Substation information determination method and system

Abstract

The invention provides a transformer substation information determination method and system, and relates to the technical field of transformer substations, and the method comprises the steps: constructing a community graph structure; processing the cell graph structure based on a graph neural network model to determine a target cell where the transformer substation is located; using a monitoring processing model to determine a plurality of primary selection positions with sparse people streams; inputting the power load information, the position information, the building layout information, the road layout information and the plurality of primary selection positions with sparse people streams of each cell to be supplied with power into a power transmission information determination model to determine the power transmission loss degree of each primary selection position with sparse people streams and the total length of a power transmission line of each primary selection position with sparse people streams; and determining the target installation position of the transformer substation based on the power transmission loss degree of each primary selection position with the sparse flow of people and the total length of the power transmission line of each primary selection position with the sparse flow of people. The method can quickly and accurately determine the installation position of the transformer substation.

Description

Technical Field

[0001] The present invention relates to the technical field of substations, and particularly to a method and system for determining substation information. Background Art

[0002] In the field of power supply, the location selection of substations is a key decision-making issue. A reasonable substation location is not only related to the effective supply of electricity, but also directly affects multiple aspects such as power loss, equipment investment, and operating costs. With the acceleration of urbanization and the continuous growth of power demand, how to efficiently and accurately determine the geographical location of substations has become an urgent problem to be solved in the power field. Traditional substation location selection methods mainly rely on manual evaluation and empirical judgment, which have problems such as low efficiency, strong subjectivity, and difficulty in comprehensively considering various factors.

[0003] Therefore, how to quickly and accurately determine the installation location of substations is an urgent problem to be solved currently. Summary of the Invention

[0004] The main technical problem to be solved by the present invention is how to quickly and accurately determine the installation location of substations.

[0005] According to a first aspect, the present invention provides a method for determining substation information, including: obtaining information of a plurality of communities to be powered, where the information of the plurality of communities to be powered includes power load information of each community to be powered, location information of each community to be powered, building layout information of each community to be powered, and road layout information of each community to be powered; constructing a community graph structure, where the community graph structure includes a plurality of nodes and multiple edges between the plurality of nodes, and each node in the plurality of nodes represents a community to be powered, and the node features of each node include power load information of each community to be powered, location information of each community to be powered, building layout information of each community to be powered, and road layout information of each community to be powered; processing the community graph structure based on a graph neural network model to determine the target community where the substation is located; obtaining the surveillance video of the target community where the substation is located, the building layout information of the target community, and the road layout information of the target community; using a surveillance processing model based on the surveillance video of the target community where the substation is located, the building layout information of the target community, and the road layout information of the target community to determine multiple primary selection locations with sparse pedestrian flow; inputting the power load information of each community to be powered, the location information of each community to be powered, the building layout information of each community to be powered, the road layout information of each community to be powered, and the multiple primary selection locations with sparse pedestrian flow into a power transmission information determination model to determine the power transmission loss degree of each primary selection location with sparse pedestrian flow and the total length of the transmission line of each primary selection location with sparse pedestrian flow; determining the target installation location of the substation based on the power transmission loss degree of each primary selection location with sparse pedestrian flow and the total length of the transmission line of each primary selection location with sparse pedestrian flow.

[0006] Further, the input of the graph neural network model is the community graph structure, the output of the graph neural network model is the target community where the substation is located, and the multiple edge features of each edge in the multiple edges include the distance and direction between the communities.

[0007] Further, the surveillance processing model is a gated recurrent unit, the input of the surveillance processing model is the surveillance video of the target community where the substation is located, the building layout information of the target community, and the road layout information of the target community, and the output of the surveillance processing model is multiple primary selection locations with sparse pedestrian flow.

