Adaptive configuration method, device and equipment of JP cabinet and storage medium

Abstract

This application discloses an adaptive configuration method, apparatus, device, and storage medium for JP cabinets, belonging to the field of JP cabinet load configuration technology. This application obtains a main node graph characterizing the real-time operating status of a distribution transformer area and a secondary node graph characterizing the future operating trend and adjustability of the distribution transformer area; it performs cross-graph feature fusion on the main node graph and the secondary node graph, and outputs adaptive configuration instructions for the nodes where each JP cabinet in the distribution transformer area is located based on the cross-graph feature fusion result; it sends the adaptive configuration instructions to the corresponding JP cabinets for execution and collects the operating status data after execution; based on the operating status data, it synchronously updates the main node graph and the secondary node graph, and continuously optimizes the configuration parameters of each JP cabinet based on the updated main node graph and secondary node graph, thereby achieving the effect of accurately allocating the load of the JP cabinets.

Description

Technical Field

[0001] This application relates to the field of JP cabinet load configuration technology, and in particular to an adaptive configuration method, apparatus, device and storage medium for JP cabinets. Background Technology

[0002] The JP cabinet (Junction and Protection Cabinet) is a core node device in a distribution transformer substation, undertaking important functions such as reactive power compensation, power quality regulation, data acquisition, and communication aggregation. With the large-scale integration of distributed photovoltaic systems and the random fluctuations in electric vehicle charging loads, the operating status of distribution transformer substations exhibits strong volatility and uncertainty, placing dynamic adaptation requirements on the configuration parameters of the JP cabinet (such as capacitor bank switching strategies, voltage regulation levels, and protection settings).

[0003] Currently, the configuration of JP cabinets mainly adopts preset fixed parameters or relies on manual periodic adjustments, which has the following technical defects: First, once the configuration parameters are set, they remain unchanged for a long time and cannot adapt to rapid changes in load and distributed power sources, leading to frequent problems such as voltage exceeding limits and power factor not meeting standards; Second, there is a lack of ability to predict the future operating trends of the distribution area, and configuration adjustments often lag behind actual operating condition changes, resulting in poor adjustment effects; Third, there is a lack of coordination mechanism between JP cabinets, and isolated adjustments may cause reactive power circulation or regulation oscillations within the distribution area.

[0004] Therefore, there is an urgent need for a method that can integrate real-time status with future trends and achieve adaptive configuration of JP cabinets in order to improve the operating economy and power supply reliability of distribution substations. Summary of the Invention

[0005] The main objective of this application is to provide an adaptive configuration method, apparatus, device, and storage medium for a JP cabinet, aiming to solve the technical problems of poor operating economy and power supply reliability in current distribution substations.

[0006] To achieve the above objectives, this application provides an adaptive configuration method for a JP cabinet, the adaptive configuration method of which includes the following steps: Obtain a main node diagram representing the real-time operating status of the distribution transformer area and a secondary node diagram representing the future operating trend and adjustability of the distribution transformer area; Cross-graph feature fusion is performed between the main node graph and the secondary node graph, and based on the result of the cross-graph feature fusion, an adaptive configuration instruction for the node where each JP cabinet in the distribution area is located is output. The adaptive configuration command is sent to the corresponding JP cabinet for execution, and the running status data after execution is collected; Based on the operational status data, the main node diagram and the secondary node diagram are updated synchronously, and the configuration parameters of each JP cabinet are continuously optimized based on the updated main node diagram and secondary node diagram.

[0007] In one embodiment, after the steps of obtaining the main node diagram characterizing the real-time operating status of the distribution transformer area and the secondary node diagram characterizing the future operating trend and adjustability of the distribution transformer area, the method further includes: Using all JP cabinets and related equipment in the distribution substation as nodes and electrical connection lines as edges, the operating status data of each node is collected at the first frequency as node attributes to construct a master node graph that represents the real-time operating status of the distribution substation. Construct a secondary node graph that shares the same set of nodes and edges as the main node graph. Associate each node in the secondary node graph with configurable resource attributes, predictive trend attributes, time period features, and regional characteristic labels. Update the configurable resource attributes at a second frequency lower than the first frequency. Update the predictive trend attributes at a third frequency based on the configurable resource data, the time period features, the regional characteristic labels, and the running status data in the main node graph.

[0008] In one embodiment, the step of updating the predicted trend attribute by self-inference at a third frequency based on the configurable resource data, the time period features, the regional feature labels, and the running status data in the master node graph includes: Based on the operating status data in the main node diagram, the historical load data of each node in the sub-node diagram is determined, and combined with the time cycle characteristics and regional characteristic labels of the current moment, a load prediction sequence for the future preset time period is generated. Historical irradiance data and weather forecast data matching the regional characteristic labels of each node in the sub-node graph are obtained to generate a photovoltaic power output prediction sequence for a future preset period. Based on the load forecast sequence and the photovoltaic output forecast sequence, the node voltage change trend for the future preset period is deduced; The prediction trend attribute is updated based on the load prediction sequence, the photovoltaic output prediction sequence, and the node voltage change trend.

[0009] In one embodiment, after the step of self-inferentially updating the predicted trend attribute, the method further includes: Every preset number of self-inferential update cycles, the actual running data collected by the main node graph and the predicted trend attributes of the sub-node graph at the same moment are obtained, and the residual between the two is calculated. Based on the residual, a predicted initial value is generated to correct the next self-deduction cycle; wherein, if the master node graph detects a node voltage over-limit, load change, or equipment abnormality, the corresponding node's sub-node graph is immediately triggered for emergency calibration to directly update the predicted trend attribute of that node.

[0010] In one embodiment, the step of outputting adaptive configuration instructions for the nodes where each JP cabinet is located within the distribution radio area based on the cross-graph feature fusion result includes: Obtain the pre-divided adaptive configuration blocks in the sub-node graph, wherein the adaptive configuration blocks are multiple configuration blocks obtained by dividing the graph based on the electrical distance between nodes and the similarity of corresponding regional characteristic labels. The cross-graph feature fusion step is executed independently within each adaptive configuration block, and adaptive configuration instructions for each node within each adaptive configuration block are generated.

