CN118017509B - Large-scale power distribution network parallel optimization method based on digital twin space - Google Patents
Large-scale power distribution network parallel optimization method based on digital twin space Download PDFInfo
- Publication number
- CN118017509B CN118017509B CN202410424144.3A CN202410424144A CN118017509B CN 118017509 B CN118017509 B CN 118017509B CN 202410424144 A CN202410424144 A CN 202410424144A CN 118017509 B CN118017509 B CN 118017509B
- Authority
- CN
- China
- Prior art keywords
- load
- power distribution
- distribution network
- parallel
- sets
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000009826 distribution Methods 0.000 title claims abstract description 212
- 238000005457 optimization Methods 0.000 title claims abstract description 112
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000004891 communication Methods 0.000 claims description 44
- 238000004458 analytical method Methods 0.000 claims description 41
- 230000005540 biological transmission Effects 0.000 claims description 39
- 238000010276 construction Methods 0.000 claims description 19
- 230000001052 transient effect Effects 0.000 claims description 15
- 238000012549 training Methods 0.000 claims description 13
- 241000219112 Cucumis Species 0.000 claims description 9
- 235000015510 Cucumis melo subsp melo Nutrition 0.000 claims description 9
- FJJCIZWZNKZHII-UHFFFAOYSA-N [4,6-bis(cyanoamino)-1,3,5-triazin-2-yl]cyanamide Chemical compound N#CNC1=NC(NC#N)=NC(NC#N)=N1 FJJCIZWZNKZHII-UHFFFAOYSA-N 0.000 claims description 9
- 238000013528 artificial neural network Methods 0.000 claims description 5
- 238000013507 mapping Methods 0.000 claims description 4
- 238000012706 support-vector machine Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 abstract description 10
- 230000004044 response Effects 0.000 abstract description 8
- 238000004364 calculation method Methods 0.000 abstract description 6
- 238000004422 calculation algorithm Methods 0.000 description 16
- 238000007726 management method Methods 0.000 description 11
- 238000012545 processing Methods 0.000 description 9
- 230000003993 interaction Effects 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 4
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 238000004140 cleaning Methods 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 238000012795 verification Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- QVFWZNCVPCJQOP-UHFFFAOYSA-N chloralodol Chemical compound CC(O)(C)CC(C)OC(O)C(Cl)(Cl)Cl QVFWZNCVPCJQOP-UHFFFAOYSA-N 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000002068 genetic effect Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 239000013307 optical fiber Substances 0.000 description 2
- 239000002245 particle Substances 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 230000000717 retained effect Effects 0.000 description 2
- 238000002922 simulated annealing Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 238000003860 storage Methods 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000011217 control strategy Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000007599 discharging Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000005538 encapsulation Methods 0.000 description 1
- 238000004146 energy storage Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000013450 outlier detection Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Abstract
The application provides a digital twin space-based large-scale power distribution network parallel optimization method, which relates to the technical field of edge calculation and comprises the following steps: obtaining power distribution network data assets of a target power distribution network, constructing and generating a power distribution network topological structure, performing specialized splitting, performing edge terminal configuration, pre-constructing a parallel optimization cloud end and communicating with an edge autonomous terminal set to obtain a plurality of twin load data, generating a plurality of load autonomous strategy sets, feeding back the load autonomous strategy sets to the parallel optimization cloud end, and finally performing parallel optimization on the power distribution network. The application can solve the problems that the prior art is difficult to effectively process huge data volume and complex power distribution network topological structure, and simultaneously lacks real-time sensing and quick response capability, so that the overall efficiency is low, and particularly the power distribution network optimization efficiency is seriously affected when the emergency is handled, and the operation efficiency, the power supply reliability and the expandability and flexibility of the system of the power distribution network are improved.
Description
Technical Field
The application relates to the technical field of edge calculation, in particular to a large-scale power distribution network parallel optimization method based on a digital twin space.
Background
As power demand increases and the size of distribution networks expands, the operation and maintenance management of distribution network areas becomes increasingly complex. Along with the continuous increase of power demand, the scale of the power distribution network is continuously expanded, the number of related devices is large, the types of the related devices are complex, and the management difficulty is obviously increased.
At present, the traditional power distribution network management method is often based on manual inspection and recording, has low efficiency and is easy to make mistakes, and is difficult to meet the management requirements of a large-scale power distribution network. Traditional power distribution network optimization methods often only focus on a single target, such as reducing loss or improving voltage stability, and are difficult to meet the requirements of multiple optimization targets at the same time. The problems of insufficient power supply, voltage fluctuation and the like possibly occur in the operation process of the power distribution network, and the electricity utilization experience of users is affected.
In summary, in the prior art, it is difficult to effectively process huge data volume and complex power distribution network topology structure, and meanwhile, the real-time sensing and quick response capability is lacking, so that the overall efficiency is low, and particularly, the power distribution network is weak when handling emergency, so that the optimization efficiency of the power distribution network is seriously affected.
Disclosure of Invention
The application aims to provide a large-scale power distribution network parallel optimization method based on a digital twin space, which is used for solving the problems that the prior art is difficult to effectively process huge data volume and complex power distribution network topological structure, and meanwhile, the overall efficiency is low due to the lack of real-time sensing and quick response capability, and particularly, the power distribution network parallel optimization method is weak when dealing with emergency situations, so that the optimization efficiency of the power distribution network is seriously affected.
In view of the above problems, the application provides a digital twin space-based parallel optimization method for a large-scale power distribution network.
In a first aspect, the present application provides a digital twin space-based parallel optimization method for a large-scale power distribution network, the method is implemented by a digital twin space-based parallel optimization system for the large-scale power distribution network, wherein the method comprises: the method comprises the steps of interactively obtaining power distribution network data assets of a target power distribution network, and constructing and generating a power distribution network topological structure according to the power distribution network data assets, wherein the power distribution network topological structure comprises a plurality of nodes and a plurality of power transmission lines; according to the node type and the transmission line connection relation among the nodes, specialized splitting is carried out on the power distribution network topological structure, and a plurality of parallel sub-network structure sets are obtained; edge terminal configuration is carried out on the plurality of parallel sub-network structure sets to obtain a plurality of edge autonomous terminal sets; pre-constructing a parallel optimized cloud, connecting the parallel optimized cloud with the plurality of nodes in the power distribution network topological structure in a communication manner, and connecting the parallel optimized cloud with the plurality of edge autonomous terminal sets in a communication manner to complete the construction of a parallel optimized twin space; in the parallel optimization twin space, the parallel optimization cloud receives real-time load data of the target power distribution network and transmits the real-time load data to a plurality of nodes of the power distribution network topological structure to obtain a plurality of twin load data; the plurality of edge autonomous terminal sets generate a plurality of load autonomous policy sets according to the plurality of twin load data, and feed back the plurality of load autonomous policy sets to the parallel optimized cloud; and the parallel optimization cloud performs parallel optimization of the target power distribution network based on the plurality of load autonomous policy sets.
