CN116613754A - Power distribution system reliability assessment method, model training method, device and equipment - Google Patents
Power distribution system reliability assessment method, model training method, device and equipment Download PDFInfo
- Publication number
- CN116613754A CN116613754A CN202310899972.8A CN202310899972A CN116613754A CN 116613754 A CN116613754 A CN 116613754A CN 202310899972 A CN202310899972 A CN 202310899972A CN 116613754 A CN116613754 A CN 116613754A
- Authority
- CN
- China
- Prior art keywords
- data
- sample
- reliability
- power distribution
- distribution system
- 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
- 238000009826 distribution Methods 0.000 title claims abstract description 185
- 238000012549 training Methods 0.000 title claims abstract description 111
- 238000000034 method Methods 0.000 title claims abstract description 77
- 238000003062 neural network model Methods 0.000 claims abstract description 98
- 238000000547 structure data Methods 0.000 claims abstract description 73
- 238000011156 evaluation Methods 0.000 claims abstract description 29
- 239000011159 matrix material Substances 0.000 claims description 31
- 230000008569 process Effects 0.000 claims description 20
- 238000004590 computer program Methods 0.000 claims description 19
- 238000003860 storage Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 description 11
- 238000013528 artificial neural network Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 7
- 230000008859 change Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 239000004973 liquid crystal related substance Substances 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 238000013500 data storage Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000008439 repair process Effects 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 239000013598 vector Substances 0.000 description 2
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 229910021389 graphene Inorganic materials 0.000 description 1
- 230000006872 improvement Effects 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
- 238000005457 optimization Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 235000002020 sage Nutrition 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- 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
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/10—Power transmission or distribution systems management focussing at grid-level, e.g. load flow analysis, node profile computation, meshed network optimisation, active network management or spinning reserve management
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
-
- 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)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Power Engineering (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The application relates to a power distribution system reliability assessment method, a model training method, a device and equipment. The method comprises the following steps: obtaining basic information data of a target power distribution system, determining graph structure data by taking each electrical node in the basic information data as a node in a graph domain, inputting the graph structure data into a graph neural network model, and obtaining a reliability data set corresponding to each electrical node in the target power distribution system, wherein the reliability data set corresponding to each electrical node is used for representing the reliability performance of the node level of the target power distribution system, the graph neural network model is obtained by training according to sample graph structure data corresponding to a sample power distribution system and reliability data label sets of each sample electrical node, and sample connection relation data in the sample power distribution system is identical with connection relation data in the target power distribution system. By adopting the method, the evaluation efficiency can be improved, and the reliability of the power distribution system can be rapidly calculated.
Description
Technical Field
The application relates to the technical field of reliability of power systems, in particular to a power distribution system reliability assessment method, a model training method, a device and equipment.
Background
With the popularization of electric power sources and the continuous improvement of requirements on electric power supply safety, the requirements on the reliability of an electric power system are also increasing. The distribution system is located at the end of the power system, and is an important link for connecting the power supply system or the power transmission and transformation system with the user facilities, distributing electric energy to users and supplying electric energy. The power distribution system generally comprises the whole power distribution network including a power distribution substation, a power distribution line and a service line and equipment thereof, so that improving the reliability level of the power distribution system is one of main and important means for ensuring the reliability level of the power system.
The reliability evaluation of the power distribution system refers to calculating various reliability index data of the power distribution system by using topology information of the power distribution system and reliability parameters of equipment elements, such as failure rate of the equipment elements, average repair time and the like. Because of the rapid development of the power distribution system, the topology structure of the power distribution system is more complex, the data volume is increased rapidly, and the reliability evaluation method of the current power distribution system is easy to sink into dimension disasters and the calculation time is long.
Disclosure of Invention
In view of the foregoing, it is desirable to provide a power distribution system reliability evaluation method, a model training method, a device, and equipment that can improve evaluation efficiency and quickly calculate the reliability of the power distribution system.
In a first aspect, the present application provides a method for evaluating reliability of a power distribution system. The method comprises the following steps:
basic information data of a target power distribution system are acquired;
determining graph structure data by taking each electrical node in the basic information data as a node in a graph domain, wherein the graph structure data comprises connection relation data among the electrical nodes and attribute characteristic data of the electrical nodes in a preset target period;
inputting the graph structure data into a graph neural network model to obtain reliability data sets corresponding to all electrical nodes in a target power distribution system, wherein the reliability data sets corresponding to all electrical nodes are used for representing the reliability performance of the node level of the target power distribution system, and the graph neural network model is obtained by training according to the sample graph structure data corresponding to the sample power distribution system and the reliability data label sets of all sample electrical nodes, and the sample connection relation data in the sample power distribution system is identical to the connection relation data in the target power distribution system.
In one embodiment, the attribute feature data includes type feature data corresponding to each electrical node, where the type feature data is used to characterize whether the electrical node is a load point; in the reliability data label set of each sample electrical node, the reliability data label of the sample electrical node corresponding to the non-load point is zero.
In one embodiment, the training process of the graph neural network model includes:
acquiring a training sample, wherein the training sample comprises corresponding records between sample graph structure data and a reliability data tag set;
based on the training sample, carrying out iterative training on the initial graph neural network model, and outputting a trained graph neural network model; wherein, in each iterative training process, the following operations are performed:
and inputting the sample graph data structure into an initial graph neural network model, obtaining a reliability prediction data set of each sample electrical node, determining a loss value through reliability prediction data and reliability data labels of the sample electrical nodes corresponding to load points in the reliability prediction data set, and adopting the loss value to carry out parameter adjustment.
In one embodiment, based on the training samples, the initial graph neural network model is iteratively trained to output a trained graph neural network model, and the method further comprises:
in each iterative training process, determining the accuracy of prediction of the initial graph neural network model through reliability prediction data and reliability data labels of sample electrical nodes corresponding to load points in the reliability prediction data set;
And outputting the trained graphic neural network model under the condition that the accuracy rate reaches a preset threshold value.