[0008] Further, determining the target installation location of the substation based on the power transmission loss degree of each primary selection location with sparse pedestrian flow and the total length of the transmission lines at each primary selection location with sparse pedestrian flow includes: after assigning different weights to the power transmission loss degree of each primary selection location with sparse pedestrian flow and the total length of the transmission lines at each primary selection location with sparse pedestrian flow respectively, performing weighted summation to obtain the power cost coefficient of each primary selection location with sparse pedestrian flow, and taking the installation location of the substation with the least power cost coefficient as the target installation location of the substation.

[0009] Further, the power transmission information determination model is a deep neural network model. The inputs of the power transmission information determination model are the power load information of each power supply target community, the location information of each power supply target community, the building layout information of each power supply target community, the road layout information of each power supply target community, and the multiple primary selection locations with sparse pedestrian flow. The outputs of the power transmission information determination model are the power transmission loss degree of each primary selection location with sparse pedestrian flow and the total length of the transmission lines at each primary selection location with sparse pedestrian flow.

[0010] According to a second aspect, the present invention provides a system for determining substation information, including: a first acquisition module, configured to acquire information of a plurality of power supply target communities, where the information of the plurality of power supply target communities includes the power load information of each power supply target community, the location information of each power supply target community, the building layout information of each power supply target community, and the road layout information of each power supply target community; a construction module, configured to construct a community graph structure, where the community graph structure includes a plurality of nodes and multiple edges between the plurality of nodes, and each node in the plurality of nodes represents a power supply target community, and the node features of each node include the power load information of each power supply target community, the location information of each power supply target community, the building layout information of each power supply target community, and the road layout information of each power supply target community; a target community determination module, configured to process the community graph structure based on a graph neural network model to determine the target community where the substation is located; a second acquisition module, configured to acquire the surveillance video of the target community where the substation is located, the building layout information of the target community, and the road layout information of the target community; a surveillance processing module, configured to use a surveillance processing model to determine multiple primary selection locations with sparse pedestrian flow based on the surveillance video of the target community where the substation is located, the building layout information of the target community, and the road layout information of the target community; A first acquisition module, configured to input the power load information of each power supply target cell, the location information of each power supply target cell, the building layout information of each power supply target cell, the road layout information of each power supply target cell, and the multiple preliminary candidate locations with sparse pedestrian flow into a power transmission information determination model to determine the power transmission loss degree of each preliminary candidate location with sparse pedestrian flow and the total length of the power transmission line of each preliminary candidate location with sparse pedestrian flow. An installation location determination module, configured to determine the target installation location of the substation based on the power transmission loss degree of each preliminary candidate location with sparse pedestrian flow and the total length of the power transmission line of each preliminary candidate location with sparse pedestrian flow.

[0011] Furthermore, the input of the graph neural network model is the cell graph structure, the output of the graph neural network model is the target cell where the substation is located, and the multiple edge features of each edge in the multiple edges include the distance and direction between cells.

[0012] Furthermore, the monitoring and processing model is a gated recurrent unit. The input of the monitoring and processing model is the monitoring video of the target cell where the substation is located, the building layout information of the target cell, and the road layout information of the target cell. The output of the monitoring and processing model is multiple preliminary candidate locations with sparse pedestrian flow.

[0013] Furthermore, the installation location determination module is further configured to: after assigning different weights to the power transmission loss degree of each preliminary candidate location with sparse pedestrian flow and the total length of the power transmission line of each preliminary candidate location with sparse pedestrian flow, perform weighted summation to obtain the power cost coefficient of each preliminary candidate location with sparse pedestrian flow, and use the installation location of the substation with the least power cost coefficient as the target installation location of the substation.

[0014] Furthermore, the power transmission information determination model is a deep neural network model. The input of the power transmission information determination model is the power load information of each power supply target cell, the location information of each power supply target cell, the building layout information of each power supply target cell, the road layout information of each power supply target cell, and the multiple preliminary candidate locations with sparse pedestrian flow. The output of the power transmission information determination model is the power transmission loss degree of each preliminary candidate location with sparse pedestrian flow and the total length of the power transmission line of each preliminary candidate location with sparse pedestrian flow.