In a second aspect, the present application also provides a digital twin space-based parallel optimization system for performing the digital twin space-based parallel optimization method for a large-scale power distribution network, wherein the system comprises: the power distribution network topology construction module is used for interactively obtaining power distribution network data assets of a target power distribution network and constructing and generating a power distribution network topology according to the power distribution network data assets, wherein the power distribution network topology comprises a plurality of nodes and a plurality of power transmission lines; the structure splitting module is used for performing specialized splitting on the power distribution network topological structure according to the node type and the transmission line connection relation among the nodes to obtain a plurality of parallel sub-network structure sets; the edge terminal configuration module is used for carrying out edge terminal configuration on the plurality of parallel sub-network structure sets to obtain a plurality of edge autonomous terminal sets; the twin space construction module is used for pre-constructing a parallel optimized cloud, connecting the parallel optimized cloud with the plurality of nodes in the power distribution network topological structure in a communication manner, and connecting the parallel optimized cloud with the plurality of edge autonomous terminal sets in a communication manner to finish the construction of the parallel optimized twin space; the twin load data acquisition module is used for receiving real-time load data of the target power distribution network in the parallel optimization twin space and transmitting the real-time load data to a plurality of nodes of the power distribution network topological structure to acquire a plurality of twin load data; the load autonomous policy set generation module is used for generating a plurality of load autonomous policy sets by the plurality of edge autonomous terminal sets according to the plurality of twin load data and feeding the plurality of load autonomous policy sets back to the parallel optimization cloud; and the parallel optimization module is used for the parallel optimization cloud end to perform parallel optimization of the target power distribution network based on the plurality of load autonomous policy sets.
One or more technical schemes provided by the application have at least the following technical effects or advantages:
Acquiring power distribution network data assets of a target power distribution network through interaction, and constructing a power distribution network topological structure according to the power distribution network data assets, wherein the power distribution network topological structure comprises a plurality of nodes and a plurality of power transmission lines; according to the node type and the transmission line connection relation among the nodes, specialized splitting is carried out on the power distribution network topological structure, and a plurality of parallel sub-network structure sets are obtained; edge terminal configuration is carried out on the plurality of parallel sub-network structure sets to obtain a plurality of edge autonomous terminal sets; pre-constructing a parallel optimized cloud, connecting the parallel optimized cloud with the plurality of nodes in the power distribution network topological structure in a communication manner, and connecting the parallel optimized cloud with the plurality of edge autonomous terminal sets in a communication manner to complete the construction of a parallel optimized twin space; in the parallel optimization twin space, the parallel optimization cloud receives real-time load data of the target power distribution network and transmits the real-time load data to a plurality of nodes of the power distribution network topological structure to obtain a plurality of twin load data; the plurality of edge autonomous terminal sets generate a plurality of load autonomous policy sets according to the plurality of twin load data, and feed back the plurality of load autonomous policy sets to the parallel optimized cloud; and the parallel optimization cloud performs parallel optimization of the target power distribution network based on the plurality of load autonomous policy sets. The method effectively solves the problems that the prior art is difficult to effectively process huge data volume and complex power distribution network topological structure, and meanwhile, the real-time sensing and quick response capability is lacked, so that the overall efficiency is low, and particularly the fatigue is represented when the emergency is handled, thereby seriously affecting the optimization efficiency of the power distribution network, and improving the operation efficiency, the power supply reliability and the expandability and the flexibility of the system of the power distribution network.
The foregoing description is only an overview of the present application, and is intended to be implemented in accordance with the teachings of the present application in order that the same may be more clearly understood and to make the same and other objects, features and advantages of the present application more readily apparent. It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the application or to delineate the scope of the application. Other features of the present application will become apparent from the description that follows.
Drawings
In order to more clearly illustrate the application or the technical solutions of the prior art, the following brief description will be given of the drawings used in the description of the embodiments or the prior art, it being obvious that the drawings in the description below are only exemplary and that other drawings can be obtained from the drawings provided without the inventive effort for a person skilled in the art.
FIG. 1 is a flow diagram of a digital twin space-based parallel optimization method for a large-scale power distribution network;
Fig. 2 is a schematic structural diagram of a parallel optimization system of a large-scale power distribution network based on a digital twin space.
Reference numerals illustrate:
The system comprises a power distribution network topological structure building module 11, a structure splitting module 12, an edge terminal configuration module 13, a twin space building module 14, a twin load data acquisition module 15, a load autonomous policy set generation module 16 and a parallel optimization module 17.
Detailed Description
The application solves the problems that the prior art is difficult to effectively process huge data volume and complex power distribution network topological structure, and meanwhile, the whole efficiency is low due to the lack of real-time sensing and quick response capability, and the power distribution network optimization efficiency is seriously affected by the fact that the whole efficiency is low particularly when the power distribution network is in emergency response, and the running efficiency, the power supply reliability and the expandability and the flexibility of the system of the power distribution network are improved by providing the digital twin space-based large-scale power distribution network parallel optimization method.
In the following, the technical solutions of the present application will be clearly and completely described with reference to the accompanying drawings, and it should be understood that the described embodiments are only some embodiments of the present application, but not all embodiments of the present application, and that the present application is not limited by the exemplary embodiments described herein. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application. It should be further noted that, for convenience of description, only some, but not all of the drawings related to the present application are shown.
Example 1
Referring to fig. 1, the application provides a digital twin space-based parallel optimization method for a large-scale power distribution network, wherein the method is applied to a digital twin space-based parallel optimization system for the large-scale power distribution network, and specifically comprises the following steps:
Step one: the method comprises the steps of interactively obtaining power distribution network data assets of a target power distribution network, and constructing and generating a power distribution network topological structure according to the power distribution network data assets, wherein the power distribution network topological structure comprises a plurality of nodes and a plurality of power transmission lines;
In particular, the data assets of the distribution network are interactively obtained from various sources such as smart meters, sensors, SCADA systems, GIS systems, and the like. These data assets may include, but are not limited to, equipment parameters, operating conditions, operational metrics, and the like. The equipment parameters comprise transformer capacity and line impedance, the operation states comprise voltage, current and power factor, and the operation indexes comprise line loss rate and power supply reliability. The data interaction may be implemented by standardized communication protocols such as IEC 60870-5-104, IEC 61850, etc. or by dedicated data interfaces. And then data preprocessing and data cleaning are performed, including but not limited to data filtering, outlier detection and correction, missing value interpolation and the like. Based on the acquired power distribution network data assets, constructing a topology structure for generating the power distribution network. The topological structure is an abstract representation of the power distribution network, comprises a plurality of nodes and a plurality of power transmission lines, and intuitively displays the structural characteristics and the running state of the power distribution network. Each node represents a key point or equipment in the power distribution network, such as a transformer substation, a switching station and the like, and the power transmission lines are connected with the nodes to form a skeleton of the power distribution network. Various nodes in the power distribution network such as transformers, switches, load points and the like and power transmission lines are shown in the form of a graph. In this topology, the connection relationship between the nodes reflects the actual operation condition of the power distribution network, including the current flow direction, the voltage distribution, and the like. By constructing the topological structure, the whole layout and the running condition of the power distribution network can be clearly known, and powerful support can be provided for subsequent parallel optimization work. For example, in the optimization process, bottleneck and potential risk points of the power distribution network can be analyzed according to the relationship between nodes in the topological structure and the power transmission line, so that a more accurate and effective optimization strategy is formulated.
