System and method for analyzing actual user number of power grid of power supply company

Abstract

The invention discloses a power supply company power grid actual user number analysis system and method, and the system comprises an actual user load demand dynamic analysis module which is used for generating an actual user load demand dynamic adaptation instruction containing a power supply region; the remote cloud data end information supervision module is used for presetting parameters of nodes and edges of the graph neural network and actual user load demand adaptation parameters to an actual user load demand remote cloud data end of a power supply company; the power consumption cost dynamic analysis module is used for determining a power consumption cost dynamic analysis process and a power consumption cost dynamic analysis result by using an adaptive algorithm preset by a random forest; the graph neural network dynamic analysis process and result module is used for determining the graph neural network dynamic analysis process and result matched with the actual user load demand of the power supply company; the power supply area dynamic analysis process and result module is used for determining a power supply area dynamic analysis process and result; therefore, an electric energy supply size adjusting instruction is generated. The working efficiency and safety of a power supply company are improved.

Description

Technical Field

[0001] The present invention relates to the field of dynamic data analysis of power supply companies, and in particular to a system and method for analyzing the actual number of users of a power grid of a power supply company. Background Art

[0002] In modern power systems, accurate analysis of the actual number of users in the power grid is crucial to optimizing the operation of power supply companies, improving power grid stability, and ensuring power supply to users. To meet this challenge, power supply companies need to rely on advanced power grid user number analysis systems that can track and analyze the power consumption of each user in the power grid in real time in order to make timely scheduling and management decisions. Such systems are usually based on complex data acquisition and analysis technologies, including smart meters, data integration platforms, and advanced analysis tools, designed to provide accurate user load data, power consumption patterns, and demand forecasts, thereby optimizing the operating efficiency and service quality of the power grid.

[0003] Existing grid user data analysis technologies mainly rely on smart meters and data acquisition systems, which can record users' electricity consumption data in real time and upload them to a central data platform. Through comprehensive analysis of these data, power supply companies can obtain users' electricity consumption habits, load changes and forecast information. Although these technologies have made significant progress in data collection and preliminary analysis, they still have several significant shortcomings. Existing analysis systems usually have limited capabilities for large-scale data processing, and may experience delays or performance bottlenecks when processing high-frequency and large-scale data. In addition, existing technologies often lack sufficient flexibility and accuracy when dealing with complex grid topologies and dynamic load changes, making it difficult to cope with various emergencies that occur in practical applications. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides a system and method for analyzing the actual number of users of a power grid of a power supply company to solve the above-mentioned technical problems.

[0005] In the first aspect, the present application provides a power supply company's power grid actual user number analysis system, the system comprising: an actual user load demand dynamic analysis module, used to obtain dynamic analysis instructions; identify the power supply area matching the dynamic analysis instructions and the actual user load demand dynamic analysis graph neural network parameters, and generate a dynamic adaptation instruction for the actual user load demand containing the power supply area; a remote cloud data terminal information supervision module, used to preset the nodes and edge parameters of the graph neural network and the actual user load demand adaptation parameters to the power supply company's actual user load demand remote cloud data terminal; based on the actual user load demand dynamic adaptation instruction, obtain the power supply company's actual user load demand pre-processed by the remote cloud data terminal; send the power supply company's actual user load demand to the matching dynamic analysis module according to the power supply area; wherein the data storage and analysis expression of the actual user load demand of the power supply company is power supply area + actual user load demand + average population per household; a power consumption cost dynamic analysis module, used to receive the power supply area as the power consumption cost power supply area The actual user load demand of the power supply company is determined by using the preset adaptation algorithm of random forest to determine the dynamic analysis process and results of power consumption cost that matches the actual user load demand of the power supply company; the graph neural network dynamic analysis process and result module is used to receive the actual user load demand of the power supply company whose power supply area is the public power supply area, and determine the graph neural network dynamic analysis process and results that match the actual user load demand of the power supply company by preset dynamic analysis graph neural network parameters; the power supply area dynamic analysis process and result module is used to determine the power supply area dynamic analysis process and results through the power consumption cost dynamic analysis process and results, the graph neural network dynamic analysis process and results, and the actual user load demand dynamic analysis graph neural network parameters; determine the actual user load demand that matches the dynamic analysis process and results of power consumption cost, the graph neural network dynamic analysis process and results, and the power supply area dynamic analysis process and results; generate power supply size adjustment instructions by analyzing the actual user load demand and transformation efficiency.