In one embodiment, the method further comprises:
extracting reliability data of the electrical nodes corresponding to all load points from the reliability data set corresponding to each electrical node;
and determining reliability data of a system in the target power distribution system based on the reliability data of the electrical nodes corresponding to the load points, wherein the reliability data of the system is used for representing the reliability performance of the system level of the target power distribution system.
In one embodiment, determining the graph structure data by taking each electrical node in the basic information data as a node in the graph domain includes:
constructing an adjacency matrix through connection relation data among the electrical nodes;
constructing a node characteristic matrix through attribute characteristic data of each electrical node in a preset target period;
and combining the adjacency matrix and the node characteristic matrix to obtain the graph structure data.
In a second aspect, the present application also provides a model training method, which includes:
obtaining training samples, wherein the training samples comprise corresponding records between sample graph structure data of a sample power distribution system and reliability data label sets of all sample electrical nodes, and the sample graph structure data comprises connection relation data among all sample electrical nodes in the sample power distribution system and sample attribute characteristic data of all sample electrical nodes in a preset sample period;
And based on the training samples, performing iterative training on the initial graph neural network model, and outputting a trained graph neural network model.
In a third aspect, the application further provides a power distribution system reliability evaluation device. The device comprises:
the acquisition module is used for acquiring basic information data of the target power distribution system;
the determining module is used for determining graph structure data by taking each electrical node in the basic information data as a node in a graph domain, wherein the graph structure data comprises connection relation data among the electrical nodes and attribute characteristic data of the electrical nodes in a preset target period;
the evaluation module is used for inputting the graph structure data into the graph neural network model to obtain reliability data sets of all the electrical nodes in the target power distribution system, wherein the reliability data sets of all the electrical nodes are used for representing the reliability performance of the node level of the target power distribution system, the graph neural network model is obtained by training according to the sample graph structure data corresponding to the sample power distribution system and the reliability data label sets of all the sample electrical nodes, and the sample connection relation data in the sample power distribution system are identical with the connection relation data in the target power distribution system.
In a fourth aspect, the present application also provides a computer device. The computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to realize the reliability evaluation method of the power distribution system provided in the first aspect or the model training method provided in the second aspect.
In a fourth aspect, the present application also provides a computer-readable storage medium. The computer readable storage medium has stored thereon a computer program which, when executed by a processor, implements the power distribution system reliability assessment method provided in the first aspect or the model training method provided in the second aspect.
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 power distribution system reliability assessment method provided in the first aspect or the model training method provided in the second aspect.
In the power distribution system reliability evaluation method, the basic information data of the target power distribution system is acquired, each electrical node in the basic information data is taken as a node in a graph domain, graph structure data is determined, the graph structure data comprises connection relation data among the electrical nodes and attribute characteristic data of the electrical nodes in a preset target period, the graph structure data is input into a graph neural network model to obtain a reliability data set corresponding to each electrical node in the target power distribution system, the reliability data set corresponding to each electrical node is used for representing the reliability performance of the node level of the target power distribution system, the graph neural network model is obtained through training according to sample graph structure data corresponding to the sample power distribution system and reliability data label sets of the electrical nodes, and the sample connection relation data in the sample power distribution system is identical to the connection relation data in the target power distribution system. By adopting the method, the sample power distribution system with the same connection relation, namely structure, with the electrical nodes of the target power distribution system is selected, sample connection relation data among the electrical nodes of the samples of the sample power distribution system, sample attribute characteristic data of the electrical nodes in a preset target period and reliability data label sets of the electrical nodes of the samples are used as training samples, the training samples are used for repeated iterative training to obtain a graph neural network model, and the reliability data sets corresponding to the electrical nodes are determined through the graph neural network model, so that the reliability performance of the node level of the target power distribution system is evaluated. The structural connection relation of the power distribution system, the attribute characteristics comprising load characteristics and equipment element parameters and the association relation among reliability attributes can be rapidly extracted by utilizing the graph neural network, the dimension disaster is avoided, and the reliability data set corresponding to each electrical node is rapidly determined for rapid evaluation; the reliability data of each electrical node is predicted by combining the connection relation and attribute characteristics of each electrical node of the power distribution system, the reliability performance and granularity of the node level are evaluated, moreover, the consistency of the connection relation, namely the structure, of the electrical nodes of the sample power distribution system and the target power distribution system is utilized, the same training sample corresponding to the sample power distribution system is utilized for repeated iterative training, one training sample comprises the structure and data association of each sample electrical node, the association can be fully mined, the structure and the data of each electrical node are learned, the granularity of the reliability data of each electrical node is predicted, the dimension disaster is avoided while the calculation accuracy is ensured, the calculation speed is high, the granularity of the calculation result is fine, and the reference and the basis are provided for corresponding decision development of planners and operators.
Drawings
FIG. 1 is a diagram of an application environment for a method of evaluating reliability of a power distribution system in one embodiment;
FIG. 2 is a flow diagram of a method of evaluating reliability of a power distribution system in one embodiment;
FIG. 3 is a schematic diagram of a topology construction graph adjacency matrix in one embodiment;
FIG. 4 is a schematic diagram of a neural network model architecture of one embodiment;
FIG. 5 is a flow chart of a model training method in another embodiment;
FIG. 6 is a block diagram of a power distribution system reliability assessment device in one embodiment;
fig. 7 is an internal structural diagram of a computer device in one embodiment.
Detailed Description
The present application will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present application more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.
The power distribution system reliability evaluation method 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 that the server 104 needs to process. The data storage system may be integrated on the server 104 or may be located on a cloud or other network server. The data storage system may store basic information data of the power distribution system. 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 evaluating reliability of a power distribution system is provided, and the method is applied to the server in fig. 1 for illustration, and includes the following steps:
step 202, obtaining basic information data of a target power distribution system.
The target power distribution system refers to a power distribution system with reliability to be evaluated.
The power distribution system is an electrical system which is used for distributing electric energy to each user through transformation and distribution of a transformer substation after the electric energy is transmitted to an alternating current power grid end by the power transmission system. The embodiment of the application firstly acquires basic information data of a target power distribution system.