Step two: according to the node type and the transmission line connection relation among the nodes, specialized splitting is carried out on the power distribution network topological structure, and a plurality of parallel sub-network structure sets are obtained;
Specifically, the type of the nodes in the power distribution network topology is identified. By identifying the type of node, it can be categorized into different functional levels such as source node, transmission node, distribution node, load node, etc. And analyzing the connection relation of the transmission lines among the nodes, wherein the connection relation comprises direct connection and indirect connection. The direct connection means that two nodes are directly connected through a transmission line; the indirect connection means that two nodes are connected with a power transmission line through other nodes. And dividing the power distribution network topological structure into a plurality of parallel sub-networks according to the node types and the connection relation. Each subnetwork should contain nodes of the same type or similar functionality and be tightly connected internally, relatively weak with external connections.
Step three: edge terminal configuration is carried out on the plurality of parallel sub-network structure sets to obtain a plurality of edge autonomous terminal sets;
Specifically, according to characteristics and requirements of the sub-network, a proper edge terminal device is selected. The devices have the functions of data acquisition, processing, communication, control and the like, and can adapt to the severe operating environment of the power distribution network. The selected terminal equipment is configured, including parameters such as sampling frequency, communication protocol, control strategy and the like, so as to ensure that the operating state of the sub-network can be accurately perceived and controlled in real time. And the edge terminal equipment is used for collecting the electric quantity such as voltage, current, power and the like and the non-electric quantity such as temperature, humidity, switching state and the like in the sub-network in real time. And carrying out on-site processing, such as filtering, transformation, compression and the like, on the acquired data so as to extract key information reflecting the operation state of the sub-network. Each edge autonomous terminal set is a relatively independent autonomous terminal, and can make decisions and execute according to local information without depending on control. The distributed autonomous structure is beneficial to reducing communication delay and improving response speed.
Step four: pre-constructing a parallel optimized cloud, connecting the parallel optimized cloud with the plurality of nodes in the power distribution network topological structure in a communication manner, and connecting the parallel optimized cloud with the plurality of edge autonomous terminal sets in a communication manner to complete the construction of a parallel optimized twin space;
Specifically, an infrastructure of the parallel optimization cloud is built, and the infrastructure comprises computing resources, storage resources, network resources and the like, and special parallel computing software, a database management system, an optimization algorithm library and the like are installed and operated. The optimization algorithm library includes, but is not limited to, such as genetic algorithms, particle swarm optimization algorithms, simulated annealing algorithms, and the like. And analyzing the topological structure of the power distribution network, and determining nodes which need to be in communication connection with the cloud. The communication connection between the node and the cloud can be established by adopting a wired or wireless communication mode, and the communication connection depends on the position and communication condition of the node. Preferably, a wired communication mode such as optical fiber, ethernet and the like is used for a scene that the node position is fixed and the communication distance is relatively close. Wireless communication means such as Wi-Fi, 4G/5G, loRa, etc. are used for scenes where the node positions are scattered or the communication distances are long. And formulating a unified communication protocol and a data interaction format. The communication protocol prescribes rules and procedures of communication between the nodes and the cloud, including aspects of data encapsulation, decapsulation, transmission control, error processing and the like. The data interaction format defines the structure and the representation mode of the data, so that the nodes and the cloud can accurately analyze and identify the data sent by each other. The cloud end and a plurality of edge autonomous terminals are connected in parallel through wired communication or wireless communication, and preferably, the wired communication can use optical fiber or Ethernet connection for the situations of short communication distance and large data volume and occasions with higher safety requirements. Wireless communication can use Wi-Fi, 4G/5G, loRa, etc., and is more flexible and used in a widely distributed terminal set or an environment difficult to wire. And (3) formulating a communication protocol and a data interaction format between the edge autonomous terminal set and the cloud, and ensuring the accuracy and consistency of data. And integrating nodes and edge autonomous terminal sets in the cloud and power distribution network topological structures to form a unified parallel optimization twin space. The parallel optimization twin space can be accurately simulated and predicted by a digital twin technology. Meanwhile, the parallel computing technology can be utilized to quickly and accurately optimize decision and cooperatively control the power distribution network.
Step five: in the parallel optimization twin space, the parallel optimization cloud receives real-time load data of the target power distribution network and transmits the real-time load data to a plurality of nodes of the power distribution network topological structure to obtain a plurality of twin load data;
Specifically, in the parallel optimization twin space, a parallel optimization cloud serves as a core for data processing and optimization, and real-time communication and data interaction are carried out on the parallel optimization cloud and a plurality of nodes and an edge autonomous terminal set of a power distribution network topological structure. When the parallel optimization cloud receives real-time load data of the target power distribution network, the real-time load data are transmitted to a plurality of nodes of the power distribution network topological structure to generate a plurality of twin load data. And the parallel optimization cloud receives real-time load data from the target power distribution network through the established communication connection. Such data may include critical information about the power requirements, voltage levels, current levels, etc. of the individual nodes. The cloud performs preprocessing on the data, including data cleaning, format conversion, standardization and the like. According to the distribution network topology structure and the configuration of the communication network, the parallel optimization cloud transmits the processed real-time load data to a plurality of nodes through a proper communication protocol. After receiving the data, the node performs local storage and processing. And generating corresponding twin load data on each node by utilizing the received real-time load data and a local power distribution network model. These data are virtual maps of the actual distribution network loads used to simulate and optimize the operating conditions of the distribution network. The twin load data may include voltage, current, power factor of the node. And the data of each node is synchronized with the cloud end periodically or in real time. The cloud can verify and check the synchronized twin load data, for example, check the rationality of the data and the running state of the power distribution network through methods such as power flow calculation and state estimation.
Step six: the plurality of edge autonomous terminal sets generate a plurality of load autonomous policy sets according to the plurality of twin load data, and feed back the plurality of load autonomous policy sets to the parallel optimized cloud;
Specifically, the edge autonomous terminal set receives twin load data of the power distribution network nodes associated with the edge autonomous terminal set through a communication network. The data are virtual mapping of actual distribution network loads, and key information such as real-time power demands, voltage and current of the nodes is included. The edge autonomous terminal set analyzes the received data, extracts key parameters, and verifies the integrity and accuracy of the data. If there is an anomaly or a loss of data, retransmission may need to be requested or interpolation processing may need to be performed. The edge autonomous terminal set first analyzes local power demand and supply conditions and the current running state of the power distribution network. And running a built-in optimization algorithm to generate a load autonomy strategy by combining the twin load data, the historical data and preset optimization targets such as cost minimization, energy efficiency maximization and the like. These strategies may include adjusting the priority of the load, adjusting the mutual load between nodes, adjusting the charging and discharging strategies of the energy storage system, etc. And evaluating the generated strategies, and selecting the optimal or most suitable strategies to the current requirements to form a load autonomous strategy set. Evaluation criteria may include feasibility of the strategy, economy, impact on the stability of the distribution network, etc. The edge autonomous terminal sets pack the generated load autonomous policy sets and ensure the safety of data transmission by using an encryption technology. And sending the policy set data to the parallel optimization cloud through a reliable communication network. During transmission, acknowledgement and retransmission mechanisms may be employed to ensure reliable transmission of data. And after receiving the load autonomous policy set from each edge autonomous terminal set, the parallel optimization cloud terminal performs decryption, verification and unpacking operations on the data, and ensures the integrity and accuracy of the data. These policy sets are then stored in a data repository in the cloud for use in subsequent global optimization and decision making.