[0006] Furthermore, the actual user load demand dynamic analysis module includes a dynamic analysis instruction acquisition unit, a dynamic analysis model establishment unit, and an actual user load demand confirmation unit; the dynamic analysis instruction acquisition unit is used to acquire dynamic analysis instructions; wherein the dynamic analysis instructions include the power supply company's actual user load demand dynamic analysis process, the power supply company's benchmark user load demand, and the variable user load demand; based on the power supply company's actual user load demand dynamic analysis process, determine whether there are matching actual user load demand dynamic analysis model parameters; when they can be matched, determine that the dynamic analysis graph neural network parameters in the actual user load demand dynamic analysis model parameters are the actual user load demand dynamic analysis graph neural network parameters; when they cannot be matched, extract the data in the power supply company's actual user load demand dynamic analysis process and the power supply company's benchmark user load demand. power supply area to obtain preset dynamic analysis graph neural network parameters for different time periods that are satisfied by both data and power supply area; input the power supply company's benchmark user load demand into the preset dynamic analysis graph neural network parameters to obtain the actual user load demand of dynamic analysis results in different time periods; determine the preset dynamic analysis graph neural network parameters that are consistent with the dynamic analysis result actual user load demand and the changing user load demand as the actual user load demand dynamic analysis graph neural network parameters; a dynamic analysis model establishment unit, used to establish the power supply company's actual user load demand dynamic analysis process and the power supply company's actual user load demand dynamic analysis process into actual user load demand dynamic analysis model parameters; an actual user load demand confirmation unit, used to confirm the power supply area in the power supply company's benchmark user load demand, and generate actual user load demand dynamic adaptation instructions according to the power supply area.

[0007] Furthermore, the remote cloud data terminal information supervision module includes a parameter transmission unit; the parameter transmission unit is used to detect whether there are nodes, edge parameters and actual user load demand adaptation parameters of the graph neural network in the remote cloud data terminal matching the power supply area; when it is detected that there are no nodes, edge parameters and actual user load demand adaptation parameters of the graph neural network, the nodes, edge parameters and actual user load demand adaptation parameters of the graph neural network are transmitted to the remote cloud data terminal.

[0008] Furthermore, the power supply area dynamic analysis process and result module includes a dynamic analysis process and result determination unit and an actual user load demand interaction curve generation unit; the dynamic analysis process and result determination unit is used to determine the power supply areas that match the power consumption cost dynamic analysis process and result, and the graph neural network dynamic analysis process and result respectively; determine the actual user load demand analysis data set that matches the power consumption cost dynamic analysis process and result, and the graph neural network dynamic analysis process and result respectively according to the power supply area; wherein the actual user load demand analysis data set includes the allocation principle of training set 6: verification set 2 and test set 2 in the data set; the power consumption cost dynamic analysis process and result, and the graph neural network dynamic analysis process and result are brought into the training set 6: verification set 2 and test set 2 allocation principle in the data set, and the respective matching analysis training data are determined from the respective matching actual user load demand analysis data sets; the actual user load demand interaction curve generation unit is used to use the actual user load demand dynamic analysis graph neural network parameters that match the final dynamic analysis results to determine the historical results of different time periods that match the actual user load demand dynamic analysis graph neural network parameters; based on the historical results and the power supply area dynamic analysis process and results, generate the actual user load demand interaction curve.

[0009] The present application provides a method for analyzing the actual number of users of a power grid of a power supply company. The system operation includes: determining the power supply area that matches the acquired dynamic analysis instruction and the actual user load demand dynamic analysis graph neural network parameters, and generating a dynamic adaptation instruction containing the actual user load demand of the power supply area; in response to the existence of nodes and edge parameters of the graph neural network and the actual user load demand adaptation parameters on the remote cloud data end that matches the power supply area, obtaining the actual user load demand of the power supply company after preprocessing by the remote cloud data end; sending the actual user load demand of the power supply company to the matching cloud control end according to the power supply area; wherein the data storage and analysis expression of the actual user load demand of the power supply company is power supply area + actual user load demand + average population per household; and the cloud control end operation matches the power supply area; receiving the actual user load demand of the power supply company The power consumption cost dynamic analysis process and results and the graph neural network dynamic analysis process and results that match the user load demand; among them, the power supply area that matches the power consumption cost dynamic analysis process and results is the power consumption cost power supply area, and the power supply area that matches the graph neural network dynamic analysis process and results is the public power supply area; the power consumption cost dynamic analysis process and results, the graph neural network dynamic analysis process and results and the actual user load demand dynamic analysis graph neural network parameters are used to determine the power supply area dynamic analysis process and results; determine the power consumption cost dynamic analysis process and results, the graph neural network dynamic analysis process and results and the power supply area dynamic analysis process and results that match each other to analyze the actual user load demand; use the analysis of the actual user load demand and transformation efficiency to generate power supply size adjustment instructions.

[0010] Further, the power supply area and the actual user load demand dynamic analysis graph neural network parameters that match the obtained dynamic analysis instructions are determined, specifically including: obtaining dynamic analysis instructions; wherein the dynamic analysis instructions include the power supply company's actual user load demand dynamic analysis process, the power supply company's benchmark user load demand, and the variable user load demand; based on the power supply company's actual user load demand dynamic analysis process, determine whether there are matching actual user load demand dynamic analysis model parameters; when a match can be made, determine the dynamic analysis graph neural network parameters in the actual user load demand dynamic analysis model parameters as the actual user load demand dynamic analysis graph neural network parameters; when a match cannot be made, extract the data in the power supply company's actual user load demand dynamic analysis process and the power supply area in the power supply company's benchmark user load demand to obtain preset dynamic analysis graph neural network parameters for different time periods that both the data and the power supply area meet; input the power supply company's benchmark user load demand into the preset dynamic analysis graph neural network parameters to obtain the actual user load demand of the dynamic analysis results for different time periods; determine the preset dynamic analysis graph neural network parameters whose actual user load demand in the dynamic analysis results is consistent with the variable user load demand as the actual user load demand dynamic analysis graph neural network parameters.