For example, the underlying information data of the target power distribution system may include a topology of the target power distribution system, which may include electrical nodes, which may refer to nodes at both ends of the feeder line, including load points and non-load points, and which may also include equipment elements distributed over the line.
As another example, the base information data of the target power distribution system may further include load information data within a predetermined target period, which may include data of a user type, a number of users, location information, peak load, average load, and the like of each load point.
As another example, the basic information data of the target power distribution system may further include parameter information data of equipment elements within a predetermined target period, the equipment elements may include circuit breakers, distribution transformers, disconnectors, tie switches, fuses, converters, and feeder lines, the parameter information data may include reliability parameters (failure rate, time to repair failure) of the above-described equipment elements, position information, parameter information data about switching elements such as circuit breakers, disconnectors, and the like may further include reliable action probabilities, operation times, and parameter information data about feeder lines may further include types, lengths, and the like.
The embodiment of the application is not limited to the acquisition mode of the basic information data, for example, the basic information data can be original or existing data which are directly acquired, or the basic information data can also be data obtained by arranging the original or existing data.
For example, raw or existing data of a target power distribution system is collected directly from a power system platform as basic information data.
Or after the original or existing data of the target power distribution system are collected from the power system platform, the original or existing data are subjected to data preprocessing, including data cleaning, missing value compensation and the like, so that basic information data are obtained.
And 204, determining the graph structure data by taking each electrical node in the basic information data as a node in the graph domain.
Wherein the graph structure data includes connection relation data between the electrical nodes and attribute characteristic data of the electrical nodes within a predetermined target period.
In the embodiment of the application, the reliability of the target power distribution system is evaluated by adopting the graph neural network, and in order to enable the data information to be suitable for the prediction of the graph neural network, the electrical nodes of the target power distribution system are mapped into the nodes of the graph domain, and the lines between the electrical nodes are mapped into the edges of the graph domain.
For example, connection relationship data between electrical nodes may be extracted from a topology of the target power distribution system base information data to describe connection relationships of nodes and edges in a graph domain.
Further, attribute characteristic data of each electrical node in a predetermined target period can be extracted from the target power distribution system basic information data, and the attribute characteristic data is used for describing the attribute characteristics of the nodes in the graph domain. Wherein the attribute characteristic data may include device element characteristic data and load characteristic data.
For example, load characteristic data of each electrical node within a predetermined target period is extracted from the load information data.
For another example, device element characteristic data of each electrical node within a predetermined target period is extracted from parameter information data of the device element.
The electrical node is divided into a load point and a non-load point, the load characteristic data of the non-load point may be zero, and the load characteristic data of the load point may be a specific value extracted from the load information data.
In addition, regarding the equipment element characteristic data, considering that the power distribution system generally adopts a closed-loop design and open-loop operation, the structure of the power distribution system is radial, and the embodiment of the application stores the parameter information data of the equipment element in the equipment element characteristic data of the adjacent electrical nodes.
For example, the device elements distributed on the feeder line may have their parameter information data selectively stored on the head or end electrical node of the feeder line in which they are located.
For another example, if a load branch on which an electrical node is located is a load point has a distribution transformer, the equipment component characteristic data for that node may include parameter information data for that distribution transformer.
And 206, inputting the graph structure data into a graph neural network model to obtain a reliability data set corresponding to each electrical node in the target power distribution system.
In the embodiment of the application, after the graph structure data for evaluating the reliability of the target power distribution system is determined, the graph structure data is input into a graph neural network model, and the reliability of the target power distribution system is predicted through the graph neural network model.
The graph neural network model is obtained through training according to sample graph structure data corresponding to a sample power distribution system and reliability data label sets of all sample electrical nodes.
It should be noted that, because the graph neural network model processes the data as a graph structure, it is different from the traditional neural network based on the assumption that the samples are independent, and the training set, the verification set and the test set can be divided by the traditional data set more randomly. The graph structure data often has interdependent structure and information, so in the training process of the embodiment of the application, the initial graph neural network model is input with sample graph structure data, namely the structure and data of the whole graph including sample connection relation data and sample attribute characteristic data.
In addition, the reliability prediction data of each sample electrical node is obtained by predicting the initial graph neural network model according to the sample graph structure data, and the parameter optimization is carried out on the initial graph neural network model according to the reliability prediction data and the reliability data label set of each sample electrical node. The sample graph structure data corresponding to the sample distribution system and the reliability data label set corresponding to each sample electrical node are used as a training sample, the training sample is used for carrying out multiple iterations, namely, the training sample input in each iteration is the same sample, the reliability prediction data set corresponding to each sample electrical node is output in each iteration, the association among the connection structure, the load characteristic, the equipment element parameter characteristic and the reliability data corresponding to each sample electrical node in the training sample can be fully utilized, the initial graph neural network model can be quickly trained, and accordingly, the reliability evaluation can be quickly carried out on the trained graph neural network model through the graph neural network model.
Further, the sample connection relation data in the sample power distribution system is the same as the connection relation data in the target power distribution system.
For example, the sample power distribution system and the target power distribution system may be two different power distribution systems, but the connection relationship between the electrical nodes is the same. The initial graph neural network model is trained by the same sample connection relationship data as the connection relationship data, sample attribute feature data within a predetermined sample period, and a reliability data tag set, wherein the predetermined sample period and the predetermined target period may be the same, or the predetermined sample period and the predetermined target period may be different.
For another example, the sample power distribution system and the target power distribution system may be the same power distribution system, and the initial graph neural network model is trained by the same sample connection relationship data as the connection relationship data, sample attribute feature data within a predetermined sample period, and a reliability data tag set, wherein the predetermined sample period and the predetermined target period are different.
Assuming that the failure rate of a certain equipment element N in sample attribute feature data corresponding to the training sample is X, when the graph neural network model is applied after training, aiming at the same power distribution system or the same structural connection relation, if the failure rate of a certain equipment element N in the attribute feature data is Y, compared with the training sample, the connection relation data of the target power distribution system is unchanged, and the attribute feature data in a preset target period is changed.