Step seven: and the parallel optimization cloud performs parallel optimization of the target power distribution network based on the plurality of load autonomous policy sets.
Specifically, the parallel optimization cloud firstly gathers load autonomous policy sets from all edge autonomous terminal sets to form a comprehensive policy library. Because different terminal sets may adopt different data formats or units, the policy sets are standardized first, and then data cleaning work is performed, including removing duplicate policies, processing missing values, identifying and correcting abnormal policies, etc. According to the operation targets and optimization requirements of the power distribution network, such as cost minimization, energy efficiency maximization, power supply reliability improvement and the like, the parallel optimization cloud can adopt a proper optimization algorithm to calculate, such as a genetic algorithm, a particle swarm optimization algorithm and a simulated annealing algorithm. And by utilizing high-performance computing resources, computing and evaluating a plurality of optimization scenes or strategy combinations simultaneously, and remarkably improving the efficiency of optimization computing. After a certain time of calculation iteration, a series of global optimization results are generated, wherein the global optimization results comprise an optimal load distribution scheme, an optimal equipment scheduling strategy and the like. Before the optimization results are applied to the actual distribution network, the cloud can verify the feasibility and effectiveness of the results through simulation or other verification means. After executing the control decision issued by the cloud, the edge autonomous terminal set feeds back the execution result and actual operation data to the cloud in real time.
Further, the second step of the present application further comprises:
the historical load curve data of the target power distribution network are obtained interactively;
according to the node type and the transmission line connection relation among the nodes, performing primary specialized splitting on the power distribution network topological structure to obtain a plurality of local sub-network structures;
And carrying out secondary specialized splitting on the plurality of local subnetwork structures according to the historical load curve data to obtain a plurality of parallel subnetwork structure sets.
Specifically, load change conditions of the power distribution network in a past period of time are obtained, and historical load curve data are data of the power distribution network for recording operation load conditions of each node. Historical load data is obtained by interaction with a power distribution network management system or an energy management system. The data exist in time series, i.e., a load curve, showing the load amount at each time point. The acquired data is preprocessed, such as denoising, missing value filling and the like, so as to ensure the data quality. And analyzing the topological structure of the power distribution network, wherein the topological structure comprises nodes such as substations, feeder nodes, load nodes and the like and the connection relation between the nodes. Splitting is performed according to the types and the importance of the nodes, such as hub nodes, common nodes and the like, and the characteristics of capacity, length and the like of the power transmission line. Nodes with relatively close electrical distances and similar load characteristics are divided into the same local subnetwork. Each local area sub-network is ensured to be relatively independent in structure, but can be connected with other sub-networks through connecting lines. The complex distribution network can be split into several smaller, relatively independent local subnetworks for more refined management and optimization. And analyzing historical load curve data of each local subnetwork, and identifying key characteristics such as load mode, peak period, load rate and the like. According to the load characteristics, the local area sub-network is further split into a plurality of parallel sub-networks. These parallel subnetworks may be electrically connected in parallel, but may differ in load characteristics. The load characteristics of each parallel sub-network are ensured to be relatively consistent so as to facilitate the subsequent optimization calculation. And verifying and testing the split parallel sub-network structure set to ensure that the split parallel sub-network structure set can meet the operation requirement and the optimization target of the power distribution network. Through the layering splitting, the complex power distribution network can be converted into a group of relatively simple and easily-managed parallel sub-network structure sets, and a foundation is provided for subsequent parallel optimization calculation.
Further, the application also comprises:
node identification is carried out on the power distribution network topological structure according to the same type of power distribution equipment, and a plurality of same type scattered point sets are obtained;
Carrying out transmission line restoration on the plurality of scattered point sets of the same type according to the power distribution network topological structure to obtain a plurality of groups of local sub-network structures;
And stripping the packet labels of the multiple groups of local area sub-network structures to obtain the multiple groups of local area sub-network structures.
Specifically, traversing the power distribution network topology identifies all types of power distribution equipment, such as transformers, switches, load nodes, and the like. For each type of device, nodes which have no connection relationship are identified and classified into corresponding scattering point sets of the same type. Ensuring that each scatter set contains only the same type of device node. And analyzing the topological structure of the power distribution network, and determining the connection relation among the nodes and parameters such as impedance, capacity and the like of the power transmission line. And for each scattered point set of the same type, restoring a corresponding power transmission line according to the connection relation among the nodes to form a local area sub-network. Each local subnetwork is ensured to remain structurally complete and reflect the electrical characteristics of the corresponding portion of the original distribution network. Each local area subnetwork structure carries corresponding packet labels that identify the category or characteristic to which the subnetwork belongs. Stripping is performed according to the packet labels, and the nodes among the nodes of different types are stripped. And after the packet labels are stripped, obtaining a plurality of pure local area sub-network structures.
Further, the application also comprises:
the historical load curve data includes a plurality of historical node load curve data for the plurality of nodes;
The target power distribution network is interacted to obtain a plurality of rated load ranges of the plurality of nodes, and the plurality of nodes and the plurality of rated load ranges are associated and stored based on a knowledge graph to obtain a rated load inquirer;
according to K nodes of the first local subnetwork, mapping and calling in the historical load curve data to obtain K historical node load curve data;
synchronizing the K nodes to the rated load inquirer, and calling to obtain K rated load ranges;
Carrying out load dynamic balancing melon division on the K nodes according to the K rated load ranges and the K historical node load curve data to obtain a first parallel sub-network structure set;
and by analogy, carrying out secondary specialized splitting on the plurality of local subnetwork structures according to the historical load curve data to obtain the plurality of parallel subnetwork structure sets.
In particular, historical load curve data for a plurality of nodes is obtained from a distribution network management system, and the data reflects the load condition of each node at different time points. And interacting with the power distribution network, and acquiring the rated load range of each node. And (5) storing the nodes and the corresponding rated load ranges in an associated manner by utilizing a knowledge graph technology. A querier such as a database query system or API is constructed to quickly retrieve the rated load information of the node when needed. A local area sub-network, such as a first local area sub-network, is selected and the number of nodes, such as K nodes, it contains is determined. Historical load data of the K nodes are extracted from the historical load curve data. The rated load ranges of the K nodes are obtained using a rated load querier. And applying a load balancing algorithm such as statistical analysis based on historical data, a machine learning algorithm and the like, and carrying out dynamic balancing and melon distribution on the loads of K nodes according to the rated load range of the nodes and the historical load data. For example, the first local area subnetwork includes A, B, C, D nodes, where the load a is higher, the load B is lower, the load C is balanced, and the load D is higher, where A, B, C may be divided into subnetworks, and D may be put back into all nodes for the next round of dynamic melon division. The load distribution of each node is ensured to meet the rated load range, and the load balance in the whole local area sub-network is realized. Repeating the step of dynamically balancing the load for each local sub-network. And splitting each local area sub-network into smaller parallel sub-network structures according to the melon division result. These parallel subnetwork structures may be electrically parallel but more refined in load characteristics and management. Finally, a power distribution network model comprising a plurality of parallel sub-network structure sets is formed, and a foundation is laid for subsequent parallel optimization and management.
