with historical risk or fault event.
8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 6 when the computer program is executed.
9. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.
10. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310504717.9A CN116523001A (en) | 2023-05-06 | 2023-05-06 | Method, device and computer equipment for constructing weak line identification model of power grid |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310504717.9A CN116523001A (en) | 2023-05-06 | 2023-05-06 | Method, device and computer equipment for constructing weak line identification model of power grid |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116523001A true CN116523001A (en) | 2023-08-01 |
Family
ID=87397270
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310504717.9A Pending CN116523001A (en) | 2023-05-06 | 2023-05-06 | Method, device and computer equipment for constructing weak line identification model of power grid |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116523001A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117239743A (en) * | 2023-11-15 | 2023-12-15 | 青岛鼎信通讯股份有限公司 | Electric energy meter electricity load acquisition method, device, equipment and medium |
| CN118780397B (en) * | 2024-09-06 | 2024-11-19 | 云和恩墨(北京)信息技术有限公司 | Data enhancement method, device, electronic equipment and storage medium |
-
2023
- 2023-05-06 CN CN202310504717.9A patent/CN116523001A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117239743A (en) * | 2023-11-15 | 2023-12-15 | 青岛鼎信通讯股份有限公司 | Electric energy meter electricity load acquisition method, device, equipment and medium |
| CN117239743B (en) * | 2023-11-15 | 2024-02-27 | 青岛鼎信通讯股份有限公司 | Electric energy meter electricity load acquisition method, device, equipment and medium |
| CN118780397B (en) * | 2024-09-06 | 2024-11-19 | 云和恩墨(北京)信息技术有限公司 | Data enhancement method, device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11042802B2 (en) | System and method for hierarchically building predictive analytic models on a dataset | |
| Barzola-Monteses et al. | Energy consumption of a building by using long short-term memory network: a forecasting study | |
| CN114065653A (en) | Construction method of power load prediction model and power load prediction method | |
| CN117674119A (en) | Power grid operation risk assessment method, device, computer equipment and storage medium | |
| CN116451093A (en) | Training method of circuit fault analysis model and circuit fault analysis method | |
| CN116523001A (en) | Method, device and computer equipment for constructing weak line identification model of power grid | |
| CN111192158A (en) | Transformer substation daily load curve similarity matching method based on deep learning | |
| CN113010774B (en) | Click rate prediction method based on dynamic deep attention model | |
| CN109697511B (en) | Data reasoning method and device and computer equipment | |
| CN112330044A (en) | Support vector regression model based on iterative aggregation grid search algorithm | |
| CN114338416A (en) | Space-time multi-index prediction method and device and storage medium | |
| CN118014352A (en) | Dynamic early warning method and device for manufacturing cost of power transformation project based on LIME model | |
| CN117671278A (en) | Power image data processing method, apparatus, device, medium, and program product | |
| Aruna et al. | Sparrow Search Optimization with Deep Belief Network based Wind Power Prediction Model | |
| Lu et al. | Laplacian deep echo state network optimized by genetic algorithm | |
| CN115169465A (en) | Transformer data anomaly detection method, device, equipment, medium and product | |
| CN113723660A (en) | Specific behavior type prediction method and system based on DNN-LSTM fusion model | |
| CN114254828A (en) | Power load prediction method based on hybrid convolution feature extractor and GRU | |
| CN116526582B (en) | Combined dispatching method and system for electric power unit based on artificial intelligence combined driving | |
| Hsieh | Employing data mining technique to achieve the parameter optimization based on manufacturing intelligence | |
| CN115577857B (en) | Method and device for predicting output data of energy system and computer equipment | |
| CN117973647B (en) | Comprehensive energy system load prediction method and device and computer equipment | |
| CN116454890B (en) | Combined control method, device and equipment for unit based on SCUC model | |
| Yunianta et al. | Modelling of network traffic usage using self-organizing maps techniques | |
| CN117976238A (en) | Infectious disease trend prediction method, infectious disease trend prediction device, infectious disease trend prediction computer device, and infectious disease trend prediction storage medium |
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 |