In the embodiment of the application, the topology structure and the load access position of the power distribution system are not easy to change in actual conditions, but the attribute characteristics corresponding to each electrical node are changed, for example, the failure rate of equipment elements in the attribute characteristics is changed along with the change of the service life and the running environment of the equipment. When the sample power distribution system and the target power distribution system are the same power distribution system, training samples can be built by using attribute characteristic data of the power distribution system in a preset sample period for training, and then the reliability evaluation is carried out on the power distribution system in the preset target period by using a trained graph neural network model.
According to the embodiment of the application, according to the trained graph neural network model, graph structure data corresponding to a target power distribution system is input, and a reliability data set corresponding to each electrical node in the target power distribution system is obtained, wherein the reliability data set is used for representing the reliability performance of the node level of the target power distribution system.
For example, the reliability data set includes reliability data for each electrical node, which may include failure rate, average power down time of failure, and the like.
According to the reliability evaluation method for the power distribution system, basic information data of the target power distribution system is obtained, all electric nodes in the basic information data are used as nodes in a graph domain, graph structure data are determined, the graph structure data comprise connection relation data among all the electric nodes and attribute characteristic data of all the electric nodes within a preset target period, the graph structure data are input into a graph neural network model, a reliability data set corresponding to all the electric nodes in the target power distribution system is obtained, wherein the reliability data set corresponding to all the electric nodes is used for representing reliability performance of a node level of the target power distribution system, the graph neural network model is obtained through training according to sample graph structure data corresponding to a sample power distribution system and reliability data label sets of all the sample electric nodes, and the sample connection relation data in the sample power distribution system are identical to the connection relation data in the target power distribution system. By adopting the method, the sample power distribution system with the same connection relation, namely structure, with the electrical nodes of the target power distribution system is selected, sample connection relation data among the electrical nodes of the samples of the sample power distribution system, sample attribute characteristic data of the electrical nodes in a preset target period and reliability data label sets of the electrical nodes of the samples are used as training samples, the training samples are used for repeated iterative training to obtain a graph neural network model, and the reliability data sets corresponding to the electrical nodes are determined through the graph neural network model, so that the reliability performance of the node level of the target power distribution system is evaluated. The structural connection relation of the power distribution system, the attribute characteristics comprising load characteristics and equipment element parameters and the association relation among reliability attributes can be rapidly extracted by utilizing the graph neural network, the dimension disaster is avoided, and the reliability data set corresponding to each electrical node is rapidly determined for rapid evaluation; the reliability data of each electrical node is predicted by combining the connection relation and attribute characteristics of each electrical node of the power distribution system, the reliability performance and granularity of the node level are evaluated, moreover, the consistency of the connection relation, namely the structure, of the electrical nodes of the sample power distribution system and the target power distribution system is utilized, the same training sample corresponding to the sample power distribution system is utilized for repeated iterative training, one training sample comprises the structure and data association of each sample electrical node, the association can be fully mined, the structure and the data of each electrical node are learned, the granularity of the reliability data of each electrical node is predicted, the dimension disaster is avoided while the calculation accuracy is ensured, the calculation speed is high, the granularity of the calculation result is fine, and the reference and the basis are provided for corresponding decision development of planners and operators.
In one embodiment, the attribute characteristic data includes type characteristic data corresponding to each electrical node, the type characteristic data being used to characterize whether the electrical node is a load point. In the reliability data label set of each sample electrical node, the reliability data label of the sample electrical node corresponding to the non-load point is zero.
The electrical node comprises a load point and a non-load point, and the reliability of the power distribution system is evaluated through the reliability data of the load point. The reliability data of the load points may include failure rate, annual average power failure time and failure average power failure duration of each load point, and in the reliability data tag set of each sample electrical node, the failure rate, annual average power failure time and failure average power failure duration of the electrical node corresponding to the load point are used as reliability data tags, and the reliability data tag of the sample electrical node corresponding to the non-load point is zero.
Accordingly, during training, the sample attribute feature data may further include type feature data, where the type feature data is used to characterize whether the sample electrical node is a load point, and during evaluating the target power distribution system, the attribute feature data may also include type feature data, where the type feature data is used to characterize whether the electrical node is a load point.
The attribute characteristic data of the embodiment includes type characteristic data corresponding to each electrical node, the type characteristic data is used for representing whether the electrical node is a load point, in the reliability data label set of each sample electrical node, the reliability data label of the sample electrical node corresponding to a non-load point is zero, the electrical node corresponding to the load point can be more focused, and the reliability data of the electrical node corresponding to the load point can be obtained more quickly by combining the connection relation between the load point and the non-load point, the attribute characteristic and the reliability data label of the load point.
In one embodiment, the training process of the graph neural network model includes:
and A1, acquiring a training sample.
The training samples comprise corresponding records between sample graph structure data and reliability data tag sets of the sample power distribution system.
The embodiment of the application firstly acquires a topological structure of a sample power distribution system, load information data in a preset sample period, parameter information data of equipment elements in the preset sample period and historical power outage information in the preset sample period, wherein the historical power outage information comprises fault rate, annual average power outage time and fault average power outage duration time of each load point.
In order to make the data information suitable for the graphic neural network training, the sample graphic structure data suitable for the graphic neural network training is constructed according to the standard data format. Specifically, each sample electrical node of the sample power distribution system is mapped to a node of a graph domain, a line between the nodes is mapped to an edge of the graph domain, and sample connection relation data between each sample electrical node and sample attribute characteristic data of each sample electrical node in a preset sample period are extracted as sample graph structure data.
The sample attribute feature data includes sample load feature data, sample device element feature data, and sample type feature data. Specifically, the sample electrical node is divided into a load point and a non-load point, sample load characteristic data of the non-load point can be taken as zero, and the sample load characteristic data of the load point comprises specific values extracted from load information data in a preset sample period; parameter information data of the device elements within a predetermined sample period is stored in sample device element characteristic data of adjacent sample electrical nodes; the sample type characteristic data is used for representing whether the sample electrical node is a load point or not, and can be coded in binary, wherein 0 represents a non-load point and 1 represents a load point.