Further, the application also comprises:
Traversing the K historical node load curve data correspondingly by adopting the K rated load ranges to obtain K groups of load deviation time stamps, and stripping grouping labels from the K groups of load deviation time stamps to obtain a plurality of load deviation time stamps;
The same-time stamp data call is carried out on the K historical node load curve data according to the load deviation time stamps, so that instantaneous loads of multiple groups of nodes are obtained;
Obtaining a first set of node transient loads from the multi-set node transient load call, wherein the first set of node transient loads includes K node transient loads;
Calculating to obtain K load deviation parameters according to the K rated load ranges and the K node instantaneous loads;
Carrying out load dynamic balancing melon division according to the K load deviation parameters to obtain a first load balancing scheme set, and so on to obtain a plurality of load balancing scheme sets of instantaneous loads of a plurality of groups of nodes;
node combination extraction is carried out on the plurality of load balancing scheme sets, so that a plurality of node combination sets are obtained;
Counting the occurrence frequency of the same combination of the plurality of node combination sets to obtain a plurality of same node combination frequencies;
And reserving a plurality of node combinations meeting a preset frequency threshold in the same node combination frequencies as the first parallel sub-network structure set.
Specifically, for K historical nodes in the distribution network, each node has its load curve data, i.e., load records over time. These load curve data are traversed according to the preset K rated load ranges. The load range may be a range set based on the node design capacity or the load level at normal operation. When the actual load of the node exceeds or falls below its corresponding rated load range, the timestamp of this point in time is recorded. Thus, each node will have a set of time stamps marking the time points on its load curve that deviate from the nominal load range. Next, the load curve data of K history nodes are queried again using these time stamps. The purpose of this query is to obtain the instantaneous load value of each node at these particular time stamps. Since each time stamp corresponds to a specific point in time, the load status of all nodes, i.e. the instantaneous load data of groups of nodes, at these points in time can be derived. From the plurality of sets of node instantaneous load data, a first set is selected as a starting point for analysis. This set of data contains the instantaneous load values of all K nodes at the same time stamp. The instantaneous load values of the K nodes and their corresponding rated load ranges are used to calculate the load deviation parameters. This parameter is a quantitative indicator that measures the degree of difference or deviation between the actual load of the node and its rated load. It may be a simple difference, percentage deviation, or other more complex measure. Using these bias parameters as inputs, the load of each node is dynamically adjusted or redistributed, such as by optimization algorithms, heuristics, etc., so that the load distribution across the system is more balanced. The load distribution schemes thus obtained constitute a first set of load balancing schemes. This set of schemes is calculated for the first set of node transient load data. After processing the first set of node transient load data, we continue to process the remaining sets of node transient load data according to the same flow. And calculating load deviation parameters for each group of data, and then carrying out dynamic balanced distribution of the load to obtain a load balancing scheme set. For example, the first local area subnetwork includes A, B, C, D nodes, where the load a is higher, the load B is lower, the load C is balanced, and the load D is higher, where A, B, C may be divided into a temporary subnetwork, and D and other nodes continue to dynamically split. Finally, a plurality of load balancing scheme sets corresponding to the instantaneous load data of the plurality of groups of nodes are obtained. Multiple sets of load balancing schemes are analyzed to extract node combinations that occur in different schemes. These node combinations represent node groups that need to work cooperatively to achieve load balancing at different points in time and under load conditions. By extracting these combinations, it is possible to know which nodes often act together in maintaining system load balancing. and carrying out statistical analysis on the extracted multiple node combination sets. The frequency of occurrence of each specific node combination in all the load balancing scheme sets is calculated. The node combinations that occur at high frequencies may be groups that can be more heavily loaded with each other. And screening the node combinations according to a preset frequency threshold. Only those node combinations whose frequency of occurrence is greater than or equal to a preset threshold are retained. These retained combinations of nodes form a first set of parallel sub-network structures. For example, A, B, C is more frequent in a sub-network, greater than a predetermined value, the group is reserved, and so on, to obtain a plurality of groups of node groups. This set represents a group of nodes in the network that often need to be loaded together to achieve equilibrium, with significant guiding implications for subsequent network optimization and autonomous control.
Further, the application also comprises:
randomly calling from the first parallel sub-network structure set to obtain a first parallel sub-network structure;
Taking the first parallel sub-network structure as constraint, and calling from the plurality of load balancing scheme sets to obtain a first parallel balancing scheme set;
pre-constructing a load balancing analysis model, and training the load balancing analysis model by adopting the first parallel balancing scheme set;
Pre-constructing a first edge autonomous terminal, synchronizing the trained load balance analysis model to the first edge autonomous terminal, and configuring the first edge autonomous terminal to the first parallel sub-network structure;
And by analogy, carrying out edge terminal configuration on the plurality of parallel sub-network structure sets to obtain the plurality of edge autonomous terminal sets.
Specifically, one parallel sub-network structure is randomly called from the first parallel sub-network structure set, and is named as a first parallel sub-network structure. And searching matched schemes in the plurality of load balancing schemes by taking the first parallel sub-network structure as a constraint condition. A first set of parallel equalization schemes matching the first parallel sub-network structure is invoked. A load balancing analysis model is pre-constructed, which may be based on machine learning or optimization algorithms. The first parallel equalization scheme set is used as training data to train the load equalization analysis model. After training is completed, the model can propose corresponding load balancing suggestions according to the input parallel sub-network structure. An edge autonomous terminal is pre-built and named as a first edge autonomous terminal. And synchronizing the trained load balance analysis model to the first edge autonomous terminal. And configuring the first edge autonomous terminal to a first parallel sub-network structure so that the first edge autonomous terminal can monitor and manage the load balance of the sub-network in real time. Similarly, the above steps are repeated for each parallel subnetwork structure in the set of parallel subnetwork structures. Finally, a plurality of edge autonomous terminal sets are obtained, and terminals in each set are matched with the corresponding parallel sub-network structure and are provided with a trained load balancing analysis model.
Further, the application also comprises:
constructing three load balancing analysis sub-channels based on the BP neural network, the support vector machine and the generation type countermeasure network;
training the three load balancing analysis sub-channels one by adopting the first parallel balancing scheme set until the output precision of the three load balancing analysis sub-channels meets the preset requirement;
pre-constructing an output equalization sub-channel;
And arranging the three load balance analysis sub-channels in parallel, and connecting the output ends of the three load balance analysis sub-channels with the input ends of the output balance sub-channels to complete the construction of the load balance analysis model.