[0015] A method and system for determining substation information provided by the present invention. The method includes obtaining information of multiple communities to be powered, where the information of the multiple communities to be powered includes power load information of each community to be powered, location information of each community to be powered, building layout information of each community to be powered, and road layout information of each community to be powered; constructing a community graph structure, where the community graph structure includes multiple nodes and multiple edges between the nodes, and each node in the multiple nodes represents a community to be powered, and the node features of each node include power load information of each community to be powered, location information of each community to be powered, building layout information of each community to be powered, and road layout information of each community to be powered; processing the community graph structure based on a graph neural network model to determine the target community where the substation is located; obtaining the surveillance video of the target community where the substation is located, building layout information of the target community, and road layout information of the target community; using a surveillance processing model based on the surveillance video of the target community where the substation is located, building layout information of the target community, and road layout information of the target community to determine multiple initial positions with sparse pedestrian flow; inputting the power load information of each community to be powered, location information of each community to be powered, building layout information of each community to be powered, road layout information of each community to be powered, and the multiple initial positions with sparse pedestrian flow into a power transmission information determination model to determine the power transmission loss degree of each initial position with sparse pedestrian flow and the total length of the transmission line of each initial position with sparse pedestrian flow; determining the target installation position of the substation based on the power transmission loss degree of each initial position with sparse pedestrian flow and the total length of the transmission line of each initial position with sparse pedestrian flow. This method can quickly and accurately determine the installation position of the substation. Description of the Drawings

[0016] Figure 1 It is a schematic flowchart of a method for determining substation information provided by an embodiment of the present invention; Figure 2 It is a schematic diagram of a system for determining substation information provided by an embodiment of the present invention. Detailed Embodiments

[0017] In an embodiment of the present invention, there is provided a method for determining substation information as shown in Figure 1 The method for determining substation information includes steps S1 to S7: Step S1, obtain information of multiple communities to be powered, where the information of the multiple communities to be powered includes power load information of each community to be powered, location information of each community to be powered, building layout information of each community to be powered, and road layout information of each community to be powered.

[0018] In some embodiments, the power load information of each cell to be powered includes power consumption, load curve, power load distribution in different time periods, power load distribution in different locations, load types of different devices in the cell, etc.

[0019] In some embodiments, the location information of each cell to be powered can be represented as the geographical location coordinate information of each cell to be powered in the city, including longitude, latitude, etc.

[0020] The building layout information includes information such as the distribution, type, and floor area ratio of buildings in each cell to be powered.

[0021] The road layout information includes information such as the layout, width, and connection of roads in each cell to be powered.

[0022] In some embodiments, the power load information, location information, building layout information, and road layout information of each cell to be powered can be obtained through power companies, geographic information systems, planning drawings provided by urban planning departments, road planning maps, etc.

[0023] Step S2, construct a cell graph structure. The cell graph structure includes multiple nodes and multiple edges between the nodes. Each node in the multiple nodes represents a cell to be powered, and the node features of each node include the power load information, location information, building layout information, and road layout information of each cell to be powered.

[0024] The cell graph structure refers to a graph structure that takes each cell to be powered as a node and represents the relationship between cells through the edges between the nodes.

[0025] The graph structure includes multiple nodes and multiple edges. The graph structure is a data structure composed of nodes and edges.

[0026] Step S3, process the cell graph structure based on the graph neural network model to determine the target cell where the substation is located.

[0027] The input of the graph neural network model is the cell graph structure, the output of the graph neural network model is the target cell where the substation is located, and the multiple edge features of each edge in the multiple edges include the distance and direction between cells. The graph neural network model includes a Graph Neural Network (GNN) and a fully connected layer. The graph neural network is a neural network that directly acts on graph-structured data.

[0028] In the power supply network planning, the spatial relationship between communities is a very important factor. By constructing a graph structure, spatial information such as the distance and direction between communities can be encoded into edge features, so as to better consider the spatial relationship between communities.