[0011] Furthermore, in response to the existence of nodes, edge parameters and actual user load demand adaptation parameters of the graph neural network in the remote cloud data end that matches the power supply area, it specifically includes: detecting whether there are nodes, edge parameters and actual user load demand adaptation parameters of the graph neural network in the remote cloud data end that matches the power supply area; when it is detected that there are no nodes, edge parameters and actual user load demand adaptation parameters of the graph neural network, the nodes, edge parameters and actual user load demand adaptation parameters of the graph neural network are transmitted to the remote cloud data end.

[0012] Furthermore, the analysis training data that match the dynamic analysis process and results of electricity consumption cost, the dynamic analysis process and results of graph neural network, and the dynamic analysis process and results of power supply area are determined, specifically including: determining the power supply area that matches the dynamic analysis process and results of electricity consumption cost, and the dynamic analysis process and results of graph neural network; determining the actual user load demand analysis data set that matches the dynamic analysis process and results of electricity consumption cost, and the dynamic analysis process and results of graph neural network according to the power supply area; wherein the actual user load demand analysis data set includes the allocation principle of training set 6: validation set 2 and test set 2 in the data set; the dynamic analysis process and results of electricity consumption cost, and the dynamic analysis process and results of graph neural network are brought into the allocation principle of training set 6: validation set 2 and test set 2 in the data set, and the matching analysis training data are determined from the matching actual user load demand analysis data sets; the actual user load demand dynamic analysis graph neural network parameters that match the final dynamic analysis results are used to determine the historical results of different time periods that match the actual user load demand dynamic analysis graph neural network parameters; based on the historical results and the dynamic analysis process and results of the power supply area, the actual user load demand interaction curve is generated.

[0013] Beneficial effects:

[0014] The present invention proposes a power supply company power grid actual user number analysis system and method, through the actual user load demand dynamic analysis module, realizes the selective acquisition of matching power supply company actual user load demand, and selectively obtains appropriate actual user load demand dynamic analysis graph neural network parameters. Through the remote cloud data terminal information supervision module, the nodes and edge parameters of the graph neural network installed in the power system that uploads the actual user load demand of the power supply company and the actual user load demand adaptation parameters are matched, so as to realize the technical effect of pre-processing the actual user load demand in the remote cloud data terminal, avoid the time loss caused by the actual user load demand processing in the analysis system, and directly perform the actual user load demand adaptation processing on the actual user load demand through the remote cloud data terminal, so as to avoid the omission of the actual user load demand. Through the power consumption cost dynamic analysis module, the actual user load demand with power consumption cost is introduced into the adaptation algorithm to obtain the power consumption cost dynamic analysis process and results, and the process of encrypting the actual user load demand and decrypting the dynamic analysis is not involved in the dynamic analysis process, and the power consumption cost dynamic analysis process and results are directly analyzed by encrypting the actual user load demand, so as to avoid the omission of the actual user load demand of power consumption cost in the dynamic analysis process. Through the dynamic analysis process and result module of the power supply area, the automatic generation of power supply size adjustment instructions is realized, thereby improving the work efficiency of the power supply company and improving the safety of power supply. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 It is a schematic diagram of the module composition of the system of the present invention;

[0017] Figure 2 It is an operation flow chart of the system of the present invention. DETAILED DESCRIPTION

[0018] It should be understood by those skilled in the art that the embodiments described below are only preferred embodiments of the present disclosure, and do not mean that the present disclosure can only be implemented through the preferred embodiments. The preferred embodiments are only used to explain the technical principles of the present disclosure, and are not used to limit the protection scope of the present disclosure. Based on the preferred embodiments provided by the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work should still fall within the protection scope of the present disclosure.

[0019] The technical solution proposed in the embodiments of the present application is described in detail below with reference to the accompanying drawings.

[0020] Figure 1 A power supply company power grid actual user number analysis system is provided in the embodiment of the present application. Figure 1 As shown, the analysis system provided in the embodiment of the present application mainly includes: an actual user load demand dynamic analysis module, a remote cloud data terminal information supervision module, a power consumption cost dynamic analysis module, a graph neural network dynamic analysis process and result module, and a power supply area dynamic analysis process and result module.

[0021] Among them, the actual user load demand dynamic analysis module is any feasible device or apparatus that can obtain and identify signals, etc., which is used to obtain dynamic analysis instructions; identify the power supply area matching the dynamic analysis instructions and the neural network parameters of the actual user load demand dynamic analysis graph, and generate dynamic adaptation instructions for the actual user load demand including the power supply area.