In one implementation, a sample distribution system topology is extracted to construct a sample graph adjacency matrix that is used to represent sample connection relationship data between sample electrical nodes. An example of constructing a graph adjacency matrix from a topology is shown in FIG. 3, where the distribution system, without considering the internal component information, can be abstracted into a graph of vertices and edges, where the left side of FIG. 3 is a simple distribution system topology with vertices representing electrical nodes and edges representing overhead lines or cabling, with a total of 7 vertices and 6 edges, the graphIn 3In order to construct a graph adjacency matrix from the topological graph of the power distribution system, the values of elements in the graph adjacency matrix represent the connection relation between the electrical nodes corresponding to the element row numbers and the electrical nodes corresponding to the element column numbers, wherein 1 represents that the connection exists between two nodes, and 0 represents that the connection does not exist between the two nodes.
In one implementation, sample load feature data, sample device element feature data, and sample type feature data for each sample electrical node are employed as input feature vectors for each sample electrical node to construct a sample node feature matrix (for non-numerical data, such as switch type, etc., using a one-hot code characterization) that is used to represent sample attribute feature data for each sample electrical node over a predetermined sample period. The sample node characteristic matrix can be adopted Representation, i.e. comprisingNOf individual nodesdThe feature is input in dimensions. And determining sample graph structure data by the sample graph adjacent matrix and the sample node characteristic matrix, and taking the sample graph structure data as input of a follow-up initial graph neural network model to effectively combine a topological connection structure and corresponding electrical information.
The reliability data label set of the embodiment of the application comprises reliability data labels of all sample electrical nodes, wherein the reliability data (fault rate, annual average power failure time and fault average power failure duration time) of the sample electrical nodes which are load points are used as labels of the nodes, and the reliability data labels of the sample electrical nodes which are non-load points are 0 in value, so that the reliability data is 0.
And step A2, based on the training sample, performing iterative training on the initial graph neural network model, and outputting the trained graph neural network model.
According to the embodiment of the application, an initial graph neural network model is initialized according to standard graph structure data, namely, a basic structure of the graph neural network model to be trained is built.
The graph neural network model may include an input layer, a feature layer, a full connection layer and an output layer, where the input layer provides input data of the graph neural network model, the input layer includes a graph adjacency matrix and a node feature matrix, the feature layer is responsible for extracting features, the graph neural network model is composed of multiple graph neural network layers, each graph neural network layer uses node feature vectors of a node and its neighboring nodes to perform weighted summation, feature extraction of local information is achieved, as the number of network layers deepens, learning capacity grows along with the deepening, information feature extraction of the whole network is finally completed, the full connection layer is located behind the feature layer, output of the feature layer is used as input of the self, feature extraction of global information is achieved, and the output layer is responsible for specific regression tasks and outputs final prediction results.
Referring to fig. 4, in the neural network model architecture, the type of the feature layer may adopt a graph roll-up network GCN, a graph annotation network GAT, or a graph sage algorithm, and the layer number may be freely set according to actual requirements.
After initializing an initial graph neural network model, performing iterative training on the initial graph neural network model based on a training sample, wherein in each iterative training process, the following operations are performed: and inputting the sample graph structure data into an initial graph neural network model, obtaining a reliability prediction data set corresponding to each sample electrical node, determining a loss value through the reliability prediction data and the reliability data labels of the sample electrical nodes corresponding to the load points in the reliability prediction data set, and adopting the loss value to carry out parameter adjustment.
For example, the loss function may select a Mean Square Error (MSE), where the mean square error refers to an expected value of the parameter estimation value and the parameter true value error squared, and the specific calculation formula is as follows:
；
wherein, the liquid crystal display device comprises a liquid crystal display device,F loss the smaller the model loss value is, the higher the accuracy of the prediction model description experimental data is;nis negativeThe number of load points;y i is the load pointiThe reliability data labels of (a) are parameters true values (such as the true fault rate of the load point, the true year average power failure time and the true fault average power failure duration); The reliability prediction data of the load point i is the parameter estimation value; the regression prediction task of the model is considered to obtain the reliability data of the load points, so that the function does not calculate the loss for all the electrical nodes, but calculates the loss for the nodes of the load points, the model can pay more attention to the calculation result of the load nodes, and finally the reliability index effect calculated at the load nodes is better.
And after determining the loss value, selecting an optimizer, calculating the initial graph neural network model by adopting a back propagation algorithm to obtain a parameter gradient value, and updating parameters of the initial graph neural network model.
In the embodiment of the application, in the process of carrying out iterative training on the initial graph neural network model and updating parameters, the same training sample, namely the corresponding record between the sample graph structure data and the reliability data label set, is always adopted to carry out iterative training on the initial graph neural network model.
According to the embodiment, the training samples are obtained, the training samples comprise the corresponding records between the sample graph structure data and the reliability data label sets of the sample power distribution system, the initial graph neural network model is subjected to iterative training based on the training samples, the trained graph neural network model is output, the sample connection relation data among all sample electrical nodes of the sample power distribution system, the sample attribute characteristic data of all electrical nodes in a preset target period and the association relation among the reliability data label sets of all sample electrical nodes can be learned, and specific design is conducted on the loss function in the training process, so that the model is more in accordance with the prediction task requirements, the calculation result of the load nodes is more focused, and the reliability data effect calculated at the load nodes is better.
In one embodiment, based on the training samples, iteratively training the initial graph neural network model to output a trained graph neural network model, further comprising:
and B1, determining the prediction accuracy of the initial graph neural network model through reliability prediction data and reliability data labels of the sample electrical nodes corresponding to the load points in the reliability prediction data set in each iterative training process.
The accuracy of the prediction of the initial graph neural network model can be expressed as:
；
wherein, the liquid crystal display device comprises a liquid crystal display device,F acc the prediction accuracy, namely a model accuracy value, is obtained by taking an average value of the prediction accuracy of all load nodes;nthe number of the load nodes;y i is a load nodeiReliability data labels of (a), i.e. parameter true values;is a load nodeiReliability prediction data of (a), i.e. parameter estimation values; similarly, the function does not calculate the precision for all nodes, but for the nodes that are load points, and finally averages.
And B2, outputting a trained graphic neural network model under the condition that the accuracy rate reaches a preset threshold value.