[0011] In one embodiment, after the step of generating adaptive configuration instructions for each node within each adaptive configuration block, the method further includes: Identify the boundary nodes between adjacent adaptive configuration blocks and determine the predicted trend attributes of each boundary node; If the predicted trend attribute of the adjacent adaptive configuration blocks is power supply and demand imbalance, a collaborative configuration instruction for allocating the adjacent adaptive configuration blocks is generated, taking the adjacent adaptive configuration blocks as a whole and aiming to minimize the total network loss of the distribution area.

[0012] In one embodiment, the step of performing cross-graph feature fusion between the main node graph and the secondary node graph includes: Based on the master node graph, the state attributes of each master node's neighbors are aggregated to determine the real-time state embedding vector of each master node; Based on the sub-node graph, the predicted trend attributes of each sub-node's neighbors are aggregated to determine the predicted trend embedding vector of each sub-node. Based on the correspondence between each master node and each sub-node, the attention weight between the real-time state embedding vector and the predicted trend embedding vector at each node is calculated, and the predicted trend embedding vector is weighted and fused into the real-time state embedding vector according to the attention weight to generate the fused feature vector of each node.

[0013] Furthermore, to achieve the above objectives, this application also provides an adaptive configuration device for a JP cabinet, the adaptive configuration device for the JP cabinet comprising: The data acquisition module is used to collect real-time status data of the smart lock and image data of its surrounding environment in real time through a sensor array deployed on the smart lock when any person triggers the security verification process of the smart lock of the metering box. The reasoning module is used to determine whether there is an abnormal state at the smart lock based on the real-time status data and the image data. The early warning module is used to extract key information of the abnormal state if it exists, generate corresponding operation and maintenance early warning information based on the key information, and send the operation and maintenance early warning information to relevant personnel so that the relevant personnel can perform real-time operation and maintenance on the metering box.

[0014] In addition, to achieve the above objectives, this application also provides an adaptive configuration device for a JP cabinet, the adaptive configuration device for the JP cabinet comprising: a memory, a processor, and an adaptive configuration program for the JP cabinet stored in the memory and executable on the processor, the adaptive configuration program for the JP cabinet being configured to implement the steps of the adaptive configuration method for the JP cabinet as described above.

[0015] In addition, to achieve the above objectives, this application also provides a computer-readable storage medium storing an adaptive configuration program for a JP cabinet, wherein the adaptive configuration program for the JP cabinet, when executed by a processor, implements the steps of the adaptive configuration method for the JP cabinet as described above.

[0016] One or more technical solutions proposed in this application have at least the following technical effects: By acquiring a main node diagram representing the real-time operating status of the distribution substation and a secondary node diagram representing the future operating trend and adjustability of the distribution substation; performing cross-graph feature fusion on the main node diagram and the secondary node diagram, and outputting adaptive configuration instructions for the nodes where each JP cabinet is located within the distribution substation based on the result of the cross-graph feature fusion; issuing the adaptive configuration instructions to the corresponding JP cabinets for execution, and collecting the operating status data after execution; synchronously updating the main node diagram and the secondary node diagram based on the operating status data, and continuously optimizing the configuration parameters of each JP cabinet based on the updated main node diagram and secondary node diagram. That is, by acquiring a main node diagram representing the real-time operating status of the substation and a secondary node diagram representing the future operating trend and adjustability, a digital node diagram model of the distribution substation is constructed, providing adaptive configuration for the JP cabinets. Based on the spatiotemporal fusion data foundation, the system achieves deep coupling between real-time operating conditions and future prediction information when performing cross-graph feature fusion between the main node graph and the secondary node graph. This enables the output adaptive configuration commands to not only respond to the current state but also predict and proactively address upcoming load fluctuations or changes in distributed power output, significantly improving the accuracy and timeliness of configuration decisions. Simultaneously, the configuration commands are sent to the corresponding JP cabinets for execution, and post-execution data is collected, ensuring the effective implementation of the configuration commands. Furthermore, the main node graph and secondary node graph are synchronously updated based on the post-execution operating status data, allowing the graph model to continuously absorb the latest operating information and iteratively optimize subsequent configuration parameters. This endows the JP cabinet configuration with the adaptive capability of dynamically evolving with the operating characteristics of the distribution area, effectively solving the lag and mismatch problems of the traditional static configuration mode, reducing distribution area line losses, and improving voltage qualification rate and distributed energy absorption capacity. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating the adaptive configuration method for the JP cabinet in this application (Example 1). Figure 2 A flowchart illustrating the adaptive configuration method for the JP cabinet in this application, as provided in Embodiment 2. Figure 3This is a schematic diagram of the module structure of the adaptive configuration device for the JP cabinet in an embodiment of this application; Figure 4 This is a schematic diagram of the hardware operating environment involved in the adaptive configuration method of the JP cabinet in this application embodiment.

[0020] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0022] Reference Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the adaptive configuration method for the JP cabinet in this application.

[0023] In the first embodiment, the adaptive configuration method of the JP cabinet includes the following steps: S10, Obtain the main node diagram representing the real-time operating status of the distribution radio station area and the secondary node diagram representing the future operating trend and adjustability of the distribution radio station area. S20, perform cross-graph feature fusion between the main node graph and the secondary node graph, and output the adaptive configuration instruction of the node where each JP cabinet in the distribution area is located based on the result of the cross-graph feature fusion. S30, the adaptive configuration command is sent to the corresponding JP cabinet for execution, and the running status data after execution is collected; S40, based on the operating status data, synchronously update the main node diagram and the secondary node diagram, and continuously optimize the configuration parameters of each JP cabinet based on the updated main node diagram and secondary node diagram. In this embodiment, a dual-map model of the distribution transformer area is first constructed. The distribution transformer area refers to the power supply area from the low-voltage side of the distribution transformer to the user, such as a residential community or industrial park covered by a 10kV / 0.4kV transformer.

[0024] The master node diagram is a graph structure used to represent the real-time operating status of the distribution area. Its nodes are all JP cabinets and other key equipment (such as transformers and branch boxes) within the distribution area, and the edges are electrical connection lines. The node attributes of the master node diagram are real-time operating data collected at a frequency of seconds or minutes, including three-phase voltage, three-phase current, active power, reactive power, power factor, etc.

[0025] The secondary node graph is another graph that shares the same set of nodes and edges as the primary node graph. It is used to characterize the future operating trend and adjustability of the transformer area. Its node attributes include configurable resource attributes (such as capacitor bank capacity, photovoltaic inverter adjustable reactive power margin), prediction trend attributes (such as load prediction curve for the next 15 minutes to 2 hours, photovoltaic output prediction curve), time period characteristics (such as season, weekday / holiday), and regional characteristic labels (such as residential area, commercial area, industrial area).

[0026] Subsequently, after acquiring the two graphs, the main node graph and the secondary node graph are input into the cross-graph feature fusion module. This module extracts real-time state features from the main node graph and predicted trend features from the secondary node graph using a graph neural network (GNN), and then uses an attention mechanism to weightedly fuse the predicted trend features into the real-time state features, generating a fused feature vector for each node. The fused feature vector comprehensively reflects both "the current state" and "what is about to happen." Based on the fused features, adaptive configuration instructions for each node (i.e., the location of each JP cabinet) are output, such as capacitor bank switching instructions (connecting 1 group, disconnecting 2 groups), on-load tap changer adjustment instructions, and protection setting modification instructions.

[0027] Commands are sent to the corresponding JP cabinet's intelligent controller via wireless communication networks (such as 4G, 5G, or fiber optics) for execution. After the JP cabinet executes the command, its internal sensors collect post-execution operational status data, such as voltage, power factor, and actual capacitor bank switching status, and transmit this data back to the system. Simultaneously, based on the transmitted operational status data, the main node diagram (replacing old data with new real-time data) and the secondary node diagram (re-deducing and predicting trend attributes based on the new operational status) are updated synchronously. The updated dual diagrams are then input into the fusion module, forming a closed-loop iteration that continuously optimizes the configuration parameters of each JP cabinet, enabling the configuration to dynamically adjust according to changes in the operating status of the distribution area.

[0028] In this embodiment, the step of performing cross-graph feature fusion between the main node graph and the secondary node graph includes: Based on the master node graph, the state attributes of each master node's neighbors are aggregated to determine the real-time state embedding vector of each master node; Based on the sub-node graph, the predicted trend attributes of each sub-node's neighbors are aggregated to determine the predicted trend embedding vector of each sub-node. Based on the correspondence between each master node and each sub-node, the attention weight between the real-time state embedding vector and the predicted trend embedding vector at each node is calculated, and the predicted trend embedding vector is weighted and fused into the real-time state embedding vector according to the attention weight to generate the fused feature vector of each node.

[0029] In this embodiment, a method combining graph neural networks and attention mechanisms can be used to dynamically combine and apply the feature information from the main node graph and the secondary node graph.

[0030] For the master node graph, a graph convolution operation is performed on each node to aggregate the state attributes of each master node's neighbors. For example, for node i, its neighbor set is N(i), and the aggregation method is: h_i_main = σ( W_main * (x_i ||aggregate({x_j for j in N(i)})) ), where x_i is the real-time state attribute of the node (such as voltage, power), aggregate is the summation or averaging function, and σ is the activation function. The output h_i_main is the real-time state embedding vector of the node, with a dimension of 64, which integrates the real-time state information of itself and its neighbors.

[0031] For the sub-node graph, a graph neural network with the same structure can be used to aggregate the predicted trend attributes of each sub-node's neighbors. The predicted trend attributes of the nodes include load prediction sequences, photovoltaic prediction sequences, voltage trend sequences, etc., and these sequences are concatenated as the initial features of the nodes. After graph convolution, a predicted trend embedding vector h_i_sub for each node is obtained, which is also 64-dimensional and integrates the predicted trend information of itself and its neighbors.

[0032] Then, an attention mechanism is used to fuse the two embedding vectors. For each node i, the attention weight between the real-time state embedding vector h_i_main and the predicted trend embedding vector h_i_sub is calculated. Specifically, α_i = softmax( v^T * tanh( W_main * h_i_main + W_sub * h_i_sub ) ), where v, W_main, and W_sub are learnable parameters. α_i is a scalar representing the weight that the predicted trend information should carry during fusion.

[0033] Finally, the predicted trend embedding vector is weighted by attention weights and fused to the real-time state embedding vector to generate a fused feature vector: z_i = (1 - α_i) * h_i_main + α_i * h_i_sub. The fused feature vector z_i combines the current real-time state (reflecting "what is happening now") and the future predicted trend (reflecting "what will happen soon"), serving as input for subsequent generation of adaptive configuration instructions. This fusion mechanism enables configuration decisions to respond to immediate operating conditions while also proactively adjusting, avoiding the problem of configuration lagging behind operational changes in traditional methods.

[0034] This embodiment acquires a main node diagram representing the real-time operating status of a distribution transformer substation and a secondary node diagram representing the future operating trend and adjustability of the substation. It then performs cross-graph feature fusion on the main and secondary node diagrams and outputs adaptive configuration instructions for the nodes where each JP cabinet within the distribution transformer substation is located based on the fusion results. These adaptive configuration instructions are then sent to the corresponding JP cabinets for execution, and post-execution operating status data is collected. Based on the operating status data, the main and secondary node diagrams are synchronously updated, and the configuration parameters of each JP cabinet are continuously optimized based on the updated diagrams. In other words, by acquiring the main node diagram representing the real-time operating status of the substation and the secondary node diagram representing the future operating trend and adjustability, a digital node diagram model of the distribution transformer substation is constructed, providing a spatiotemporal fusion data foundation for the adaptive configuration of JP cabinets. By fusing features across the main node graph and the secondary node graph, a deep coupling between real-time operating conditions and future prediction information is achieved. This enables the output adaptive configuration commands to not only respond to the current state but also to predict and proactively address upcoming load fluctuations or changes in distributed power output, significantly improving the accuracy and timeliness of configuration decisions. Simultaneously, the configuration commands are sent to the corresponding JP cabinets for execution, and post-execution data is collected, ensuring the effective implementation of the configuration commands. Furthermore, the main node graph and secondary node graph are synchronously updated based on the post-execution operating status data, allowing the graph model to continuously absorb the latest operating information and iteratively optimize subsequent configuration parameters. This endows the JP cabinet configuration with the adaptive capability of dynamically evolving with the operating characteristics of the distribution area, effectively solving the lag and mismatch problems of the traditional static configuration mode, reducing distribution area line losses, and improving voltage qualification rate and distributed energy absorption capacity. like Figure 2 As shown, based on the first embodiment, a second embodiment of the adaptive configuration method for the JP cabinet of this application is proposed. In this embodiment, the method further includes: S110, taking all JP cabinets and related equipment in the distribution substation area as nodes, electrical connection lines as edges, and collecting the operating status data of each node at the first frequency as node attributes, a master node graph representing the real-time operating status of the distribution substation area is constructed. S120, construct a secondary node graph that shares the same set of nodes and edges as the main node graph, associate each node in the secondary node graph with configurable resource attributes, prediction trend attributes, time period features, and regional characteristic labels, update the configurable resource attributes at a second frequency lower than the first frequency, and update the prediction trend attributes at a third frequency based on the configurable resource data, the time period features, the regional characteristic labels, and the running status data in the main node graph through self-inference.

[0035] In this embodiment, when constructing the main node graph, all JP cabinets, distribution transformers, branch boxes, and other equipment within the distribution substation area are abstracted as nodes, for example, with 15 nodes. The electrical connection lines (including cables and busbars) between the devices are abstracted as edges, forming an undirected or directed graph. Operating status data for each node, including RMS voltage, RMS current, active power, and reactive power, are collected at a first frequency (e.g., once every 15 seconds). This data is used as node attributes to construct the main node graph G_main.

[0036] Furthermore, when constructing the secondary node graph, a secondary node graph G_sub is created that shares the exact same set of nodes and edges as the primary node graph. The following attributes are associated with each node in G_sub: Among them, the configurable resource attributes are the adjustable resources of the JP cabinet at this node, including the capacity of the capacitor bank (e.g., a total of 120kVar, divided into 4 groups), the reactive power capacity of the photovoltaic inverter (e.g., ±50kVar), the capacity of the energy storage system (e.g., 200kWh), the number of adjustable transformer taps (e.g., ±5 taps), etc. This attribute is updated at a second frequency, for example, once every 15 minutes, because the resource capacity changes slowly (e.g., changes in the switching status of the capacitor bank, changes in the SOC of energy storage).

[0037] Among them, the predicted trend attribute refers to the operating trend of this node over a future period of time, such as the load forecast curve for the next hour (one point every 5 minutes), the photovoltaic output forecast curve, and the node voltage change trend. This attribute is updated automatically at a third frequency (e.g., every 5 minutes).

[0038] Among them, the time cycle characteristics are: the time type to which the current moment belongs, including season (spring / summer / autumn / winter), month, weekday / weekend, and time period of day (peak / off-peak / valley), etc.

[0039] Among them, the regional characteristic label is the load type label of the region where the node is located, such as residential area, commercial area, industrial area, agricultural area, etc., which is used to guide the selection of parameters for the prediction model.

[0040] It should be noted that configurable resource attributes are updated at a lower frequency (second frequency) because the physical capacity of the resources themselves changes slowly; predictive trend attributes are updated at a medium frequency (third frequency) through self-inference, balancing prediction accuracy and computational cost. Self-inference updates rely on configurable resource data, time period characteristics, regional characteristic labels, and the latest operational status data in the master node graph.

[0041] In this embodiment, the step of updating the predicted trend attribute by self-inference at a third frequency based on the configurable resource data, the time period features, the regional feature labels, and the running status data in the master node graph includes: Based on the operating status data in the main node diagram, the historical load data of each node in the sub-node diagram is determined, and combined with the time cycle characteristics and regional characteristic labels of the current moment, a load prediction sequence for the future preset time period is generated. Historical irradiance data and weather forecast data matching the regional characteristic labels of each node in the sub-node graph are obtained to generate a photovoltaic power output prediction sequence for a future preset period. Based on the load forecast sequence and the photovoltaic output forecast sequence, the node voltage change trend for the future preset period is deduced; The prediction trend attribute is updated based on the load prediction sequence, the photovoltaic output prediction sequence, and the node voltage change trend.

[0042] In this embodiment, the main focus is on the self-inferential update process of the predicted trend attribute. Specifically, an update is triggered at a third frequency (e.g., every 5 minutes). For each node in the sub-node graph, the following steps are performed: The first step is load forecasting. Extract the node's past operating status data (e.g., load data from the past 24 hours) from the master node graph. Combine this with the current time period characteristics (e.g., "10 AM on a summer weekday") and regional characteristic labels (e.g., "commercial area"), and call upon a load forecasting model matching that node (e.g., a long-short cycle neural network or a time series model; this is a general model for power grid load forecasting and is not specifically limited). The model outputs a load forecast sequence for a future preset time period (e.g., the next hour, with a step size of 5 minutes), which is a vector of 12 points, where each point represents the predicted average active power of the node within those 5 minutes.

[0043] The second step is photovoltaic (PV) output prediction. If the node is connected to distributed PV, historical irradiance data (obtained from the local weather station) and future weather forecast data (such as total irradiance forecast) matching the node's regional characteristic labels are obtained. Combined with the PV module's installed capacity and conversion efficiency model, a PV output prediction sequence (also a vector of 12 points) for the future preset period is generated. If the node is not connected to PV, the output prediction sequence is zero.

[0044] The third step is voltage trend extrapolation. Using the current measured node voltage as the initial state, the load forecast sequence and photovoltaic output forecast sequence are input into the power flow calculation model of the distribution network (such as the forward backward substitution method or the Newton-Raphson method; this model is also a general model and is not specifically limited). Combined with the line impedance parameters of the feeder where the node is located, the voltage change trend of the node in the future preset period is extrapolated to obtain the voltage prediction sequence.

[0045] The fourth step is to update the predicted trend attribute. The load forecast sequence, photovoltaic output forecast sequence, and voltage change trend sequence are combined and stored as the predicted trend attribute of this node for subsequent cross-graph feature fusion.

[0046] In this embodiment, after the step of self-inferentially updating the predicted trend attribute, the method further includes: Every preset number of self-inferential update cycles, the actual running data collected by the main node graph and the predicted trend attributes of the sub-node graph at the same moment are obtained, and the residual between the two is calculated. Based on the residual, a predicted initial value is generated to correct the next self-deduction cycle; wherein, if the master node graph detects a node voltage over-limit, load change, or equipment abnormality, the corresponding node's sub-node graph is immediately triggered for emergency calibration to directly update the predicted trend attribute of that node.

[0047] In this embodiment, a closed-loop correction mechanism for the prediction model is introduced to improve prediction accuracy. By setting a correction cycle, for example, every 5 self-inference update cycles (i.e. every 25 minutes), residual calculation and correction are performed once.

[0048] Specifically, at the correction time, the actual operating data (such as actual load power and actual voltage) collected from the main node diagram at that time is first obtained, along with the predicted trend attributes (i.e., the values ​​predicted before that time) from the corresponding time in the sub-node diagram. The residual between the two is then calculated (actual value minus predicted value). For example, for load forecasting, the difference ΔP between the actual load and the predicted load is calculated; for voltage forecasting, the difference ΔU between the actual voltage and the predicted voltage is calculated.

[0049] Based on the calculated residuals, initial prediction values ​​are generated to correct the next self-inference cycle. Correction methods can include adding the residuals as bias terms to the initial state of the next prediction model, or using them to update the prediction model parameters online (e.g., fine-tuning model weights through Kalman filtering or gradient descent), enabling subsequent predictions to gradually eliminate systematic prediction biases.

[0050] In addition, an emergency calibration mechanism is set up. When the master node diagram detects an abnormal event in real time, such as node voltage exceeding the limit (voltage below 198V or above 242V), load change (load changes by more than 30% within 1 minute), or equipment abnormality (JP cabinet communication interruption, capacitor bank failure to operate, etc.), the corresponding sub-node diagram is immediately triggered to perform emergency calibration.

[0051] This emergency calibration skips regular periodic updates and directly re-infers the predictive trend attributes of the node based on the latest actual data and abnormal event types. This ensures that the prediction can respond quickly in abnormal situations and provides accurate trend information for subsequent configuration decisions.

[0052] In this embodiment, the step of outputting adaptive configuration instructions for the nodes of each JP cabinet within the distribution radio area based on the cross-graph feature fusion result includes: Obtain the pre-divided adaptive configuration blocks in the sub-node graph, wherein the adaptive configuration blocks are multiple configuration blocks obtained by dividing the graph based on the electrical distance between nodes and the similarity of corresponding regional characteristic labels. The cross-graph feature fusion step is executed independently within each adaptive configuration block, and adaptive configuration instructions for each node within each adaptive configuration block are generated.

[0053] In this embodiment, an adaptive configuration block partitioning mechanism is also introduced to decompose large-scale transformer areas into multiple local optimization regions, reduce computational complexity and improve the targeting of configuration instructions. Adaptive configuration block partitioning is performed in advance in the sub-node graph.

[0054] Specifically, this classification is based on two dimensions: The first is the electrical distance between nodes, which is the degree of proximity between nodes in the electrical topology. It is usually measured by the impedance matrix or the shortest path length. Nodes with close electrical distances have a greater influence on each other and are suitable to be assigned to the same block. Second, the similarity of regional characteristic labels, that is, whether adjacent nodes belong to the same load type (such as both being residential areas or both being industrial areas). Nodes of the same type have similar load characteristics and are suitable for unified optimization.

[0055] Among them, clustering algorithms (such as spectral clustering or similarity values ​​based on electrical distance) can be used to divide the nodes of the entire station area into multiple configuration blocks, and the nodes in each block have high cohesion in terms of electrical and load characteristics.

[0056] Furthermore, after partitioning, the cross-graph feature fusion step is executed independently within each adaptive configuration block. This means that the cross-graph feature fusion module processes only the nodes and their internal edges within a block, without performing feature fusion across blocks. Within each block, the fused feature vectors of each node are generated according to the spatiotemporal fusion method of the main node graph and the secondary node graph in the first embodiment. Based on the fused features, the adaptive configuration instructions for each node are output. The independent computation within the block significantly reduces the scale of a single fusion inference, while avoiding mutual interference between different blocks, making the configuration instructions more aligned with the operating characteristics of the local region.

[0057] In this embodiment, after the step of generating adaptive configuration instructions for each node within each adaptive configuration block, the method further includes: Identify the boundary nodes between adjacent adaptive configuration blocks and determine the predicted trend attributes of each boundary node; If the predicted trend attribute of the adjacent adaptive configuration blocks is power supply and demand imbalance, a collaborative configuration instruction for allocating the adjacent adaptive configuration blocks is generated, taking the adjacent adaptive configuration blocks as a whole and aiming to minimize the total network loss of the distribution area.

[0058] In this embodiment, in order to prevent the overall performance of the substation from degrading due to independent block optimization, after the blocks are divided, the boundary nodes between adjacent adaptive configuration blocks are identified. The boundary nodes are those nodes that have electrical connections with nodes in adjacent blocks (i.e., there are edges in the graph). For example, if node N1 in block A and node N2 in block B have a line connection, then N1 and N2 are both boundary nodes.

[0059] In addition, it is necessary to obtain the predicted trend attributes of these boundary nodes, including load forecast, photovoltaic output forecast, and voltage forecast. By analyzing the power exchange at the boundaries of adjacent blocks, it is possible to determine whether there is a power supply-demand imbalance. For example, if block A predicts a large photovoltaic output surplus (supply exceeds demand) in the next 15 minutes, while block B predicts high load demand (demand exceeds supply) in the next 15 minutes, then there is a power supply-demand imbalance between the adjacent blocks.

[0060] Specifically, when an imbalance is detected, collaborative optimization is performed, treating two adjacent blocks (such as block A and block B) as a whole, and setting the optimization objective as minimizing the total network loss of the distribution substation (or other objectives, such as minimizing voltage deviation). Under overall constraints (including adjustable resource capacity of JP cabinets and line transmission capacity limits within each block), collaborative configuration instructions are generated.

[0061] The coordinated configuration command may include: activating the capacitor banks of some JP cabinets within Block A to support voltage; adjusting the voltage regulation settings of some JP cabinets within Block B to accommodate power from Block A; or adjusting the state of boundary switches to achieve reasonable power flow. Furthermore, after the coordinated command is issued to the relevant JP cabinets for execution, it can eliminate power imbalances between blocks, reduce network losses across the entire distribution area, and prevent overall performance degradation due to localized optimization.

[0062] This embodiment uses all JP cabinets and associated equipment within the distribution substation as nodes and electrical connection lines as edges. Operating status data of each node is collected at a first frequency as node attributes to construct a main node graph representing the real-time operating status of the distribution substation. A secondary node graph sharing the same set of nodes and edges as the main node graph is constructed. Each node in the secondary node graph is associated with configurable resource attributes, predicted trend attributes, time period characteristics, and regional characteristic labels. The configurable resource attributes are updated at a second frequency lower than the first frequency. The predicted trend attributes are updated at a third frequency based on the configurable resource data, the time period characteristics, the regional characteristic labels, and the operating status data in the main node graph. This method of updating the main and secondary node graphs with varying frequencies, combined with corresponding feature configurations, ensures the accuracy of the main and secondary node graphs while reducing their resource requirements.

[0063] Furthermore, embodiments of this application also propose an adaptive configuration device for a JP cabinet, referring to... Figure 3 The adaptive configuration device of the JP cabinet includes: The acquisition module 10 is used to acquire the main node diagram representing the real-time operating status of the distribution radio area and the sub-node diagram representing the future operating trend and adjustability of the distribution radio area. The processing module 20 is used to perform cross-graph feature fusion between the main node graph and the secondary node graph, and output adaptive configuration instructions for the nodes where each JP cabinet in the distribution area is located based on the cross-graph feature fusion result. The acquisition module 30 is used to send the adaptive configuration command to the corresponding JP cabinet for execution and to collect the running status data after execution. The optimization module 40 is used to synchronously update the main node diagram and the secondary node diagram based on the operating status data, and to continuously optimize the configuration parameters of each JP cabinet based on the updated main node diagram and secondary node diagram.

[0064] This embodiment acquires a main node diagram representing the real-time operating status of a distribution transformer substation and a secondary node diagram representing the future operating trend and adjustability of the substation. It then performs cross-graph feature fusion on the main and secondary node diagrams and outputs adaptive configuration instructions for the nodes where each JP cabinet within the distribution transformer substation is located based on the fusion results. These adaptive configuration instructions are then sent to the corresponding JP cabinets for execution, and post-execution operating status data is collected. Based on the operating status data, the main and secondary node diagrams are synchronously updated, and the configuration parameters of each JP cabinet are continuously optimized based on the updated diagrams. In other words, by acquiring the main node diagram representing the real-time operating status of the substation and the secondary node diagram representing the future operating trend and adjustability, a digital node diagram model of the distribution transformer substation is constructed, providing a spatiotemporal fusion data foundation for the adaptive configuration of JP cabinets. By fusing features across the main node graph and the secondary node graph, a deep coupling between real-time operating conditions and future prediction information is achieved. This enables the output adaptive configuration commands to not only respond to the current state but also to predict and proactively address upcoming load fluctuations or changes in distributed power output, significantly improving the accuracy and timeliness of configuration decisions. Simultaneously, the configuration commands are sent to the corresponding JP cabinets for execution and data is collected after execution, ensuring the effective implementation of the configuration commands. Furthermore, the main node graph and the secondary node graph are updated synchronously based on the operational status data after execution, allowing the graph model to continuously absorb the latest operational information and continuously correct subsequent configuration parameters through iterative optimization. This endows the JP cabinet configuration with the adaptive capability of dynamically evolving with the operating characteristics of the distribution area, effectively solving the lag and mismatch problems of the traditional static configuration mode, reducing distribution area line losses, and improving voltage qualification rate and distributed energy absorption capacity.

[0065] It should be noted that each module in the above-mentioned device can be used to implement each step in the above-mentioned method and achieve the corresponding technical effect. This embodiment will not elaborate further here.

[0066] Reference Figure 4 , Figure 4 This is a schematic diagram of the hardware operating environment of the device involved in the embodiments of this application.

[0067] like Figure 4As shown, the device may include: a processor 1001, such as a CPU; a communication bus 1002; a user interface 1003; a network interface 1004; and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen or an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The memory 1005 may be high-speed RAM or non-volatile memory, such as a disk drive. Optionally, the memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0068] Those skilled in the art will understand that Figure 4 The structure shown does not constitute a limitation on the device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0069] like Figure 4 As shown, the memory 1005, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an adaptive configuration program for the JP cabinet.

[0070] exist Figure 4 In the device shown, network interface 1004 is mainly used for data communication with external networks; user interface 1003 is mainly used for receiving user input commands; the device calls the adaptive configuration program of the JP cabinet stored in memory 1005 through processor 1001 and performs the following operations: Obtain a main node diagram representing the real-time operating status of the distribution transformer area and a secondary node diagram representing the future operating trend and adjustability of the distribution transformer area; Cross-graph feature fusion is performed between the main node graph and the secondary node graph, and based on the result of the cross-graph feature fusion, an adaptive configuration instruction for the node where each JP cabinet in the distribution area is located is output. The adaptive configuration command is sent to the corresponding JP cabinet for execution, and the running status data after execution is collected; Based on the operational status data, the main node diagram and the secondary node diagram are updated synchronously, and the configuration parameters of each JP cabinet are continuously optimized based on the updated main node diagram and secondary node diagram.

[0071] Furthermore, the processor 1001 can call the adaptive configuration program of the JP cabinet stored in the memory 1005, and also perform the following operations: Using all JP cabinets and related equipment in the distribution substation as nodes and electrical connection lines as edges, the operating status data of each node is collected at the first frequency as node attributes to construct a master node graph that represents the real-time operating status of the distribution substation. Construct a secondary node graph that shares the same set of nodes and edges as the main node graph. Associate each node in the secondary node graph with configurable resource attributes, predictive trend attributes, time period features, and regional characteristic labels. Update the configurable resource attributes at a second frequency lower than the first frequency. Update the predictive trend attributes at a third frequency based on the configurable resource data, the time period features, the regional characteristic labels, and the running status data in the main node graph.

[0072] Furthermore, the processor 1001 can call the adaptive configuration program of the JP cabinet stored in the memory 1005, and also perform the following operations: Based on the operating status data in the main node diagram, the historical load data of each node in the sub-node diagram is determined, and combined with the time cycle characteristics and regional characteristic labels of the current moment, a load prediction sequence for the future preset time period is generated. Historical irradiance data and weather forecast data matching the regional characteristic labels of each node in the sub-node graph are obtained to generate a photovoltaic power output prediction sequence for a future preset period. Based on the load forecast sequence and the photovoltaic output forecast sequence, the node voltage change trend for the future preset period is deduced; The prediction trend attribute is updated based on the load prediction sequence, the photovoltaic output prediction sequence, and the node voltage change trend.

[0073] Furthermore, the processor 1001 can call the adaptive configuration program of the JP cabinet stored in the memory 1005, and also perform the following operations: Every preset number of self-inferential update cycles, the actual running data collected by the main node graph and the predicted trend attributes of the sub-node graph at the same moment are obtained, and the residual between the two is calculated. Based on the residual, a predicted initial value is generated to correct the next self-deduction cycle; wherein, if the master node graph detects a node voltage over-limit, load change, or equipment abnormality, the corresponding node's sub-node graph is immediately triggered for emergency calibration to directly update the predicted trend attribute of that node.

[0074] Furthermore, the processor 1001 can call the adaptive configuration program of the JP cabinet stored in the memory 1005, and also perform the following operations: Obtain the pre-divided adaptive configuration blocks in the sub-node graph, wherein the adaptive configuration blocks are multiple configuration blocks obtained by dividing the graph based on the electrical distance between nodes and the similarity of corresponding regional characteristic labels. The cross-graph feature fusion step is executed independently within each adaptive configuration block, and adaptive configuration instructions for each node within each adaptive configuration block are generated.

[0075] Furthermore, the processor 1001 can call the adaptive configuration program of the JP cabinet stored in the memory 1005, and also perform the following operations: Identify the boundary nodes between adjacent adaptive configuration blocks and determine the predicted trend attributes of each boundary node; If the predicted trend attribute of the adjacent adaptive configuration blocks is power supply and demand imbalance, a collaborative configuration instruction for allocating the adjacent adaptive configuration blocks is generated, taking the adjacent adaptive configuration blocks as a whole and aiming to minimize the total network loss of the distribution area.

[0076] Furthermore, the processor 1001 can call the adaptive configuration program of the JP cabinet stored in the memory 1005, and also perform the following operations: Based on the master node graph, the state attributes of each master node's neighbors are aggregated to determine the real-time state embedding vector of each master node; Based on the sub-node graph, the predicted trend attributes of each sub-node's neighbors are aggregated to determine the predicted trend embedding vector of each sub-node. Based on the correspondence between each master node and each sub-node, the attention weight between the real-time state embedding vector and the predicted trend embedding vector at each node is calculated, and the predicted trend embedding vector is weighted and fused into the real-time state embedding vector according to the attention weight to generate the fused feature vector of each node.

[0077] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0078] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0079] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the adaptive configuration method of the JP cabinet in the above embodiments.

[0080] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0081] The aforementioned computer-readable storage medium may be included in the adaptive configuration device of the JP cabinet; or it may exist independently and not be assembled into the adaptive configuration device of the JP cabinet.

[0082] The aforementioned computer-readable storage medium carries one or more programs, which, when executed by the adaptive configuration device of the JP cabinet, cause the adaptive configuration device of the JP cabinet to: Obtain a main node diagram representing the real-time operating status of the distribution transformer area and a secondary node diagram representing the future operating trend and adjustability of the distribution transformer area; Cross-graph feature fusion is performed between the main node graph and the secondary node graph, and based on the result of the cross-graph feature fusion, an adaptive configuration instruction for the node where each JP cabinet in the distribution area is located is output. The adaptive configuration command is sent to the corresponding JP cabinet for execution, and the running status data after execution is collected; Based on the operational status data, the main node diagram and the secondary node diagram are updated synchronously, and the configuration parameters of each JP cabinet are continuously optimized based on the updated main node diagram and secondary node diagram.

[0083] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0084] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0085] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0086] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the adaptive configuration method of the JP cabinet described above, thereby solving the technical problem of adaptive configuration of the JP cabinet. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the adaptive configuration method of the JP cabinet provided in the above embodiments, and will not be repeated here.

[0087] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0088] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0089] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0090] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. An adaptive configuration method for a JP cabinet, characterized in that, The adaptive configuration method for the JP cabinet includes the following steps: obtaining a main node diagram representing the real-time operating status of the distribution transformer area and a secondary node diagram representing the future operating trend and adjustability of the distribution transformer area; performing cross-graph feature fusion on the main node diagram and the secondary node diagram, and outputting an adaptive configuration instruction for the node where each JP cabinet is located within the distribution transformer area based on the result of the cross-graph feature fusion; sending the adaptive configuration instruction to the corresponding JP cabinet for execution, and collecting the operating status data after execution; synchronously updating the main node diagram and the secondary node diagram based on the operating status data, and continuously optimizing the configuration parameters of each JP cabinet based on the updated main node diagram and secondary node diagram.

2. The method as described in claim 1, characterized in that, After the step of obtaining the main node graph representing the real-time operating status of the distribution substation and the secondary node graph representing the future operating trend and adjustability of the distribution substation, the method further includes: taking all JP cabinets and associated equipment in the distribution substation as nodes and electrical connection lines as edges, collecting the operating status data of each node at a first frequency as node attributes, and constructing a main node graph representing the real-time operating status of the distribution substation; constructing a secondary node graph that shares the same set of nodes and edge sets as the main node graph, associating each node in the secondary node graph with configurable resource attributes, predicted trend attributes, time period characteristics, and regional characteristic labels, updating the configurable resource attributes at a second frequency lower than the first frequency, and updating the predicted trend attributes at a third frequency based on the configurable resource data, the time period characteristics, the regional characteristic labels, and the operating status data in the main node graph through self-inference.

3. The method as described in claim 2, characterized in that, The step of updating the prediction trend attribute by self-deduction at a third frequency based on the configurable resource data, the time period characteristics, the regional characteristic labels, and the operating status data in the main node graph includes: determining the historical load data of each node in the sub-node graph based on the operating status data in the main node graph, and generating a load prediction sequence for a future preset period by combining the time period characteristics and regional characteristic labels of the current moment; obtaining historical irradiance data and weather forecast data of each node in the sub-node graph that match the regional characteristic labels, and generating a photovoltaic output prediction sequence for the future preset period; deducing the node voltage change trend for the future preset period based on the load prediction sequence and the photovoltaic output prediction sequence; and updating the prediction trend attribute based on the load prediction sequence, the photovoltaic output prediction sequence, and the node voltage change trend.

4. The method as described in claim 2, characterized in that, After the step of updating the predicted trend attribute through self-inference, the method further includes: every preset number of self-inference update cycles, acquiring the actual operating data collected by the main node graph and the predicted trend attribute of the sub-node graph at the same time, and calculating the residual between the two; generating a predicted initial value for correcting the next self-inference cycle based on the residual; wherein, if the main node graph detects a node voltage exceeding the limit, a sudden load change, or an equipment abnormality, it immediately triggers the sub-node graph of the corresponding node to perform emergency calibration to directly update the predicted trend attribute of that node.

5. The method as described in claim 2, characterized in that, The step of outputting adaptive configuration instructions for the nodes of each JP cabinet in the distribution area based on the cross-graph feature fusion result includes: obtaining pre-divided adaptive configuration blocks in the sub-node graph, wherein the adaptive configuration blocks are multiple configuration blocks divided according to the electrical distance between nodes in the sub-node graph and the similarity of corresponding regional characteristic labels; independently executing the cross-graph feature fusion step in each adaptive configuration block, and generating adaptive configuration instructions for each node in each adaptive configuration block.

6. The method as described in claim 5, characterized in that, After the step of generating adaptive configuration instructions for each node within each adaptive configuration block, the method further includes: determining the boundary nodes between adjacent adaptive configuration blocks and determining the predicted trend attributes of each boundary node; if the predicted trend attribute of the adjacent adaptive configuration blocks is power supply and demand imbalance, taking the adjacent adaptive configuration blocks as a whole and aiming to minimize the total network loss of the distribution area, generating a coordinated configuration instruction for coordinating the adjacent adaptive configuration blocks.

7. The method as described in claim 1, characterized in that, The step of performing cross-graph feature fusion between the main node graph and the sub-node graph includes: aggregating the state attributes of each main node's neighbors according to the main node graph to determine the real-time state embedding vector of each main node; aggregating the predicted trend attributes of each sub-node's neighbors according to the sub-node graph to determine the predicted trend embedding vector of each sub-node; calculating the attention weight between the real-time state embedding vector and the predicted trend embedding vector at each node according to the correspondence between the main nodes and the sub-nodes, and fusion the predicted trend embedding vector into the real-time state embedding vector according to the attention weight to generate the fused feature vector of each node.

8. An adaptive configuration device for a JP cabinet, characterized in that, The adaptive configuration device of the JP cabinet includes: an acquisition module for acquiring a main node diagram representing the real-time operating status of the distribution transformer area and a secondary node diagram representing the future operating trend and adjustability of the distribution transformer area; a processing module for performing cross-graph feature fusion on the main node diagram and the secondary node diagram, and outputting adaptive configuration instructions for the nodes where each JP cabinet in the distribution transformer area is located based on the cross-graph feature fusion result; an acquisition module for sending the adaptive configuration instructions to the corresponding JP cabinet for execution, and acquiring the operating status data after execution; and an optimization module for synchronously updating the main node diagram and the secondary node diagram based on the operating status data, and continuously optimizing the configuration parameters of each JP cabinet based on the updated main node diagram and secondary node diagram.

9. An adaptive configuration device for a JP cabinet, characterized in that, The adaptive configuration device of the JP cabinet includes: a memory, a processor, and an adaptive configuration program for the JP cabinet stored in the memory and executable on the processor, wherein the adaptive configuration program for the JP cabinet is configured to implement the steps of the adaptive configuration method for the JP cabinet as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium stores a program that implements the adaptive configuration method of the JP cabinet, and the program that implements the adaptive configuration method of the JP cabinet is executed by a processor to implement the steps of the adaptive configuration method of the JP cabinet as described in any one of claims 1 to 7.

Images