[0022] It should be noted that the dynamic analysis instruction can be a dynamic analysis process of the actual user load demand of the power supply company, the benchmark user load demand of the power supply company, and the establishment of the variable user load demand; the power supply area that matches the signal and the neural network parameters of the dynamic analysis graph of the actual user load demand are determined by identifying the content in the establishment. After the power supply area is identified, there is a dynamic adaptation instruction for the actual user load demand of the preset missing power supply area. After the power supply area is filled in the signal, the dynamic adaptation instruction for the actual user load demand is activated. The generation of the dynamic adaptation instruction for the actual user load demand is completed.

[0023] The actual user load demand dynamic analysis module includes a dynamic analysis instruction acquisition unit, a dynamic analysis model building unit, and an actual user load demand confirmation unit. The dynamic analysis instruction acquisition unit is any feasible device or apparatus capable of acquiring signals, etc., and is mainly used to acquire dynamic analysis instructions; wherein, the dynamic analysis instructions include the power supply company's actual user load demand dynamic analysis process, the power supply company's benchmark user load demand, and the variable user load demand; based on the power supply company's actual user load demand dynamic analysis process, determine whether there are matching actual user load demand dynamic analysis model parameters; when they can be matched, determine the dynamic analysis graph neural network parameters in the actual user load demand dynamic analysis model parameters as the actual user load demand dynamic analysis graph neural network parameters; when they cannot be matched, extract the data in the power supply company's actual user load demand dynamic analysis process and the power supply area in the power supply company's benchmark user load demand to obtain data and power supply areas that meet both the data and the power supply area. The preset dynamic analysis graph neural network parameters for different time periods are sufficient; the benchmark user load demand of the power supply company is input into the preset dynamic analysis graph neural network parameters to obtain the actual user load demand of the dynamic analysis results in different time periods; the preset dynamic analysis graph neural network parameters that are consistent with the dynamic analysis results actual user load demand and the changing user load demand are determined as the actual user load demand dynamic analysis graph neural network parameters; a dynamic analysis model establishment unit is used to establish the power supply company's actual user load demand dynamic analysis process and the power supply company's actual user load demand dynamic analysis process into actual user load demand dynamic analysis model parameters; an actual user load demand confirmation unit is used to confirm the power supply area in the benchmark user load demand of the power supply company, and generate actual user load demand dynamic adaptation instructions according to the power supply area.

[0024] Among them, the remote cloud data terminal information supervision module is any feasible equipment or device that can transmit the model, etc., which is mainly used to preset the nodes and edge parameters of the graph neural network and the actual user load demand adaptation parameters to the power supply company's actual user load demand remote cloud data terminal; based on the actual user load demand dynamic adaptation instructions, obtain the actual user load demand of the power supply company after pre-processing by the remote cloud data terminal; send the actual user load demand of the power supply company to the matching dynamic analysis module according to the power supply area; among them, the data storage and analysis expression of the actual user load demand of the power supply company is power supply area + actual user load demand + average population per household.

[0025] The remote cloud data terminal information supervision module includes a parameter transmission unit; the parameter transmission unit is used to detect whether there are nodes and edge parameters of the graph neural network and actual user load demand adaptation parameters in the remote cloud data terminal matching the power supply area; when it is detected that there are no nodes and edge parameters of the graph neural network and actual user load demand adaptation parameters, the nodes and edge parameters of the graph neural network and actual user load demand adaptation parameters are transmitted to the remote cloud data terminal.

[0026] Among them, the power consumption cost dynamic analysis module is used to receive the actual user load demand of the power supply company whose power supply area is the power consumption cost power supply area, and determine the power consumption cost dynamic analysis process and results that match the actual user load demand of the power supply company by utilizing the adaptation algorithm preset by random forest.

[0027] It should be noted that the adaptation algorithm can be any feasible neural network algorithm. By importing the actual user load demand of the power supply company and the dynamic analysis process and results of electricity consumption cost in different time periods into the neural network algorithm, the training of the neural network is realized, and then the adaptation algorithm preset by random forest is obtained.

[0028] Among them, the graph neural network dynamic analysis process and result module is used to receive the actual user load demand of the power supply company whose power supply area is the public power supply area, and determine the graph neural network dynamic analysis process and results that match the actual user load demand of the power supply company by presetting the dynamic analysis graph neural network parameters.

[0029] It should be noted that the public dynamic analysis model includes the correspondence between the power supply area and the preset dynamic analysis graph neural network parameters, and the preset dynamic analysis graph neural network parameters can be determined by technical personnel in this field through multiple experiments.

[0030] Among them, the power supply area dynamic analysis process and result module is any feasible equipment or device that can perform actual user load demand analysis, etc. It is mainly used to determine the dynamic analysis process and results of the power supply area through the dynamic analysis process and results of electricity consumption cost, the dynamic analysis process and results of the graph neural network, and the dynamic analysis graph neural network parameters of actual user load demand; determine the actual user load demand analysis that matches the dynamic analysis process and results of electricity consumption cost, the dynamic analysis process and results of the graph neural network, and the dynamic analysis process and results of the power supply area; and generate power supply size adjustment instructions by analyzing the actual user load demand and transformation efficiency.

[0031] It should be noted that in the dynamic analysis process and result module of the graph neural network, in the process of dynamic analysis to match the actual user load demand of the power supply company, the design of formulas and models includes graph structure modeling, the propagation process of the graph neural network, dynamic prediction, result matching and model optimization;

[0032] Graph structure modeling: Define the graph structure G, G = (V, E), where V is a node set, representing each user and substation in the power supply area; E is an edge set, representing the power connection relationship between nodes; the initialization of node feature X and edge feature A can be expressed as:

[0033] Where d is the feature dimension of each node;

[0034] Where f is the characteristic dimension of each edge;

[0035] The propagation process of graph neural network: The core of graph neural network is to update the features of nodes through message passing mechanism. Assuming the number of layers of graph neural network is L, the node feature update formula of the lth layer can be expressed as:

[0036]

[0037] in, is the feature representation of node v in layer l, is the set of neighbor nodes of node v, W (l) is the weight matrix of the lth layer, σ is the activation function;

[0038] Dynamic prediction: For time series data, time-dependent modeling can be integrated into the graph neural network. Assuming there is a time series feature T, it is used as an additional input feature to adjust the node features. The final feature of the node is expressed as:

[0039]

[0040] in, represents the feature concatenation operation, is the final feature representation of the L-th layer node v, T v is a time series feature;

[0041] Load demand forecasting: Use the trained graph neural network to predict the load demand at future time points It can be expressed as:

[0042]

[0043] Among them, MLP is a multi-layer perceptron, which is used to map node features to predicted values ​​of load demand;

[0044] Result matching: The error between the predicted result and the actual load demand Y can be measured by the loss function. The commonly used loss function is the mean square error:

[0045]

[0046] in, is the loss value, is the forecast load demand of node v, Y v is the actual load demand of node v;

[0047] Model optimization: by minimizing the loss function The parameters of the graph neural network can be optimized; the optimization process usually uses a gradient descent method, such as the Adam optimizer, to update the weights W in the network (l) and bias terms.

[0048] Through the above formulas and processes, the graph neural network can effectively process the actual user load demand data in the power supply area, and predict and match the load demand through a dynamic analysis process. The specific formulas and methods can be adjusted and optimized according to the needs and data characteristics of the actual application.

[0049] The power supply area dynamic analysis process and result module includes a dynamic analysis process and result determination unit and an actual user load demand interaction curve generation unit; the dynamic analysis process and result determination unit is used to determine the power supply areas that match the power consumption cost dynamic analysis process and results and the graph neural network dynamic analysis process and results respectively; determine the actual user load demand analysis data set that matches the power consumption cost dynamic analysis process and results and the graph neural network dynamic analysis process and results respectively according to the power supply area; wherein the actual user load demand analysis data set includes the allocation principle of training set 6: validation set 2 and test set 2 in the data set; the power consumption cost dynamic analysis process and results and the graph neural network dynamic analysis process and results are brought into the training set 6: validation set 2 and test set 2 allocation principle in the data set, and the respective matching analysis training data are determined from the respective matching actual user load demand analysis data sets; the actual user load demand interaction curve generation unit is used to use the actual user load demand dynamic analysis graph neural network parameters that match the final dynamic analysis results to determine the historical results of different time periods that match the actual user load demand dynamic analysis graph neural network parameters; based on the historical results and the power supply area dynamic analysis process and results, the actual user load demand interaction curve is generated.

[0050] In addition, the embodiment of the present application also provides a method for analyzing the actual number of users of the power grid of a power supply company, such as Figure 2 As shown, the system operation process provided in the embodiment of the present application mainly includes the following steps:

[0051] A100. Determine the power supply area that matches the acquired dynamic analysis instruction and the neural network parameters of the actual user load demand dynamic analysis graph, and generate a dynamic adaptation instruction for the actual user load demand including the power supply area.

[0052] Determine the power supply area and actual user load demand dynamic analysis graph neural network parameters that match the obtained dynamic analysis instructions, specifically including: obtaining dynamic analysis instructions; the dynamic analysis instructions include the power supply company's actual user load demand dynamic analysis process, the power supply company's benchmark user load demand, and the variable user load demand; based on the power supply company's actual user load demand dynamic analysis process, determine whether there are matching actual user load demand dynamic analysis model parameters; when a match can be made, determine the dynamic analysis graph neural network parameters in the actual user load demand dynamic analysis model parameters as the actual user load demand dynamic analysis graph neural network parameters; when a match cannot be made, extract the data in the power supply company's actual user load demand dynamic analysis process and the power supply area in the power supply company's benchmark user load demand to obtain preset dynamic analysis graph neural network parameters for different time periods that both the data and the power supply area meet; input the power supply company's benchmark user load demand into the preset dynamic analysis graph neural network parameters to obtain the actual user load demand of the dynamic analysis results in different time periods; determine the preset dynamic analysis graph neural network parameters whose dynamic analysis results are consistent with the variable user load demand as the actual user load demand dynamic analysis graph neural network parameters.

[0053] A200, in response to the existence of nodes and edge parameters of the graph neural network and actual user load demand adaptation parameters on the remote cloud data end that matches the power supply area, obtain the actual user load demand of the power supply company after preprocessing on the remote cloud data end.

[0054] In response to the existence of nodes, edge parameters and actual user load demand adaptation parameters of the graph neural network in the remote cloud data end that matches the power supply area, it specifically includes: detecting whether there are nodes, edge parameters and actual user load demand adaptation parameters of the graph neural network in the remote cloud data end that matches the power supply area; when it is detected that the nodes, edge parameters and actual user load demand adaptation parameters of the graph neural network do not exist, transmitting the nodes, edge parameters and actual user load demand adaptation parameters of the graph neural network to the remote cloud data end.

[0055] A300. Send the actual user load demand of the power supply company to the matching cloud control terminal according to the power supply area; the data of the actual user load demand of the power supply company is stored and analyzed in the form of power supply area + actual user load demand + average population per household; and the cloud control terminal operation matches the power supply area.

[0056] A400. Receive the dynamic analysis process and results of electricity consumption cost and the dynamic analysis process and results of graph neural network that match the actual user load demand of the power supply company; among them, the power supply area where the dynamic analysis process and results of electricity consumption cost match is the electricity consumption cost power supply area, and the power supply area where the dynamic analysis process and results of graph neural network match is the public power supply area.

[0057] A500. Determine the dynamic analysis process and results of the power supply area through the dynamic analysis process and results of electricity consumption cost, the dynamic analysis process and results of the graph neural network, and the dynamic analysis graph neural network parameters of actual user load demand; determine the actual user load demand analysis that matches the dynamic analysis process and results of electricity consumption cost, the dynamic analysis process and results of the graph neural network, and the dynamic analysis process and results of the power supply area; generate power supply size adjustment instructions by analyzing the actual user load demand and transformation efficiency.

[0058] Determine the analysis training data that match the dynamic analysis process and results of electricity consumption cost, the dynamic analysis process and results of graph neural network, and the dynamic analysis process and results of power supply area, specifically including: determining the power supply area that matches the dynamic analysis process and results of electricity consumption cost, and the dynamic analysis process and results of graph neural network; determine the actual user load demand analysis data set that matches the dynamic analysis process and results of electricity consumption cost, and the dynamic analysis process and results of graph neural network according to the power supply area; wherein the actual user load demand analysis data set includes the allocation principle of training set 6: validation set 2 and test set 2 in the data set; bring the dynamic analysis process and results of electricity consumption cost, and the dynamic analysis process and results of graph neural network into the allocation principle of training set 6: validation set 2 and test set 2 in the data set, and determine the respective matching analysis training data from the respective matching actual user load demand analysis data sets; use the actual user load demand dynamic analysis graph neural network parameters that match the final dynamic analysis results to determine the historical results of different time periods that match the actual user load demand dynamic analysis graph neural network parameters; based on the historical results and the dynamic analysis process and results of the power supply area, generate the actual user load demand interaction curve.

[0059] Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A power supply company power grid actual user number analysis system, characterized in that: The system includes: The actual user load demand dynamic analysis module is used to obtain dynamic analysis instructions; identify the power supply area that matches the dynamic analysis instructions and the neural network parameters of the actual user load demand dynamic analysis graph, and generate a dynamic adaptation instruction for the actual user load demand including the power supply area; A remote cloud data terminal information supervision module, which is used to preset the parameters of the nodes and edges of the graph neural network and the actual user load demand adaptation parameters to the actual user load demand remote cloud data terminal of the power supply company; based on the actual user load demand dynamic adaptation instruction, the actual user load demand of the power supply company preprocessed by the remote cloud data terminal is obtained; according to the power supply area, the actual user load demand of the power supply company is sent to the matching dynamic analysis module; the data of the actual user load demand of the power supply company is stored and analyzed in the form of power supply area + actual user load demand + average population per household; A power consumption cost dynamic analysis module, which is used to receive actual user load demands of a power supply company whose power supply area is a power consumption cost power supply area, and determine a power consumption cost dynamic analysis process and results that match the actual user load demands of the power supply company by using a preset adaptation algorithm of a random forest; A graph neural network dynamic analysis process and result module, which is used to receive actual user load demands of power supply companies whose power supply areas are public power supply areas, and determine the graph neural network dynamic analysis process and results that match the actual user load demands of the power supply company by presetting dynamic analysis graph neural network parameters; A power supply area dynamic analysis process and result module, which is used to determine the power supply area dynamic analysis process and result through the power consumption cost dynamic analysis process and result, the graph neural network dynamic analysis process and result, and the actual user load demand dynamic analysis graph neural network parameters; The dynamic analysis process and result graph of power consumption cost is determined by analyzing the actual user load demand that matches the dynamic analysis process and result of the neural network and the dynamic analysis process and result of the power supply area; and generating power supply size adjustment instructions by analyzing the actual user load demand and the transformer efficiency; The graph neural network dynamic analysis process and result module performs dynamic analysis to match the actual user load requirements of the power supply company. The design of formulas and models includes graph structure modeling, graph neural network propagation process, dynamic prediction, result matching and model optimization. Graph structure modeling: Define the graph structure G, G = (V, E), where V is a node set, representing each user and substation in the power supply area; E is an edge set, representing the power connection relationship between nodes; the initialization of node feature X and edge feature A is expressed as: Where d is the feature dimension of each node; Where f is the characteristic dimension of each edge; The propagation process of graph neural network: The core of graph neural network is to update the features of nodes through message passing mechanism. Assuming the number of layers of graph neural network is L, the node feature update formula of the lth layer is expressed as: in, is the feature representation of node v in layer l, is the set of neighbor nodes of node v, W( l ) is the weight matrix of the lth layer, σ is the activation function; Dynamic prediction: For time series data, time dependency modeling is integrated into the graph neural network. Assuming that there is a time series feature T, it is used as an additional input feature to adjust the node feature. The final feature of the node is expressed as: in, represents the feature concatenation operation, is the final feature representation of the L-th layer node v, T v is a time series feature; Load demand forecasting: Use the trained graph neural network to predict the load demand at future time points Expressed as: Among them, MLP is a multi-layer perceptron, which is used to map node features to predicted values ​​of load demand; Result matching: The error between the predicted result and the actual load demand Y is measured by the loss function. The commonly used loss function is the mean square error: in, is the loss value, is the forecast load demand of node v, Y v is the actual load demand of node v; Model optimization: by minimizing the loss function Optimize the parameters of the graph neural network; the optimization process usually uses the gradient descent method to update the weights W in the network (l) and bias terms.

2. A power supply company power grid actual user number analysis system according to claim 1, characterized in that: The actual user load demand dynamic analysis module includes a dynamic analysis instruction acquisition unit, a dynamic analysis model establishment unit, and an actual user load demand confirmation unit; The dynamic analysis instruction acquisition unit is used to acquire dynamic analysis instructions; wherein the dynamic analysis instructions include the power supply company's actual user load demand dynamic analysis process, the power supply company's benchmark user load demand, and the variable user load demand; based on the power supply company's actual user load demand dynamic analysis process, determine whether there are matching actual user load demand dynamic analysis model parameters; when they can be matched, determine the dynamic analysis graph neural network parameters in the actual user load demand dynamic analysis model parameters as the actual user load demand dynamic analysis graph neural network parameters; when they cannot be matched, extract the data in the power supply company's actual user load demand dynamic analysis process and the power supply area in the power supply company's benchmark user load demand, so as to obtain the preset dynamic analysis graph neural network parameters for different time periods that both the data and the power supply area meet; Inputting the power supply company's benchmark user load demand into the preset dynamic analysis graph neural network parameters to obtain the actual user load demand of the dynamic analysis results in different time periods; Determine the preset dynamic analysis graph neural network parameters that are consistent with the dynamic analysis result actual user load demand and the changed user load demand as the actual user load demand dynamic analysis graph neural network parameters; The dynamic analysis model establishment unit is used to establish the actual user load demand dynamic analysis process of the power supply company and the actual user load demand dynamic analysis process of the power supply company into actual user load demand dynamic analysis model parameters; The actual user load demand confirmation unit is used to confirm the power supply area in the power supply company's benchmark user load demand, and generate an actual user load demand dynamic adaptation instruction based on the power supply area.

3. A power supply company power grid actual user number analysis system according to claim 1, characterized in that: The remote cloud data terminal information supervision module includes a parameter transmission unit; The parameter transmission unit is used to detect whether there are nodes, edge parameters and actual user load demand adaptation parameters of the graph neural network in the remote cloud data terminal that matches the power supply area; when it is detected that there are no nodes, edge parameters and actual user load demand adaptation parameters of the graph neural network, the nodes, edge parameters and actual user load demand adaptation parameters of the graph neural network are transmitted to the remote cloud data terminal.

4. A power supply company power grid actual user number analysis system according to claim 1, characterized in that: The power supply area dynamic analysis process and result module includes a dynamic analysis process and result determination unit and an actual user load demand interaction curve generation unit; The dynamic analysis process and result determination unit is used to determine the power supply areas that match the power consumption cost dynamic analysis process and result, and the graph neural network dynamic analysis process and result respectively; Based on the power supply area, determine the actual user load demand analysis data set that matches the power consumption cost dynamic analysis process and results and the graph neural network dynamic analysis process and results respectively; wherein the actual user load demand analysis data set includes the allocation principle of training set 6: verification set 2 and test set 2 in the data set; bring the power consumption cost dynamic analysis process and results and the graph neural network dynamic analysis process and results into the allocation principle of training set 6: verification set 2 and test set 2 in the data set, and determine the respective matching analysis training data from the respective matching actual user load demand analysis data sets; The actual user load demand interaction curve generating unit is used to determine the historical results of different time periods that match the actual user load demand dynamic analysis graph neural network parameters using the actual user load demand dynamic analysis graph neural network parameters that match the final dynamic analysis results; and generate the actual user load demand interaction curve based on the historical results and the power supply area dynamic analysis process and results.

5. A method for analyzing the actual number of users of a power grid of a power supply company, characterized in that: The method includes: Determine the power supply area and the actual user load demand dynamic analysis graph neural network parameters that match the obtained dynamic analysis instructions, and generate a dynamic adaptation instruction containing the actual user load demand of the power supply area; In response to the presence of nodes and edge parameters of the graph neural network and actual user load demand adaptation parameters on the remote cloud data end that matches the power supply area, obtain the actual user load demand of the power supply company preprocessed by the remote cloud data end; The actual user load demand of the power supply company is sent to the matching cloud control terminal according to the power supply area; the data of the actual user load demand of the power supply company is stored and analyzed in the form of power supply area + actual user load demand + average population per household; and the cloud control terminal operation matches the power supply area; Receive the power consumption cost dynamic analysis process and results and the graph neural network dynamic analysis process and results that match the actual user load demand of the power supply company; the power supply area where the power consumption cost dynamic analysis process and results match is the power consumption cost power supply area, and the power supply area where the graph neural network dynamic analysis process and results match is the public power supply area; The dynamic analysis process and results of the power consumption cost dynamic analysis process and results, the graph neural network dynamic analysis process and results, and the actual user load demand dynamic analysis graph neural network parameters are used to determine the dynamic analysis process and results of the power consumption cost dynamic analysis process and results, the graph neural network dynamic analysis process and results, and the power supply area dynamic analysis process and results that match each other to analyze the actual user load demand; and the power supply size adjustment instruction is generated by analyzing the actual user load demand and the transformer efficiency.

6. A method for analyzing the actual number of users of a power supply company's power grid according to claim 5, characterized in that: Determine the power supply area that matches the obtained dynamic analysis instructions and the neural network parameters of the actual user load demand dynamic analysis graph, including: Obtain dynamic analysis instructions; wherein the dynamic analysis instructions include the dynamic analysis process of the actual user load demand of the power supply company, the benchmark user load demand of the power supply company, and the variable user load demand; Based on the actual user load demand dynamic analysis process of the power supply company, determining whether there are matching actual user load demand dynamic analysis model parameters; When the match is found, the dynamic analysis graph neural network parameters in the actual user load demand dynamic analysis model parameters are determined to be the actual user load demand dynamic analysis graph neural network parameters; When a match cannot be made, extract the data in the dynamic analysis process of the actual user load demand of the power supply company and the power supply area in the benchmark user load demand of the power supply company to obtain the preset dynamic analysis graph neural network parameters for different time periods that are satisfied by both the data and the power supply area; input the benchmark user load demand of the power supply company into the preset dynamic analysis graph neural network parameters to obtain the actual user load demand of the dynamic analysis results for different time periods; determine that the preset dynamic analysis graph neural network parameters that are consistent with the actual user load demand of the dynamic analysis result and the changed user load demand are the actual user load demand dynamic analysis graph neural network parameters.

7. A method for analyzing the actual number of users of a power supply company's power grid according to claim 5, characterized in that: In response to the existence of nodes and edge parameters of the graph neural network and actual user load demand adaptation parameters at the remote cloud data end matching the power supply area, specifically including: Detect whether there are nodes, edge parameters and actual user load demand adaptation parameters of the graph neural network in the remote cloud data terminal that matches the power supply area; when it is detected that there are no nodes, edge parameters and actual user load demand adaptation parameters of the graph neural network, adapt the nodes, edge parameters and actual user load demand adaptation parameters of the graph neural network to the remote cloud data terminal.

8. A method for analyzing the actual number of users of a power supply company's power grid according to claim 5, characterized in that: Determine the analysis training data that matches the dynamic analysis process and results of electricity consumption cost, the dynamic analysis process and results of graph neural network, and the dynamic analysis process and results of power supply area, including: Determine the power supply areas where the power consumption cost dynamic analysis process and results, and the graph neural network dynamic analysis process and results respectively match; based on the power supply areas, determine the actual user load demand analysis data set that matches the power consumption cost dynamic analysis process and results, and the graph neural network dynamic analysis process and results respectively; wherein the actual user load demand analysis data set includes the allocation principle of training set 6: validation set 2 and test set 2 in the data set; bring the power consumption cost dynamic analysis process and results, and the graph neural network dynamic analysis process and results into the allocation principle of training set 6: validation set 2 and test set 2 in the data set, and determine the respective matching analysis training data from the respective matching actual user load demand analysis data sets; Utilizing the neural network parameters of the actual user load demand dynamic analysis graph that matches the final dynamic analysis results, determine the historical results of different time periods that match the neural network parameters of the actual user load demand dynamic analysis graph; based on the historical results and the dynamic analysis process and results of the power supply area, generate an actual user load demand interaction curve.

Images