In addition, the embodiment of the application can stop training when the iteration times reach the preset maximum times, and output the trained graph neural network model.
According to the embodiment, the accuracy of prediction of the initial graph neural network model is determined through the reliability prediction data and the reliability data labels of the sample electrical nodes corresponding to the load points in the reliability prediction data set, the trained graph neural network model is output under the condition that the accuracy reaches a preset threshold value, the model precision can be calculated for the nodes serving as the load points, and when the updated accuracy of the initial graph neural network model reaches the requirement, the model precision is output as the trained graph neural network model.
In one embodiment, determining the graph structure data with each electrical node in the base information data as a node in the graph domain includes:
and C1, constructing a graph adjacency matrix through connection relation data among the electric nodes.
And C2, constructing a node characteristic matrix through attribute characteristic data of each electrical node in a preset target period.
And step C3, combining the graph adjacent matrix and the node characteristic matrix to obtain graph structure data.
According to the embodiment, the graph adjacent matrix is constructed through connection relation data among the electrical nodes, the attribute characteristic data of the electrical nodes in a preset target period are constructed, the node characteristic matrix is constructed, the graph adjacent matrix and the node characteristic matrix are combined to obtain graph structure data, and the step of obtaining sample graph structure data in a training sample is similar to that of obtaining the sample graph structure data, so that a standard data format can be obtained, and the graph neural network model is suitable for a trained graph neural network model.
In one embodiment, the power distribution system reliability assessment method further comprises:
and D1, extracting the reliability data of the electrical nodes corresponding to all the load points from the reliability data set corresponding to each electrical node.
And D2, determining reliability data of a system in the target power distribution system based on the reliability data of the electrical node corresponding to the load point, wherein the reliability data of the system is used for representing the reliability performance of the system level of the target power distribution system.
The reliability data of the electrical node corresponding to the load point output by the graph neural network model can comprise fault rate, annual average power failure time and fault average power failure duration time of each load point. Acquiring system-level reliability data according to all load point reliability data, wherein the selection and calculation modes of the system-level reliability data are as follows:
the average outage frequency index (system average interruption frequency index, SAIFI) of the system is recorded as：
；
In the method, in the process of the invention,nthe number of load points;is the load pointiIs a failure rate of (1);N i is the load pointiIs a number of users of the system.
The average power outage duration index (system average interruption duration index, SAIDI) of the system is recorded as：
；
In the method, in the process of the invention, U i Is the load pointiAverage power outage time in years.
The average power outage duration index (customer average interruption duration index, CAIDI) of the user is recorded as：
；
Average power availability indicator (average service availability index, ASAI), noted as：
；
According to the embodiment, the reliability data of the system in the target power distribution system is determined by extracting the reliability data of all the electrical nodes corresponding to the load points from the reliability data set corresponding to the electrical nodes, and the reliability data of the system is used for representing the reliability performance of the system level of the target power distribution system based on the reliability data of the electrical nodes corresponding to the load points, so that the reliability data of the system level of the target power distribution system can be quickly obtained.
Based on the same inventive concept, the embodiment of the application also provides a model training method. The implementation of the solution to the problem provided by the model training method is similar to that described in the above method, so the specific limitation in one or more embodiments of the model training method provided below may be referred to the limitation of the reliability evaluation method of the power distribution system hereinabove, and will not be described herein.
In one embodiment, as shown in FIG. 5, a model training method is provided, comprising:
Step 502, obtaining a training sample, where the training sample includes a record of correspondence between sample graph structure data of a sample power distribution system and a reliability data tag set of each sample electrical node, and the sample graph structure data includes connection relationship data between each sample electrical node in the sample power distribution system and sample attribute feature data of each sample electrical node in a predetermined sample period.
And step 504, based on the training samples, performing iterative training on the initial graph neural network model, and outputting a trained graph neural network model.
It should be understood that, although the steps in the flowcharts related to the embodiments described above 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 distribution system reliability evaluation device for realizing the power distribution system reliability evaluation method. The implementation of the solution provided by the device is similar to the implementation described in the above method, so the specific limitation in the embodiments of the reliability evaluation device for a power distribution system provided below may be referred to the limitation of the reliability evaluation method for a power distribution system hereinabove, and will not be repeated herein.
In one embodiment, as shown in fig. 6, there is provided a power distribution system reliability evaluation apparatus including: an acquisition module 602, a determination module 604, and an evaluation module 606, wherein:
and the acquisition module 602 is used for acquiring basic information data of the target power distribution system.
A determining module 604, configured to determine graph structure data with each electrical node in the basic information data as a node in a graph domain, where the graph structure data includes connection relationship data between each electrical node and attribute feature data of each electrical node within a predetermined target period.
The evaluation module 606 is configured to input the graph structure data into a graph neural network model to obtain reliability data sets of each electrical node in the target power distribution system, where the reliability data sets of each electrical node are used to characterize reliability performance of a node level of the target power distribution system, and the graph neural network model is obtained by training according to the sample graph structure data corresponding to the sample power distribution system and the reliability data label sets of each sample electrical node, and sample connection relationship data in the sample power distribution system is the same as connection relationship data in the target power distribution system.
In one embodiment, the attribute feature data includes type feature data corresponding to each electrical node, the type feature data being used to characterize whether the electrical node is a load point; in the reliability data label set of each sample electrical node, the reliability data label of the sample electrical node corresponding to the non-load point is zero.
In one embodiment, the training process of the graph neural network model in the evaluation module 606 includes: acquiring a training sample, wherein the training sample comprises corresponding records between sample graph structure data and a reliability data tag set; based on the training sample, carrying out iterative training on the initial graph neural network model, and outputting a trained graph neural network model; wherein, in each iterative training process, the following operations are performed: and inputting the sample graph structure data into an initial graph neural network model to obtain a reliability prediction data set of each sample electrical node, determining a loss value through the reliability prediction data and the reliability data labels of the sample electrical nodes corresponding to the load points in the reliability prediction data set, and adopting the loss value to carry out parameter adjustment.