Specifically, the first sub-channel is constructed based on a BP neural network, which is a multi-layer feed-forward network, which is trained by a back propagation algorithm so that it can learn the mapping relationship from input to output. In the load balancing analysis, a BP neural network can be utilized to learn a complex nonlinear relationship between the parallel sub-network structure and the load balancing scheme. The second sub-channel is constructed based on a support vector machine, which is a supervised learning method based on a statistical learning theory and is suitable for classification and regression problems. In the load balancing analysis, the SVM may be used to identify key features of the parallel sub-network structure and classify or regression-predict based on these features to determine an optimal load balancing scheme. The third sub-channel is constructed based on a generative countermeasure network, which is a deep learning model, consisting of two parts, a generator and a arbiter, which generates new data samples through countermeasure training. In the load balancing analysis, GAN can be used to generate simulation data similar to a real parallel subnetwork structure, thereby expanding the training dataset and improving the generalization ability of the model. And training three load balancing analysis sub-channels one by adopting data in the first parallel balancing scheme. The method comprises the steps of adjusting super parameters of each sub-channel, optimizing algorithm selection and the like so as to ensure that the model can effectively learn patterns in data. And monitoring the output precision of each sub-channel in the training process, and adjusting a training strategy or a model structure according to preset requirements. This includes setting early-stop mechanisms, learning rate adjustments, etc. to prevent overfitting and improve generalization performance of the model. The output equalization subchannel may employ various strategies to synthesize the outputs of the three subchannels, such as weighted averaging, voting schemes, stacked integration, and the like. In constructing the output equalization sub-channels, it is also necessary to consider how to handle output differences and conflicts between different sub-channels. Including setting weight adjustment mechanisms, introducing additional verification steps to ensure accuracy and reliability of the final output. The three load balancing analysis sub-channels are arranged in parallel, and the output ends of the three load balancing analysis sub-channels are connected with the input ends of the output balancing sub-channels. Thus, when a parallel sub-network structure is given as input, the three sub-channels will generate respective prediction results, and the output equalization sub-channels are used for comprehensive processing, so as to finally generate a unified load equalization scheme as output.
In summary, the large-scale power distribution network parallel optimization method based on the digital twin space provided by the application has the following technical effects:
Acquiring power distribution network data assets of a target power distribution network through interaction, and constructing a power distribution network topological structure according to the power distribution network data assets, wherein the power distribution network topological structure comprises a plurality of nodes and a plurality of power transmission lines; according to the node type and the transmission line connection relation among the nodes, specialized splitting is carried out on the power distribution network topological structure, and a plurality of parallel sub-network structure sets are obtained; edge terminal configuration is carried out on the plurality of parallel sub-network structure sets to obtain a plurality of edge autonomous terminal sets; pre-constructing a parallel optimized cloud, connecting the parallel optimized cloud with the plurality of nodes in the power distribution network topological structure in a communication manner, and connecting the parallel optimized cloud with the plurality of edge autonomous terminal sets in a communication manner to complete the construction of a parallel optimized twin space; in the parallel optimization twin space, the parallel optimization cloud receives real-time load data of the target power distribution network and transmits the real-time load data to a plurality of nodes of the power distribution network topological structure to obtain a plurality of twin load data; the plurality of edge autonomous terminal sets generate a plurality of load autonomous policy sets according to the plurality of twin load data, and feed back the plurality of load autonomous policy sets to the parallel optimized cloud; and the parallel optimization cloud performs parallel optimization of the target power distribution network based on the plurality of load autonomous policy sets. The method effectively solves the problems that the prior art is difficult to effectively process huge data volume and complex power distribution network topological structure, and meanwhile lacks of real-time sensing and quick response capability, so that overall efficiency is low, and particularly fatigue is shown when emergency is handled, so that the optimization efficiency of the power distribution network is seriously affected, and the operation efficiency, the power supply reliability and the expandability and flexibility of the system of the power distribution network are improved.
Example two
Based on the same inventive concept as the large-scale power distribution network parallel optimization method based on the digital twin space in the foregoing embodiment, the present application further provides a large-scale power distribution network parallel optimization system based on the digital twin space, referring to fig. 2, the system includes:
The power distribution network topology construction module 11 is used for interactively obtaining power distribution network data assets of a target power distribution network and constructing and generating a power distribution network topology according to the power distribution network data assets, wherein the power distribution network topology comprises a plurality of nodes and a plurality of power transmission lines;
The structure splitting module 12 is used for performing specialized splitting on the power distribution network topological structure according to the node type and the transmission line connection relation among the nodes to obtain a plurality of parallel sub-network structure sets;
The edge terminal configuration module 13 is configured to perform edge terminal configuration on the multiple parallel sub-network structure sets, so as to obtain multiple edge autonomous terminal sets;
The twin space construction module 14 is configured to pre-construct a parallel optimized cloud, connect the parallel optimized cloud with the plurality of nodes in the power distribution network topology structure in a communication manner, and connect the parallel optimized cloud with the plurality of edge autonomous terminal sets in a communication manner, so as to complete the construction of the parallel optimized twin space;
the twin load data acquisition module 15 is configured to receive real-time load data of the target power distribution network in the parallel optimization twin space by the parallel optimization cloud and transmit the real-time load data to a plurality of nodes of the power distribution network topology structure, so as to obtain a plurality of twin load data;
the load autonomous policy set generating module 16 is configured to generate a plurality of load autonomous policy sets by using the plurality of edge autonomous terminal sets according to the plurality of twin load data, and feed back the plurality of load autonomous policy sets to the parallel optimized cloud;
And the parallel optimization module 17 is used for the parallel optimization cloud end to perform parallel optimization of the target power distribution network based on the plurality of load autonomous policy sets.
Further, the structure splitting module 12 in the system is further configured to:
the historical load curve data of the target power distribution network are obtained interactively;
according to the node type and the transmission line connection relation among the nodes, performing primary specialized splitting on the power distribution network topological structure to obtain a plurality of local sub-network structures;
And carrying out secondary specialized splitting on the plurality of local subnetwork structures according to the historical load curve data to obtain a plurality of parallel subnetwork structure sets.
Further, the system further comprises a local area sub-network acquisition module, wherein the local area sub-network acquisition module is used for:
node identification is carried out on the power distribution network topological structure according to the same type of power distribution equipment, and a plurality of same type scattered point sets are obtained;
Carrying out transmission line restoration on the plurality of scattered point sets of the same type according to the power distribution network topological structure to obtain a plurality of groups of local sub-network structures;
And stripping the packet labels of the multiple groups of local area sub-network structures to obtain the multiple groups of local area sub-network structures.
Further, the system further includes a first parallel sub-network structure set acquisition module configured to:
the historical load curve data includes a plurality of historical node load curve data for the plurality of nodes;
The target power distribution network is interacted to obtain a plurality of rated load ranges of the plurality of nodes, and the plurality of nodes and the plurality of rated load ranges are associated and stored based on a knowledge graph to obtain a rated load inquirer;
according to K nodes of the first local subnetwork, mapping and calling in the historical load curve data to obtain K historical node load curve data;
synchronizing the K nodes to the rated load inquirer, and calling to obtain K rated load ranges;
Carrying out load dynamic balancing melon division on the K nodes according to the K rated load ranges and the K historical node load curve data to obtain a first parallel sub-network structure set;
and by analogy, carrying out secondary specialized splitting on the plurality of local subnetwork structures according to the historical load curve data to obtain the plurality of parallel subnetwork structure sets.