[0029] The target community where the substation is located is a community selected by the graph neural network model from multiple communities to be powered, considering the power supply demand and resource allocation of the entire region. The target community where the substation is located is one of the target communities where the substation is located. Power supply network planning usually needs to consider the power supply demand and resource allocation of the entire region. Using the graph neural network model can process the community graph structure globally, comprehensively consider the interactions between each community node, and comprehensively consider the power load information, location information, building layout information, and road layout information of each community to be powered, so as to achieve global optimization and finally determine the target community where the substation is located.

[0030] In the specific application scenario of determining the target community where the substation is located, constructing the community graph structure and using the graph neural network model can fully consider the spatial relationship, achieve global optimization, and adaptively learn the feature representation, so as to accurately determine the target community where the substation is located.

[0031] Step S4, obtain the surveillance video of the target community where the substation is located, the building layout information of the target community, and the road layout information of the target community.

[0032] The surveillance video of the target community where the substation is located is the surveillance video obtained by monitoring the target community. In some embodiments, the surveillance video of the target community where the substation is located can be obtained by synthesizing the surveillance videos of multiple cameras in the target community. In some embodiments, the surveillance video of the target community where the substation is located can be obtained by multi-screen splicing of the surveillance videos of multiple cameras in the target community. In some embodiments, the surveillance videos of multiple cameras in the target community can be superimposed by a video fusion algorithm to obtain the surveillance video of the target community where the substation is located.

[0033] Step S5, use the surveillance processing model to determine multiple initial positions with sparse human flow based on the surveillance video of the target community where the substation is located, the building layout information of the target community, and the road layout information of the target community.

[0034] The surveillance processing model is a gated recurrent unit. The input of the surveillance processing model is the surveillance video of the target community where the substation is located, the building layout information of the target community, and the road layout information of the target community. The output of the surveillance processing model is multiple initial positions with sparse human flow. In some embodiments, the surveillance processing model is trained by the gradient descent method.

[0035] Multiple initial candidate locations with sparse pedestrian flow are initial candidate locations where the pedestrian flow output by the gated recurrent unit is relatively sparse and the locations are relatively reasonable within the target community.

[0036] The gated recurrent unit (GRU) is used to process sequential data and temporal information. The gated recurrent unit includes three components: a memory unit, an update gate, and a reset gate. Through the gated recurrent unit model, the surveillance video of the target community where the substation is located in consecutive time periods can be processed, and the relationships in the time series of the surveillance video of the target community where the substation is located can be better captured. The features that comprehensively consider the correlation relationships between the surveillance videos of the target community where the substation is located at each time point can be output, making the output features more accurate and comprehensive.

[0037] The gated recurrent unit analyzes and extracts the trajectories and trends of pedestrian flow from the surveillance video, building layout information, and road layout information, divides the pedestrian flow within the target community where the substation is located into different time periods or spatial regions, then analyzes the pedestrian flow density and distribution in each time period or spatial region, and finds the initial candidate locations where the pedestrian flow is relatively sparse and the installation locations are relatively reasonable within the community as multiple initial candidate locations with sparse pedestrian flow.

[0038] In some embodiments, the monitoring and processing model includes a crowd distribution determination layer, a preliminary screening layer for regions, and a preliminary position determination layer. The input of the crowd distribution determination layer is the monitoring video of the target community where the substation is located, and the output of the crowd distribution determination layer is the crowd distribution map of the target community where the substation is located. The input of the preliminary screening layer for regions is the crowd distribution map of the target community where the substation is located, and the output of the preliminary screening layer for regions is multiple preliminary screening regions with sparse crowds. The input of the preliminary position determination layer is multiple preliminary screening regions with sparse crowds, the building layout information of the target community, and the road layout information of the target community. The output of the preliminary position determination layer is multiple preliminary positions with sparse crowds. The multiple preliminary screening regions with sparse crowds are regions with relatively few people output only by the preliminary screening layer for regions. The multiple preliminary positions with sparse crowds are preliminary positions that are relatively sparse in terms of crowd flow and relatively reasonable in location within the community, which are comprehensively output by the preliminary position determination layer based on the multiple preliminary screening regions with sparse crowds, the building layout information of the target community, and the road layout information of the target community. The crowd distribution determination layer is responsible for extracting crowd information from the monitoring video, the preliminary screening layer for regions is responsible for preliminarily screening the crowd distribution, and the preliminary position determination layer further comprehensively considers multiple factors to determine the final position. Each layer has clear responsibilities and objectives, avoiding task confusion and information redundancy. The output of each layer can be used as the input of the next layer, forming an end-to-end data processing flow, which is conducive to information transmission and processing. Through the hierarchical design, data can be effectively organized and utilized, improving the overall performance and expressiveness of the system.

[0039] Step S6: Input the power load information of each power supply target community, the location information of each power supply target community, the building layout information of each power supply target community, the road layout information of each power supply target community, and the multiple preliminary positions with sparse crowds into the power transmission information determination model to determine the power transmission loss degree of each preliminary position with sparse crowds and the total length of the power transmission lines of each preliminary position with sparse crowds.

[0040] The power transmission information determination model is a deep neural network model. The input of the power transmission information determination model is the power load information of each power supply target community, the location information of each power supply target community, the building layout information of each power supply target community, the road layout information of each power supply target community, and the multiple preliminary positions with sparse crowds. The output of the power transmission information determination model is the power transmission loss degree of each preliminary position with sparse crowds and the total length of the power transmission lines of each preliminary position with sparse crowds.

[0041] During the power transmission process, due to factors such as the impedance, resistance, and inductance of the wires, power loss will occur, which affects the quality and stability of power supply. The power transmission loss degree of each initial selection position with sparse pedestrian flow refers to the loss degree when each initial selection position with sparse pedestrian flow transmits power to all communities to be powered.

[0042] The deep neural network model includes a deep neural network (DNN). The deep neural network can include multiple processing layers, each processing layer consists of multiple neurons, and each neuron performs a matrix transformation on the data. The parameters used in the matrix can be obtained through training. By training the deep neural network, the model can learn the complex relationships between various features from multiple inputs such as the power load information, location information, building layout information, road layout information, and initial selection positions of the communities to be powered, so as to accurately predict the power transmission loss degree and the length of the transmission line. The deep neural network can achieve end-to-end learning and can better utilize the input information, including information such as power load, location, building layout, road layout, and initial selection positions, enabling the model to comprehensively consider the impact of these information on power loss and the length of the transmission line.

[0043] Step S7, determine the target installation position of the substation based on the power transmission loss degree of each initial selection position with sparse pedestrian flow and the total length of the transmission line of each initial selection position with sparse pedestrian flow.

[0044] In some embodiments, the determining the target installation position of the substation based on the power transmission loss degree of each initial selection position with sparse pedestrian flow and the total length of the transmission line of each initial selection position with sparse pedestrian flow includes: after assigning different weights to the power transmission loss degree of each initial selection position with sparse pedestrian flow and the total length of the transmission line of each initial selection position with sparse pedestrian flow respectively, then performing weighted summation to obtain the power cost coefficient of each initial selection position with sparse pedestrian flow, and taking the installation position of the substation with the least power cost coefficient as the target installation position of the substation.

[0045] By assigning different weights to the power transmission loss degree and the total length of the transmission line, and comprehensively considering these two factors, a power cost coefficient is obtained. By comparing the power cost coefficients of different substation positions, the substation position with the smallest power cost coefficient can be selected as the target installation position. The advantage of doing this is that it can minimize the power transmission loss and the length of the transmission line to the greatest extent, thereby reducing energy waste and improving the power supply efficiency and reliability.

[0046] The power cost coefficient is used to evaluate the power cost of different substation positions.

[0047] Based on the same inventive concept, Figure 2Schematic diagram of a system for determining substation information provided by an embodiment of the present invention. The system for determining substation information includes: A first acquisition module 21, configured to acquire information of multiple communities to be powered. The information of the multiple communities to be powered includes power load information of each community to be powered, location information of each community to be powered, building layout information of each community to be powered, and road layout information of each community to be powered; A construction module 22, configured to construct a community graph structure. The community graph structure includes multiple nodes and multiple edges between the multiple nodes. Each node in the multiple nodes represents a community to be powered, and the node features of each node include power load information of each community to be powered, location information of each community to be powered, building layout information of each community to be powered, and road layout information of each community to be powered; A target community determination module 23, configured to process the community graph structure based on a graph neural network model to determine the target community where the substation is located; A second acquisition module 24, configured to acquire the surveillance video of the target community where the substation is located, the building layout information of the target community, and the road layout information of the target community; A surveillance processing module 25, configured to use a surveillance processing model to determine multiple initial positions with sparse pedestrian flows based on the surveillance video of the target community where the substation is located, the building layout information of the target community, and the road layout information of the target community; A power transmission information determination module 26, configured to input the power load information of each community to be powered, the location information of each community to be powered, the building layout information of each community to be powered, the road layout information of each community to be powered, and the multiple initial positions with sparse pedestrian flows into a power transmission information determination model to determine the power transmission loss degree of each initial position with sparse pedestrian flow and the total length of the power transmission line of each initial position with sparse pedestrian flow; An installation position determination module 27, configured to determine the target installation position of the substation based on the power transmission loss degree of each initial position with sparse pedestrian flow and the total length of the power transmission line of each initial position with sparse pedestrian flow.

Claims

1. A method for determining substation information, characterized in that: include: Acquire information of multiple cells to be powered, wherein the information of the multiple cells to be powered includes power load information of each cell to be powered, location information of each cell to be powered, building layout information of each cell to be powered, and road layout information of each cell to be powered; Constructing a cell graph structure, the cell graph structure comprising a plurality of nodes and a plurality of edges between the plurality of nodes, each of the plurality of nodes representing a cell to be powered, and the node characteristics of each node comprising power load information of each cell to be powered, location information of each cell to be powered, building layout information of each cell to be powered, and road layout information of each cell to be powered; Processing the cell graph structure based on a graph neural network model to determine a target cell where the substation is located; Obtaining surveillance video of the target community where the substation is located, building layout information of the target community, and road layout information of the target community; Determine a plurality of preliminary locations with sparse traffic using a monitoring processing model based on monitoring video of a target cell where the substation is located, building layout information of the target cell, and road layout information of the target cell; Input the power load information of each cell to be powered, the location information of each cell to be powered, the building layout information of each cell to be powered, the road layout information of each cell to be powered, and the multiple sparsely populated preliminary locations into the power transmission information determination model to determine the power transmission loss degree of each sparsely populated preliminary location and the total length of the power transmission line of each sparsely populated preliminary location; The target installation location of the substation is determined based on the degree of power transmission loss at each of the preliminary selected locations with sparse traffic and the total length of the transmission line at each of the preliminary selected locations with sparse traffic.

2. The method for determining substation information according to claim 1, characterized in that: The input of the graph neural network model is the cell graph structure, and the output of the graph neural network model is the target cell where the substation is located. The multiple edge features of each of the multiple edges include the distance and direction between the cells.

3. The method for determining substation information according to claim 1, characterized in that: The monitoring processing model is a gated loop unit, the input of which is the monitoring video of the target cell where the substation is located, the building layout information of the target cell, and the road layout information of the target cell, and the output of the monitoring processing model is a plurality of preliminary locations with sparse traffic.

4. The method for determining substation information according to claim 1, characterized in that: The method of determining the target installation location of the substation based on the degree of power transmission loss at each of the preliminary locations with sparse traffic and the total length of the transmission lines at each of the preliminary locations with sparse traffic includes: assigning different weights to the degree of power transmission loss at each of the preliminary locations with sparse traffic and the total length of the transmission lines at each of the preliminary locations with sparse traffic, respectively, and then performing weighted summation to obtain the power cost coefficient of each of the preliminary locations with sparse traffic, and taking the installation location of the substation with the smallest power cost coefficient as the target installation location of the substation.

5. The method for determining substation information according to claim 1, characterized in that: The power transmission information determination model is a deep neural network model, and the input of the power transmission information determination model is the power load information of each cell to be powered, the location information of each cell to be powered, the building layout information of each cell to be powered, the road layout information of each cell to be powered, and the multiple preliminary locations with sparse traffic, and the output of the power transmission information determination model is the degree of power transmission loss at each preliminary location with sparse traffic and the total length of the transmission line at each preliminary location with sparse traffic.

6. A system for determining substation information, characterized in that: include: A first acquisition module is used to acquire information of multiple cells to be powered, wherein the information of the multiple cells to be powered includes power load information of each cell to be powered, location information of each cell to be powered, building layout information of each cell to be powered, and road layout information of each cell to be powered; A construction module is used to construct a cell graph structure, wherein the cell graph structure includes a plurality of nodes and a plurality of edges between the plurality of nodes, wherein each of the plurality of nodes represents a cell to be powered, and the node characteristics of each node include power load information of each cell to be powered, location information of each cell to be powered, building layout information of each cell to be powered, and road layout information of each cell to be powered; A target cell determination module, used to process the cell graph structure based on a graph neural network model to determine the target cell where the substation is located; The second acquisition module is used to acquire the monitoring video of the target cell where the substation is located, the building layout information of the target cell, and the road layout information of the target cell; A monitoring processing module, configured to determine a plurality of preliminary locations with sparse traffic using a monitoring processing model based on a monitoring video of a target cell where the substation is located, building layout information of the target cell, and road layout information of the target cell; A power transmission information determination module, used to input the power load information of each cell to be powered, the location information of each cell to be powered, the building layout information of each cell to be powered, the road layout information of each cell to be powered, and the multiple preliminary locations with sparse traffic into the power transmission information determination model to determine the power transmission loss degree of each preliminary location with sparse traffic and the total length of the transmission line of each preliminary location with sparse traffic; The installation location determination module is used to determine the target installation location of the substation based on the degree of power transmission loss at each of the preliminary selected locations with sparse traffic and the total length of the transmission line at each of the preliminary selected locations with sparse traffic.

7. The substation information determination system according to claim 6, characterized in that: The input of the graph neural network model is the cell graph structure, and the output of the graph neural network model is the target cell where the substation is located. The multiple edge features of each of the multiple edges include the distance and direction between the cells.

8. The system for determining substation information according to claim 6, characterized in that: The monitoring processing model is a gated loop unit, the input of which is the monitoring video of the target cell where the substation is located, the building layout information of the target cell, and the road layout information of the target cell, and the output of the monitoring processing model is a plurality of preliminary locations with sparse traffic.

9. The system for determining substation information according to claim 6, characterized in that: The installation location determination module is also used to assign different weights to the degree of power transmission loss at each of the preliminary locations with sparse traffic and the total length of the transmission lines at each of the preliminary locations with sparse traffic, and then perform weighted summation to obtain the power cost coefficient of each of the preliminary locations with sparse traffic, and take the installation location of the substation with the lowest power cost coefficient as the target installation location of the substation.

10. The system for determining substation information according to claim 6, characterized in that: The power transmission information determination model is a deep neural network model, and the input of the power transmission information determination model is the power load information of each cell to be powered, the location information of each cell to be powered, the building layout information of each cell to be powered, the road layout information of each cell to be powered, and the multiple preliminary locations with sparse traffic, and the output of the power transmission information determination model is the degree of power transmission loss at each preliminary location with sparse traffic and the total length of the transmission line at each preliminary location with sparse traffic.

Images