In one embodiment, the training process of the graph neural network model in the evaluation module 606 further includes: in each iterative training process, determining the accuracy of prediction of the initial graph neural network model through reliability prediction data and reliability data labels of sample electrical nodes corresponding to load points in the reliability prediction data set; and outputting the trained graphic neural network model under the condition that the accuracy rate reaches a preset threshold value.
In one embodiment, the evaluation module 606 is further configured to: extracting reliability data of the electrical nodes corresponding to all load points from the reliability data set corresponding to each electrical node; and determining reliability data of a system in the target power distribution system based on the reliability data of the electrical nodes corresponding to the load points, wherein the reliability data of the system is used for representing the reliability performance of the system level of the target power distribution system.
In one embodiment, the determining module 604, when executing the determining of the graph structure data with each electrical node in the base information data as a node in the graph domain, comprises: constructing an adjacency matrix through connection relation data among the electrical nodes; constructing a node characteristic matrix through attribute characteristic data of each electrical node in a preset target period; and combining the adjacency matrix and the node characteristic matrix to obtain the graph structure data.
The respective modules in the above-described power distribution system reliability evaluation apparatus may be implemented in whole or in part by software, hardware, and combinations 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, an Input/Output interface (I/O) and a communication interface. The processor, the memory and the input/output interface are connected through a system bus, and the communication interface is connected to the system bus through the input/output interface. 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 used for storing basic information data of the power distribution system. The input/output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used for communicating with an external terminal through a network connection. The computer program, when executed by a processor, implements a power distribution system reliability assessment method or model training method.
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 inventive arrangements and is not limiting of the computer device to which the present inventive arrangements may be applied, and that a particular computer device may include more or fewer components than shown, or may combine some of the 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 implementing the steps of the method embodiments described above when the computer program is executed.
In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored which, when executed by a processor, implements the steps of the method embodiments described above.
In an embodiment, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the steps of the method embodiments described above.
It should be noted that, the user information (including but not limited to user equipment information, user personal information, etc.) and the data (including but not limited to data for analysis, stored data, presented data, etc.) related to the present application are information and data authorized by the user or sufficiently authorized by each party, and the collection, use and processing of the related data need to comply with the related laws and regulations and standards of the related country and region.
Those skilled in the art will appreciate that implementing all or part of the above described methods 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 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 (Magnetoresistive Random Access Memory, MRAM), ferroelectric Memory (Ferroelectric Random Access Memory, FRAM), phase change Memory (Phase Change Memory, PCM), graphene Memory, and the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, and the like. By way of illustration, and not limitation, RAM can be in the form of a variety of forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), and the like. The databases referred to in the embodiments provided herein may include at least one of a relational database and a non-relational database. The non-relational database may include, but is not limited to, a blockchain-based distributed database, and the like. The processor referred to in the embodiments provided in the present application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, or the like, but is not 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 illustrate only a few embodiments of the application and are described in detail herein without thereby limiting the scope of the application. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the application, which are all within the scope of the application. Accordingly, the scope of the application should be assessed as that of the appended claims.
Claims (10)
1. A method of evaluating reliability of a power distribution system, the method comprising:
basic information data of a target power distribution system are acquired;
determining graph structure data by taking each electrical node in the basic information data as a node in a graph domain, wherein the graph structure data comprises connection relation data among the electrical nodes and attribute characteristic data of the electrical nodes in a preset target period;
Inputting the graph structure data into a graph neural network model to obtain reliability data sets corresponding to all electrical nodes in the target power distribution system, wherein the reliability data sets corresponding to all electrical nodes are used for representing the reliability performance of the node level of the target power distribution system, the graph neural network model is obtained through training according to sample graph structure data corresponding to a sample power distribution system and reliability data label sets of all sample electrical nodes, and sample connection relation data in the sample power distribution system are identical with connection relation data in the target power distribution system.
2. The method according to claim 1, wherein the attribute feature data includes type feature data corresponding to each electrical node, the type feature data being used to characterize whether the electrical node is a load point; and in the reliability data label set of each sample electric node, the reliability data label of the sample electric node corresponding to the non-load point is zero.
3. The method of claim 2, wherein the training process of the graph neural network model comprises:
acquiring a training sample, wherein the training sample comprises corresponding records between the sample graph structure data and the reliability data tag set;
Based on the training sample, performing iterative training on the initial graph neural network model, and outputting a trained graph neural network model; wherein, in each iterative training process, the following operations are performed:
and inputting the sample graph structure data into the initial graph neural network model to obtain a reliability prediction data set of each sample electrical node, determining a loss value through the reliability prediction data and the reliability data label of the sample electrical node corresponding to the load point in the reliability prediction data set, and adopting the loss value to carry out parameter adjustment.
4. The method of claim 3, wherein the iteratively training the initial graphical neural network model based on the training samples to output a trained graphical neural network model, further comprising:
in each iterative training process, determining the accuracy of prediction of the initial graph neural network model through reliability prediction data and reliability data labels of sample electrical nodes corresponding to load points in the reliability prediction data set;
and outputting the trained graphic neural network model under the condition that the accuracy rate reaches a preset threshold value.
5. The method according to claim 2, wherein the method further comprises:
extracting reliability data of the electrical nodes corresponding to all load points from the reliability data set corresponding to each electrical node;
and determining reliability data of a system in the target power distribution system based on the reliability data of the electrical node corresponding to the load point, wherein the reliability data of the system is used for representing the reliability performance of the system level of the target power distribution system.
6. The method according to claim 1, wherein determining the graph structure data using each electrical node in the basic information data as a node in a graph domain includes:
constructing an adjacency matrix through the connection relation data among the electrical nodes;
constructing a node characteristic matrix according to the attribute characteristic data of each electrical node in a preset target period;
and combining the adjacency matrix and the node characteristic matrix to obtain graph structure data.
7. A method of model training, the method comprising:
obtaining a training sample, wherein the training sample comprises corresponding records between sample graph structure data of a sample power distribution system and reliability data label sets of all sample electrical nodes, and the sample graph structure data comprises connection relation data among all sample electrical nodes in the sample power distribution system and sample attribute characteristic data of all sample electrical nodes in a preset sample period;
And based on the training sample, performing iterative training on the initial graph neural network model, and outputting a trained graph neural network model.