Further, the system also comprises a frequency statistics module, wherein the frequency statistics module is used for:
Traversing the K historical node load curve data correspondingly by adopting the K rated load ranges to obtain K groups of load deviation time stamps, and stripping grouping labels from the K groups of load deviation time stamps to obtain a plurality of load deviation time stamps;
The same-time stamp data call is carried out on the K historical node load curve data according to the load deviation time stamps, so that instantaneous loads of multiple groups of nodes are obtained;
Obtaining a first set of node transient loads from the multi-set node transient load call, wherein the first set of node transient loads includes K node transient loads;
Calculating to obtain K load deviation parameters according to the K rated load ranges and the K node instantaneous loads;
Carrying out load dynamic balancing melon division according to the K load deviation parameters to obtain a first load balancing scheme set, and so on to obtain a plurality of load balancing scheme sets of instantaneous loads of a plurality of groups of nodes;
node combination extraction is carried out on the plurality of load balancing scheme sets, so that a plurality of node combination sets are obtained;
Counting the occurrence frequency of the same combination of the plurality of node combination sets to obtain a plurality of same node combination frequencies;
And reserving a plurality of node combinations meeting a preset frequency threshold in the same node combination frequencies as the first parallel sub-network structure set.
Further, the system further comprises a plurality of edge autonomous terminal set acquisition modules, wherein the plurality of edge autonomous terminal set acquisition modules are used for:
randomly calling from the first parallel sub-network structure set to obtain a first parallel sub-network structure;
Taking the first parallel sub-network structure as constraint, and calling from the plurality of load balancing scheme sets to obtain a first parallel balancing scheme set;
pre-constructing a load balancing analysis model, and training the load balancing analysis model by adopting the first parallel balancing scheme set;
Pre-constructing a first edge autonomous terminal, synchronizing the trained load balance analysis model to the first edge autonomous terminal, and configuring the first edge autonomous terminal to the first parallel sub-network structure;
And by analogy, carrying out edge terminal configuration on the plurality of parallel sub-network structure sets to obtain the plurality of edge autonomous terminal sets.
Further, the system also comprises a load balancing analysis model construction module, wherein the load balancing analysis model construction module is used for:
constructing three load balancing analysis sub-channels based on the BP neural network, the support vector machine and the generation type countermeasure network;
training the three load balancing analysis sub-channels one by adopting the first parallel balancing scheme set until the output precision of the three load balancing analysis sub-channels meets the preset requirement;
pre-constructing an output equalization sub-channel;
And arranging the three load balance analysis sub-channels in parallel, and connecting the output ends of the three load balance analysis sub-channels with the input ends of the output balance sub-channels to complete the construction of the load balance analysis model.
The embodiments in this specification are described in a progressive manner, and each embodiment focuses on the difference from the other embodiments, and the foregoing digital twin space-based parallel optimization method and specific example for a digital twin space-based parallel optimization system for a large-scale power distribution network in the first embodiment of fig. 1 are also applicable to the digital twin space-based parallel optimization system for a large-scale power distribution network in this embodiment, and by the foregoing detailed description of the digital twin space-based parallel optimization method for a large-scale power distribution network, those skilled in the art can clearly know the digital twin space-based parallel optimization system for a large-scale power distribution network in this embodiment, so that the details of this embodiment are not described herein for brevity of the specification. For the system disclosed in the embodiment, since the system corresponds to the method disclosed in the embodiment, the description is simpler, and the relevant points refer to the description of the method section.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present application without departing from the spirit or scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the present application and the equivalent techniques thereof, the present application is also intended to include such modifications and variations.
Claims (8)
1. The parallel optimization method for the large-scale power distribution network based on the digital twin space is characterized by comprising the following steps of:
the method comprises the steps of interactively obtaining power distribution network data assets of a target power distribution network, and constructing and generating a power distribution network topological structure according to the power distribution network data assets, wherein the power distribution network topological structure comprises a plurality of nodes and a plurality of power transmission lines;
According to the node type and the transmission line connection relation among the nodes, specialized splitting is carried out on the power distribution network topological structure, and a plurality of parallel sub-network structure sets are obtained;
edge terminal configuration is carried out on the plurality of parallel sub-network structure sets to obtain a plurality of edge autonomous terminal sets;
Pre-constructing a parallel optimized cloud, connecting the parallel optimized cloud with the plurality of nodes in the power distribution network topological structure in a communication manner, and connecting the parallel optimized cloud with the plurality of edge autonomous terminal sets in a communication manner to complete the construction of a parallel optimized twin space;
in the parallel optimization twin space, the parallel optimization cloud receives real-time load data of the target power distribution network and transmits the real-time load data to a plurality of nodes of the power distribution network topological structure to obtain a plurality of twin load data;
The plurality of edge autonomous terminal sets generate a plurality of load autonomous policy sets according to the plurality of twin load data, and feed back the plurality of load autonomous policy sets to the parallel optimized cloud;
And the parallel optimization cloud performs parallel optimization of the target power distribution network based on the plurality of load autonomous policy sets.
2. The method of claim 1, wherein the power distribution network topology is split in a specialized manner according to node types and transmission line connection relationships between nodes to obtain a plurality of parallel sub-network structure sets, the method further comprising:
the historical load curve data of the target power distribution network are obtained interactively;
according to the node type and the transmission line connection relation among the nodes, performing primary specialized splitting on the power distribution network topological structure to obtain a plurality of local sub-network structures;
And carrying out secondary specialized splitting on the plurality of local subnetwork structures according to the historical load curve data to obtain a plurality of parallel subnetwork structure sets.
3. The method of claim 2, wherein performing a first-level specialized splitting on the topology of the power distribution network according to the node type and the transmission line connection relationship between the nodes to obtain a plurality of local subnetwork structures, comprising:
node identification is carried out on the power distribution network topological structure according to the same type of power distribution equipment, and a plurality of same type scattered point sets are obtained;
Carrying out transmission line restoration on the plurality of scattered point sets of the same type according to the power distribution network topological structure to obtain a plurality of groups of local sub-network structures;
And stripping the packet labels of the multiple groups of local area sub-network structures to obtain the multiple groups of local area sub-network structures.
4. The method of claim 2, wherein performing a two-level specialized splitting of the plurality of local area subnetwork structures based on the historical load curve data to obtain a plurality of sets of parallel subnetwork structures, comprising:
the historical load curve data includes a plurality of historical node load curve data for the plurality of nodes;
The target power distribution network is interacted to obtain a plurality of rated load ranges of the plurality of nodes, and the plurality of nodes and the plurality of rated load ranges are associated and stored based on a knowledge graph to obtain a rated load inquirer;
according to K nodes of the first local subnetwork, mapping and calling in the historical load curve data to obtain K historical node load curve data;
synchronizing the K nodes to the rated load inquirer, and calling to obtain K rated load ranges;
Carrying out load dynamic balancing melon division on the K nodes according to the K rated load ranges and the K historical node load curve data to obtain a first parallel sub-network structure set;
and by analogy, carrying out secondary specialized splitting on the plurality of local subnetwork structures according to the historical load curve data to obtain the plurality of parallel subnetwork structure sets.
5. The method of claim 4, wherein dynamically balancing loads of the K nodes according to the K rated load ranges and the K historical node load curve data to obtain a first set of parallel sub-network structures, comprising:
Traversing the K historical node load curve data correspondingly by adopting the K rated load ranges to obtain K groups of load deviation time stamps, and stripping grouping labels from the K groups of load deviation time stamps to obtain a plurality of load deviation time stamps;
The same-time stamp data call is carried out on the K historical node load curve data according to the load deviation time stamps, so that instantaneous loads of multiple groups of nodes are obtained;
Obtaining a first set of node transient loads from the multi-set node transient load call, wherein the first set of node transient loads includes K node transient loads;
Calculating to obtain K load deviation parameters according to the K rated load ranges and the K node instantaneous loads;
Carrying out load dynamic balancing melon division according to the K load deviation parameters to obtain a first load balancing scheme set, and so on to obtain a plurality of load balancing scheme sets of instantaneous loads of a plurality of groups of nodes;
node combination extraction is carried out on the plurality of load balancing scheme sets, so that a plurality of node combination sets are obtained;
Counting the occurrence frequency of the same combination of the plurality of node combination sets to obtain a plurality of same node combination frequencies;
And reserving a plurality of node combinations meeting a preset frequency threshold in the same node combination frequencies as the first parallel sub-network structure set.