8. A power distribution system reliability assessment apparatus, the apparatus comprising:
the acquisition module is used for acquiring basic information data of the target power distribution system;
the determining module is used for determining graph structure data by taking each electrical node in the basic information data as a node in a graph domain, wherein the graph structure data comprises connection relation data among the electrical nodes and attribute characteristic data of the electrical nodes in a preset target period;
the evaluation module is used for inputting the graph structure data into a graph neural network model to obtain reliability data sets of all electrical nodes in the target power distribution system, wherein the reliability data sets of all electrical nodes are used for representing the reliability performance of the node level of the target power distribution system, the graph neural network model is obtained through training according to sample graph structure data corresponding to a sample power distribution system and reliability data label sets of all sample electrical nodes, and sample connection relation data in the sample power distribution system are identical to connection relation data in the target power distribution system.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the power distribution system reliability assessment method of any one of claims 1 to 6 or the model training method of claim 7.
10. 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 power distribution system reliability evaluation method of any one of claims 1 to 6 or the model training method of claim 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310899972.8A CN116613754A (en) | 2023-07-21 | 2023-07-21 | Power distribution system reliability assessment method, model training method, device and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310899972.8A CN116613754A (en) | 2023-07-21 | 2023-07-21 | Power distribution system reliability assessment method, model training method, device and equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116613754A true CN116613754A (en) | 2023-08-18 |
Family
ID=87684053
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310899972.8A Pending CN116613754A (en) | 2023-07-21 | 2023-07-21 | Power distribution system reliability assessment method, model training method, device and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116613754A (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109558657A (en) * | 2018-11-20 | 2019-04-02 | 国家电网有限公司 | Powering mode reliability calculation method, device, storage medium and calculating equipment |
| CN111461392A (en) * | 2020-01-23 | 2020-07-28 | 华中科技大学 | Power failure prediction method and system based on graph neural network |
| CN114330670A (en) * | 2022-01-04 | 2022-04-12 | 京东科技信息技术有限公司 | Graph neural network training method, device, equipment and storage medium |
| CN114548762A (en) * | 2022-02-22 | 2022-05-27 | 浙江大学 | Real-time power system cascading failure risk assessment method and system based on space-time diagram neural network |
| US20220383127A1 (en) * | 2021-06-01 | 2022-12-01 | Basmah ALTAF | Methods and systems for training a graph neural network using supervised contrastive learning |
| CN116207739A (en) * | 2023-05-04 | 2023-06-02 | 南方电网数字电网研究院有限公司 | Optimal scheduling method and device for power distribution network, computer equipment and storage medium |
-
2023
- 2023-07-21 CN CN202310899972.8A patent/CN116613754A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109558657A (en) * | 2018-11-20 | 2019-04-02 | 国家电网有限公司 | Powering mode reliability calculation method, device, storage medium and calculating equipment |
| CN111461392A (en) * | 2020-01-23 | 2020-07-28 | 华中科技大学 | Power failure prediction method and system based on graph neural network |
| US20220383127A1 (en) * | 2021-06-01 | 2022-12-01 | Basmah ALTAF | Methods and systems for training a graph neural network using supervised contrastive learning |
| CN114330670A (en) * | 2022-01-04 | 2022-04-12 | 京东科技信息技术有限公司 | Graph neural network training method, device, equipment and storage medium |
| CN114548762A (en) * | 2022-02-22 | 2022-05-27 | 浙江大学 | Real-time power system cascading failure risk assessment method and system based on space-time diagram neural network |
| CN116207739A (en) * | 2023-05-04 | 2023-06-02 | 南方电网数字电网研究院有限公司 | Optimal scheduling method and device for power distribution network, computer equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Müller et al. | Genetic algorithm‐based phasor measurement unit placement method considering observability and security criteria | |
| CN114579584B (en) | Data table processing method and device, computer equipment and storage medium | |
| CN110610019A (en) | Dynamics analysis method of Markov jump system with partially unknown transition probability | |
| CN116167581A (en) | Battery demand estimation method and device, scheduling method and computer equipment | |
| Mukherjee et al. | Application of deep learning for power system state forecasting | |
| Mohamed et al. | A review on big data management and decision-making in smart grid | |
| Liu et al. | Short‐term load forecasting based on LSTNet in power system | |
| CN116757465A (en) | Line risk assessment method and device based on double training weight distribution model | |
| Albarakati et al. | Evaluation of the vulnerability in water distribution systems through targeted attacks | |
| CN117674119A (en) | Power grid operation risk assessment method, device, computer equipment and storage medium | |
| CN117458440A (en) | Method and system for predicting generated power load based on association feature fusion | |
| CN115879742A (en) | Power distribution network transformation method and system, electronic equipment and storage medium | |
| CN116613754A (en) | Power distribution system reliability assessment method, model training method, device and equipment | |
| Ramesh et al. | Electrical power system state estimation meter placement—a comparative survey report | |
| CN116191398A (en) | Load prediction method, load prediction device, computer equipment and storage medium | |
| CN115544276B (en) | Metering device knowledge graph construction method and metering device archive checking method | |
| CN116365520B (en) | Power transmission line equipment risk prediction method and device and computer equipment | |
| CN116316616B (en) | Fault processing scheme determining method and device for power distribution network and computer equipment | |
| CN117354842A (en) | Network quality difference root cause positioning method, device, computer equipment and storage medium | |
| CN117192373B (en) | Power battery fault analysis method, device, computer equipment and storage medium | |
| CN117829287A (en) | Method, device, equipment and storage medium for evaluating performance of knowledge graph embedded model | |
| CN117149750A (en) | Data processing method, apparatus, device, storage medium and computer program product | |
| CN117911004A (en) | Method, device, computer equipment and storage medium for determining overhaul information | |
| CN117828496A (en) | Abnormal data identification method, device, computer equipment and storage medium | |
| CN117041015A (en) | Network fault detection method, device, computer equipment and 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 |