6. The method of claim 5, wherein configuring the edge terminals for the plurality of parallel sub-network structure sets to obtain a plurality of edge autonomous terminal sets comprises:
randomly calling from the first parallel sub-network structure set to obtain a first parallel sub-network structure;
Taking the first parallel sub-network structure as constraint, and calling from the plurality of load balancing scheme sets to obtain a first parallel balancing scheme set;
pre-constructing a load balancing analysis model, and training the load balancing analysis model by adopting the first parallel balancing scheme set;
Pre-constructing a first edge autonomous terminal, synchronizing the trained load balance analysis model to the first edge autonomous terminal, and configuring the first edge autonomous terminal to the first parallel sub-network structure;
And by analogy, carrying out edge terminal configuration on the plurality of parallel sub-network structure sets to obtain the plurality of edge autonomous terminal sets.
7. The method of claim 6, wherein pre-building a first edge autonomous terminal and synchronizing the trained load balancing analysis model to the first edge autonomous terminal and configuring the first edge autonomous terminal to the first parallel sub-network structure comprises:
constructing three load balancing analysis sub-channels based on the BP neural network, the support vector machine and the generation type countermeasure network;
training the three load balancing analysis sub-channels one by adopting the first parallel balancing scheme set until the output precision of the three load balancing analysis sub-channels meets the preset requirement;
pre-constructing an output equalization sub-channel;
And arranging the three load balance analysis sub-channels in parallel, and connecting the output ends of the three load balance analysis sub-channels with the input ends of the output balance sub-channels to complete the construction of the load balance analysis model.
8. A digital twin space based parallel optimization system for a large-scale distribution network, characterized by the steps for implementing the method according to any of claims 1 to 7, said system comprising:
The power distribution network topology construction module is used for interactively obtaining power distribution network data assets of a target power distribution network and constructing and generating a power distribution network topology according to the power distribution network data assets, wherein the power distribution network topology comprises a plurality of nodes and a plurality of power transmission lines;
The structure splitting module is used for performing specialized splitting on the power distribution network topological structure according to the node type and the transmission line connection relation among the nodes to obtain a plurality of parallel sub-network structure sets;
The edge terminal configuration module is used for carrying out edge terminal configuration on the plurality of parallel sub-network structure sets to obtain a plurality of edge autonomous terminal sets;
the twin space construction module is used for pre-constructing a parallel optimized cloud, connecting the parallel optimized cloud with the plurality of nodes in the power distribution network topological structure in a communication manner, and connecting the parallel optimized cloud with the plurality of edge autonomous terminal sets in a communication manner to finish the construction of the parallel optimized twin space;
The twin load data acquisition module is used for receiving real-time load data of the target power distribution network in the parallel optimization twin space and transmitting the real-time load data to a plurality of nodes of the power distribution network topological structure to acquire a plurality of twin load data;
The load autonomous policy set generation module is used for generating a plurality of load autonomous policy sets by the plurality of edge autonomous terminal sets according to the plurality of twin load data and feeding the plurality of load autonomous policy sets back to the parallel optimization cloud;
And the parallel optimization module is used for the parallel optimization cloud end to perform parallel optimization of the target power distribution network based on the plurality of load autonomous policy sets.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410424144.3A CN118017509B (en) | 2024-04-10 | Large-scale power distribution network parallel optimization method based on digital twin space |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410424144.3A CN118017509B (en) | 2024-04-10 | Large-scale power distribution network parallel optimization method based on digital twin space |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118017509A CN118017509A (en) | 2024-05-10 |
| CN118017509B true CN118017509B (en) | 2024-07-16 |
Family
ID=
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114744764A (en) * | 2022-04-25 | 2022-07-12 | 南方电网科学研究院有限责任公司 | Digital twin terminal, system and terminal control method for power distribution network |
| CN117436215A (en) * | 2023-11-20 | 2024-01-23 | 国网智能电网研究院有限公司 | Distribution network digital twin body rapid construction method integrating electrical topology |
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114744764A (en) * | 2022-04-25 | 2022-07-12 | 南方电网科学研究院有限责任公司 | Digital twin terminal, system and terminal control method for power distribution network |
| CN117436215A (en) * | 2023-11-20 | 2024-01-23 | 国网智能电网研究院有限公司 | Distribution network digital twin body rapid construction method integrating electrical topology |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109613329B (en) | Fine line loss analysis system | |
| CN107330056B (en) | Wind power plant SCADA system based on big data cloud computing platform and operation method thereof | |
| CN110059356A (en) | A kind of bulk power grid intelligent control system and method based on big data and artificial intelligence | |
| CN113705085B (en) | Intelligent power grid multi-level structure modeling and risk assessment method | |
| CN111884347B (en) | Power data centralized control system for multi-source power information fusion | |
| CN110826228B (en) | Regional power grid operation quality limit evaluation method | |
| CN115859700B (en) | Power grid modeling method based on digital twin technology | |
| CN112688431A (en) | Power distribution network load overload visualization method and system based on big data | |
| Yang et al. | A novel PMU fog based early anomaly detection for an efficient wide area PMU network | |
| CN103869192A (en) | Smart power grid line loss detection method and system | |
| CN116526667B (en) | Secondary fusion distribution network feeder terminal system based on current Internet of things feedback mechanism | |
| CN112202597A (en) | Method for evaluating importance of communication network node in low-voltage distribution area | |
| CN103617447A (en) | Evaluation system and method for intelligent substation | |
| Ju et al. | The use of edge computing-based internet of things big data in the design of power intelligent management and control platform | |
| CN117374978B (en) | Grid-connected scheduling management method and system constructed by combining knowledge graph | |
| CN109378834A (en) | Large scale electric network voltage stability margin assessment system based on information maximal correlation | |
| CN118017509B (en) | Large-scale power distribution network parallel optimization method based on digital twin space | |
| CN110675276B (en) | Method and system for inversion droop control of direct current power transmission system | |
| CN111965442A (en) | Energy internet fault diagnosis method and device under digital twin environment | |
| CN118017509A (en) | Large-scale power distribution network parallel optimization method based on digital twin space | |
| CN107911763B (en) | Intelligent power distribution and utilization communication network EPON network planning method based on QoS | |
| Lydia et al. | Generation of dynamic energy management using data mining techniques basing on big data analytics isssues in smart grids | |
| Hu et al. | Optimization analysis of intelligent substation monitoring information based on improved PSO | |
| Gan et al. | Node importance ranking algorithm based on grey relational degree | |
| Yu et al. | Troubleshooting and Traceability Method Based on MapReduce Big Data Platform and Improved Genetic Reduction Algorithm for Smart Substation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |