Regional resource distribution characteristic analysis method for power grid power supply structure and load

Abstract

The invention discloses a regional resource distribution characteristic analysis method for a power grid power supply structure and a load, and relates to the technical field of characteristic analysis, and the method comprises the following steps: extracting change data of power supply output and load demand based on a regional resource distribution model, and forming a power supply-load time sequence characteristic pair; based on the power supply-load time sequence feature pair, analyzing the spatial mismatch between the power supply output and the load demand, and constructing a spatial distribution difference mode of the power supply and the load; calculating the accessibility and matching degree of power supply-load based on a spatial distribution difference mode in combination with power grid topology and power transmission constraints to obtain a realizable distribution pattern; and based on the distribution pattern, extracting coupling characteristics of the power supply structure and load distribution, and outputting a regional distribution characteristic analysis result. According to the method, the distribution pattern can be realized by extracting the time sequence coupling characteristics of the power supply and the load and combining the power grid topology constraint analysis, so that the problem that regional power supply-load mismatch is difficult to quantify and dynamically evaluate is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of feature analysis, and more particularly, to a regional resource distribution feature analysis method for power supply structure and load of a power grid. BACKGROUND

[0002] With the continuous expansion of new energy access scale and the rapid development of distributed power supply, the power supply structure and load distribution of modern power grid present high non-equilibrium and dynamic change characteristics. The traditional power grid planning and operation analysis method mainly relies on overall capacity statistics, single node load prediction or static topology analysis, usually ignoring the spatial distribution difference, time sequence characteristics and coupling relationship between adjacent regions of regional internal power supply and load. This leads to the fact that the existing method is difficult to comprehensively quantify the matching degree and potential mismatch risk between regional power supply and load, and also cannot accurately evaluate the actual effect of energy storage arrangement, renewable energy consumption capacity and load regulation strategy. In addition, the existing method lacks effective comprehensive analysis means when dealing with power output fluctuation, load demand peak-valley mismatch and dynamic change trend, and it is difficult to form a reliable index system reflecting the power supply-load coupling characteristics at the regional level. Therefore, there is an urgent need for a method that can combine power supply structure, load distribution and its time and space dynamic characteristics to systematically analyze the regional power grid resource distribution characteristics, so as to provide a scientific basis for power grid planning, operation optimization and regulation strategy, and realize accurate evaluation of regional power grid supply-demand matching, resource optimization potential and operation safety.

[0003] In the above disclosed technical solution, there are at least the following technical problems: the existing power grid planning and operation analysis method usually only focuses on the total capacity distribution of power supply and load, lacks quantitative analysis of the spatial difference and dynamic coupling characteristics of power supply and load within the region, and leads to the difficulty in identifying local supply-demand mismatch, evaluating resource optimization potential, and limiting the scientific formulation of optimization strategies such as energy storage regulation, distributed power supply dispatching and load migration.

[0004] In view of the above problems, the present application provides a solution. SUMMARY

[0005] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present application provide a regional resource distribution feature analysis method for power supply structure and load of a power grid, which can realize distribution pattern by extracting the time sequence coupling characteristics of power supply and load and combining power grid topology constraint analysis, so as to solve the problems of difficult quantification and dynamic evaluation of regional power supply-load mismatch.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solutions: The power grid power supply structure and load regional resource distribution characteristic analysis method comprises the following steps: based on a regional resource distribution model, extracting power output and load demand change data to form a power-load time sequence characteristic pair; based on the power-load time sequence characteristic pair, analyzing the spatial mismatch of power output and load demand, and constructing a power and load spatial distribution difference mode; based on the spatial distribution difference mode, combining the power grid topology and power transmission constraints, calculating the power-load reachability and matching degree to obtain an achievable distribution pattern; based on the distribution pattern, extracting the coupling characteristics of the power supply structure and the load distribution, and outputting the regional distribution characteristic analysis result.

[0007] In a preferred embodiment, the regional resource distribution model is established based on power supply data and load operation data of the target power grid, specifically: collecting operation parameters of the target power grid, the operation parameters including power supply data and load operation data; performing regional grid division according to the spatial structure and grid node layout of the target power grid; in each grid cell, constructing a quantitative feature vector reflecting the power and load state in the cell based on the operation parameters; integrating the quantitative feature vectors to generate a regional resource distribution model covering the target power grid.

[0008] In a preferred embodiment, the power-load time sequence characteristic pair is specifically formed in the following steps: Based on the regional resource distribution model, obtaining power output and load demand time sequence data of each grid cell and preprocessing the data; extracting power dynamic coupling characteristics and load dynamic coupling characteristics in each grid cell from the preprocessed time sequence data through multi-scale decomposition and nonlinear time alignment; pairing the power dynamic coupling characteristics and the load dynamic coupling characteristics according to the time correspondence to form a power-load time sequence characteristic pair.

[0009] In a preferred embodiment, the power dynamic coupling characteristics and the load dynamic coupling characteristics in each grid cell are extracted from the preprocessed time sequence data through multi-scale decomposition and nonlinear time alignment, specifically: calculating the spatial weight of the cell according to the power capacity density and load intensity of each grid cell in the regional resource distribution model; based on the spatial weight, performing multi-scale decomposition on the time sequence signals of the power output and the load demand, in the decomposition process, dynamically adjusting the weight of feature extraction according to the distribution difference of adjacent grid cells to extract initial features; performing alignment processing on the initial features to fuse the spatio-temporal correlation of adjacent cells; extracting dynamic coupling characteristics representing the coupling relationship between power and load on the aligned sequence.

[0010] In a preferred embodiment, the step of analyzing the spatial mismatch between power output and load demand based on power-time sequence feature pairs specifically involves: calculating the cell-level mismatch index vector within each grid cell based on the power-load time sequence feature pairs; and performing a difference operation on the mismatch index vectors of adjacent cells based on spatial proximity to generate an inter-grid coupling deviation matrix.

[0011] In a preferred embodiment, the construction of the spatial distribution difference pattern between power supply and load specifically involves: constructing a weighted graph model with topological constraints based on the coupling deviation matrix; performing dynamic spatial clustering on the weighted graph model within a sliding time window to obtain the temporally evolving spatial cluster partitioning results; weighted fusion of the cell-level mismatch indices of all grid cells within each spatial cluster to generate representative feature vectors for each spatial cluster; and aggregating the representative feature vectors of the dynamic clusters to form a spatiotemporal coupling difference pattern characterizing power supply and load.

[0012] In a preferred embodiment, the step of performing dynamic spatial clustering on the weighted graph model within a sliding time window to obtain a temporally evolved spatial cluster partitioning result specifically involves: performing initial clustering of grid cells based on the weighted graph edge weights of the current time window; merging adjacent cells according to a preset intra-cluster consistency condition to form an optimized spatial cluster partition; applying temporal stability constraints to the spatial cluster partitioning results of adjacent time windows to minimize changes in cluster structure over time; based on the spatial cluster partitioning of the current time window, fusing the mismatch indices of each cell within the cluster to update the representative feature vector of the cluster; and outputting the spatial cluster partitioning and its representative feature vector under each time window to form a temporally evolved spatial cluster partitioning result.

[0013] In a preferred embodiment, the step of combining the power grid topology and transmission constraints to calculate the accessibility and matching degree of power sources and loads to obtain an achievable distribution pattern specifically involves: determining the accessibility matrix between power source nodes and load nodes that take into account transmission constraints based on the spatial distribution difference pattern and the power grid topology; calculating the matching degree of connected power source-load nodes in the accessibility matrix based on the spatial distribution difference pattern to generate a matching degree matrix; and selecting power source-load combinations that meet the preset matching degree requirements based on the accessibility matrix and the matching degree matrix to form an achievable distribution pattern.

[0014] In a preferred embodiment, the coupling characteristics of the power supply structure and the load distribution based on the distribution pattern are extracted, and the regional distribution characteristic analysis result is output, specifically: based on the realizable distribution pattern, the effective supply-demand pair formed by the power supply node and the corresponding load node under the power transmission constraint is obtained; for each effective supply-demand pair, the time sequence coupling index is calculated according to the time sequence characteristics, and the spatial coupling degree of the supply-demand pair is generated in combination with the matching degree; based on the spatial coupling degree of each effective supply-demand pair, the time sequence complementarity and the local consumption potential of the power supply at the regional level are aggregated and calculated; and the regional distribution characteristic analysis result containing the spatial coupling degree, the time sequence complementarity and the local consumption potential is output.

[0015] The technical effects and advantages of the regional resource distribution characteristic analysis method of the power grid power supply structure and load of the application are as follows: 1. The application can quantify the dynamic matching of regional power supply and load in terms of peak-valley synchronization, wave complementarity and change rate, etc. by constructing a spatial coupling analysis system based on the time sequence characteristics of power output and load demand, accurately identifying the mismatch characteristics and evolution law of regional power supply-load, and providing reliable data support for power grid planning, energy storage layout and load regulation.

[0016] 2. The application can filter out the realizable power supply-load distribution pattern by combining the power grid topology and the power transmission constraint, and extract the spatial coupling degree, the time sequence complementarity and the local consumption potential of the power supply on this basis, realize the comprehensive evaluation of the regional level power supply structure and load distribution characteristics, and provide scientific basis for new energy consumption capacity analysis and operation optimization. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 The flowchart of the regional resource distribution characteristic analysis method of the power grid power supply structure and load of the application is shown. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0019] Embodiment 1, Figure 1 The regional resource distribution characteristic analysis method of the power grid power supply structure and load of the application is given, including the following steps: S1, based on the regional resource distribution model, the change data of power output and load demand are extracted to form a power-load time sequence characteristic pair; The regional resource distribution model is established based on the power supply data and load operation data of the target power grid; The power supply data includes installed capacity, output curve and geographical position of distributed power supply and centralized power supply, and the load operation data includes capacity, power consumption curve and power supply range of partitioned load; The regional resource distribution model is specifically: The power supply data and load operation data of the target power grid are collected, and regional grids are divided according to the spatial structure and grid node layout of the target power grid; In each grid unit, the owned power supply capacity, load capacity and their time sequence characteristics are counted, and a quantified feature vector of the unit is constructed; The quantified feature vectors of all grid units are integrated to form a regional resource distribution model covering the entire target region.

[0020] In this embodiment, the regional resource distribution model is used to extract the change data of power supply output and load demand to form a power load time sequence feature pair, which is specifically: Based on the regional resource distribution model, the time sequence data of power supply output and load demand in each grid unit are obtained, and the data are corrected for outliers and completed for missing values; Through multi-scale decomposition (such as within a day, within a week and within a month) and nonlinear time alignment, the dynamic coupling characteristics of power supply and load in each grid unit are extracted, including peak-valley synchronization degree, change rate difference and fluctuation complementarity; The power supply feature vector and the load feature vector of each grid unit are paired according to the time correspondence to form a power load time sequence feature pair that can depict the spatio-temporal coupling relationship between regional power supply output and load demand.

[0021] The multi-scale decomposition and nonlinear time alignment are used to extract the dynamic coupling characteristics of power supply and load in each grid unit, which is specifically: According to the power supply capacity density and load intensity of each grid unit in the regional resource distribution model, the spatial weight of the unit is calculated; Based on the spatial weight, the time sequence signals of power supply output and load demand are subjected to multi-scale decomposition, and in each time scale (such as hour, day and week), the feature extraction weight is dynamically adjusted in combination with the power load distribution difference of adjacent grid units to extract initial features, including peak-valley value, fluctuation amplitude, change rate and the like, so that the obtained feature vector can reflect both the internal dynamic change and the spatial mismatch potential at the regional level; The initial features are aligned through a nonlinear alignment algorithm based on mutual information to synchronize the peak-valley misalignment and fluctuation of each grid unit and its adjacent units; The dynamic coupling characteristics are extracted from the aligned weighted multi-scale sequence, including peak-valley synchronization degree, change rate deviation, fluctuation complementarity and cross-grid complementary potential; The power supply feature vector of each grid cell is paired with the load feature vector in a time corresponding relationship to form a power-load time sequence feature pair that can depict the space-time coupling relationship between regional power output and load demand.

[0022] The power supply capacity density is the total installed capacity of power supply in each grid cell divided by the grid area The power supply capacity density is the total installed capacity of power supply in each grid cell divided by the grid area The total installed capacity of all power supplies in grid k, The actual space area of grid k; The load intensity is the average load demand in each grid cell The load intensity is the average load demand in each grid cell The load intensity is the average load demand in each grid cell The statistical period, The instantaneous load function of grid k at continuous time t (based on discrete instantaneous load operation data), The starting time of the statistical period.

[0023] The power supply capacity density and the load intensity can be used as the weight basis for spatially weighted multi-scale decomposition: Grid power supply capacity density is high, and load intensity is low → feature extraction can increase the power fluctuation weight; Grid load intensity is high, and capacity density is low → feature extraction can increase the load response weight.

[0024] S2, based on the power-load time sequence feature pair, analyze the spatial mismatch between power output and load demand, and construct the spatial distribution difference mode of power and load; The power-load time sequence feature pair is used to analyze the spatial mismatch between power output and load demand, specifically: According to the power-load time sequence feature pair, the peak-valley synchronization degree, fluctuation complementarity and change rate difference of power output and load demand are calculated in each grid cell to obtain a unit-level mismatch index vector; Based on the spatial proximity relationship between each grid cell, the difference degree of the mismatch index vectors of adjacent cells is calculated to form a grid inter-coupling deviation matrix, which quantifies the mismatch degree between adjacent cells.

[0025] The difference degree of the mismatch index vectors of adjacent cells is calculated, specifically:

[0026]

[0027] Wherein, The grid inter-coupling deviation matrix, The Euclidean distance between cell i and cell j,​ , are mismatch index vectors of adjacent cells, respectively, is a spatial weight based on geographical proximity, is the distance between the geographical centers of cell i and cell j.

[0028] The peak-valley synchronization degree is used to characterize the synchronization of the power supply and the load at the time of peak and valley occurrence, specifically:

[0029] The fluctuation complementarity is used to measure whether there is a complementary relationship between the fluctuations of power output and load demand in a short time scale, specifically:

[0030] The change rate difference is used to measure the difference in change trend between power output and load demand, specifically:

[0031] wherein, is the peak-valley synchronization degree, is the number of aligned peak-valley, , are the occurrence times of the kth peak / valley in the power output and load demand, respectively, is the total sampling duration, is the fluctuation complementarity, , are the power output and load demand at time t in the i-th grid cell, respectively, , are the average values of the power output and load demand, respectively, is the change rate difference, is the sampling interval.

[0032] The spatial distribution difference pattern of power supply and load is constructed, specifically: The coupling deviation matrix is constructed as a weighted graph with topological constraints, wherein the edge weight not only contains the coupling deviation value between cells, but also combines the resource structure similarity and spatial proximity of cells to form a comprehensive edge weight; In the sliding time window, the weighted graph is dynamically graph partitioned, and through optimizing the consistency of intra-cluster edge weight and time continuity, the spatial cluster partition results of time evolution are obtained; The unit-level mismatch index vectors in each dynamic cluster are weighted and fused, and the weights are dynamically adjusted according to the unit power supply capacity, load size and mutual coupling strength in the cluster, to obtain a representative feature vector of the cluster; the weighted fusion weights of the unit-level mismatch index vectors are dynamically determined according to the power supply contribution degree (the proportion of the total power supply installed capacity of the unit to the total power supply installed capacity in the cluster), the load demand influence (the proportion of the average load of the unit to the total average load in the cluster) and the cooperative coupling degree (the average mutual information normalized value of the unit and other units in the cluster) in the dynamic cluster, and the greater the value of each dimension index, the higher the corresponding weight; meanwhile, the grid operation state (such as power consumption peak, power supply tension, flat operation) in the sliding time window is combined, the importance of each dimension is balanced through an adjustment coefficient, the sum of the weights of all units in the cluster is 1, so as to accurately match the comprehensive influence of the unit on the power-load mismatch characteristics in the cluster.

[0033] The representative feature vectors of each dynamic cluster are combined to form a final space-time coupling difference mode of power and load, reflecting the power-load mismatch characteristics and evolution rules at the regional level.

[0034] In the sliding time window, the weighted graph is dynamically divided, the spatial cluster division result of time evolution is obtained by optimizing the consistency of edge weight in the cluster and the time continuity, and specifically: In each time window, the grid cells are initially clustered based on the edge weight of the weighted graph, and adjacent cells with edge weight higher than a preset threshold are merged to optimize the consistency in the cluster; The cluster labels of adjacent time windows are adjusted in time continuity, and the cluster label change is minimized to maintain time sequence stability; In each cluster, the unit-level mismatch index vectors are weighted and fused to update the representative feature vector of the cluster, which is used for cluster division in the next time window; The dynamic evolution clusters and their representative features in each time window are output to form the spatial cluster division result of time evolution.

[0035] The comprehensive edge weight is specifically:

[0036]

[0037]

[0038] Wherein, is the comprehensive edge weight, is the value of the grid coupling deviation matrix, is the unit resource structure similarity, is the spatial proximity, , , are the proportionality coefficients (set according to historical data) is the scale parameter of the Gaussian kernel.

[0039] S3, based on the spatial distribution difference mode, combines the power grid topology and transmission constraints to calculate the reachability and matching degree of the power supply-load, and obtains the achievable distribution pattern; In this embodiment, the reachability and matching degree of the power supply-load are calculated to obtain the achievable distribution pattern, specifically: According to the spatial distribution difference mode and the power grid topology structure, the reachability of each power supply node to the load node is determined, and considering the transmission line capacity constraint, the reachability matrix is obtained; For each pair of power supply-load nodes in the reachability matrix, the matching degree is calculated in combination with the spatial distribution difference mode characteristics; According to the reachability matrix and the matching degree matrix, the power supply-load combination that meets the transmission constraint and has high matching degree is screened to form the final achievable distribution pattern.

[0040] The reachability of each power supply node to the load node is determined, specifically: For each power supply unit i and load unit j, it is judged whether there is a power grid topology path and all the residual capacities of the lines passed are greater than the load demand. If the condition is met, the reachability is set to 1, otherwise . .

[0041] The matching degree is specifically:

[0042]

[0043] wherein, is the matching degree, is the weight coefficient, which is between 0 and 1 (set according to the power grid planning standard), is the power supply capacity, is the load demand, is the spatial matching degree, , are the cluster representative feature vectors, , are the cluster representative feature vectors of the grid k and the grid l.

[0044] The role of the achievable distribution pattern is reflected in: Operability guarantee under constraints: By considering the grid topology and transmission line capacity constraints, the selected power-load combinations are truly achievable distribution patterns, ensuring that they will not exceed line capacity or cause power unavailability when running in the actual grid. This ensures that subsequent analysis is based on implementable real-world scenarios rather than theoretically ideal distributions, avoiding planning biases.

[0045] Quantifying spatial mismatch and resource optimization potential: Based on the achievable distribution patterns, it is possible to analyze which areas have insufficient or redundant power-load matching, quantify the degree of spatial mismatch, and provide a basis for grid operation optimization, load regulation, or resource allocation, such as adjusting energy storage, renewable energy utilization, or load migration strategies.

[0046] S4, based on the distribution pattern, extracts the coupling characteristics of power structure and load distribution, and outputs regional distribution characteristic analysis results.

[0047] The coupling characteristics of power structure and load distribution based on the distribution pattern are extracted, and the regional distribution characteristic analysis results are output, specifically: Based on the achievable distribution pattern, the connection relationship between each power node and the corresponding load node, and the effective supply-demand pair under transmission constraints are obtained. Within the effective supply-demand pair, combined with peak-valley synchronization degree, fluctuation complementarity, and change rate difference, the time sequence coupling index of the combination is constructed. The time sequence coupling index and matching degree are combined to obtain the spatial coupling degree of the power-load pair. Based on the spatial coupling degree of the power-load pair, the time sequence complementarity and power on-site consumption potential of the region are calculated. The spatial coupling degree, time sequence complementarity, and on-site consumption potential are output as regional distribution characteristic analysis results.

[0048] The spatial coupling degree is specifically:

[0049]

[0050] The time sequence complementarity is specifically:

[0051] The power on-site consumption potential is specifically:

[0052] Wherein, is the spatial coupling degree, , are preset coupling coefficients, is the matching degree, is the time sequence coupling index, , , is a weight factor (set according to entropy weight method, entropy value is calculated for the distribution of each index on different grids, the smaller the entropy value (the greater the difference), the higher the weight, used for normalization), , is a peak-valley synchronization degree, is a fluctuation complementarity, is a change rate difference, is an effective supply-demand pair quantity, is a timing complementarity, is the available capacity of the power supply unit i, is the demand of the load unit j, is the local consumption potential of the power supply.

[0053] The above formulas are all dimensionless values, and the formulas are obtained by software simulation of a large amount of data to obtain a formula closest to the actual situation, and the preset parameters in the formula are set by a person skilled in the art according to the actual situation.

[0054] The above embodiments can be realized wholly or partially by software, hardware, firmware or any combination thereof. When realized by software, the above embodiments can be realized wholly or partially in the form of a computer program product.

[0055] Those skilled in the art can realize that the modules and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. A person skilled in the art can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0056] In addition, the functional modules in each embodiment of the present application can be integrated in one processing module, or each module can exist physically alone, or two or more modules can be integrated in one module.

[0057] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0058] Finally: the above only for the preferred embodiments of the present application, and not for limiting the present application, any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application, should be included in the scope of protection of the present application.

Claims

1. A method for analyzing the regional resource distribution characteristics of power grid power supply structure and load, characterized in that, The method comprises the following steps: Based on the regional resource distribution model, the variation data of power output and load demand are extracted to form a power-load time sequence feature pair; Based on the power-load time sequence feature pair, the spatial mismatch of power output and load demand is analyzed, and a spatial distribution difference mode of power and load is constructed; Based on the spatial distribution difference mode, combined with the power grid topology and transmission constraints, the reachability and matching degree of power-load are calculated to obtain an achievable distribution pattern; Based on the distribution pattern, the coupling characteristics of power structure and load distribution are extracted, and the regional distribution feature analysis result is output.

2. The method of claim 1, wherein, The regional resource distribution model is established based on power data and load operation data of the target power grid, specifically: Collect the operation parameters of the target power grid, including power data and load operation data; According to the spatial structure and grid node layout of the target power grid, the regional grid is divided; In each grid unit, a quantitative feature vector reflecting the state of power and load in the unit is constructed based on the operation parameters; Integrate the quantitative feature vectors to generate a regional resource distribution model covering the target power grid.

3. The method of claim 2, wherein, The power-load time sequence feature pair is specifically formed as follows: Based on the regional resource distribution model, the time sequence data of power output and load demand of each grid unit are obtained, and the data are preprocessed; Through multi-scale decomposition and nonlinear time alignment, the dynamic coupling characteristics of power and load in each grid unit are extracted from the preprocessed time sequence data; The dynamic coupling characteristics of power and load are paired according to the time correspondence to form a power-load time sequence feature pair.

4. The method of claim 3, wherein, The dynamic coupling characteristics of power and load in each grid unit are extracted from the preprocessed time sequence data by multi-scale decomposition and nonlinear time alignment, specifically: According to the power capacity density and load intensity of each grid unit in the regional resource distribution model, the spatial weight of the unit is calculated; Based on the spatial weight, the time sequence signals of power output and load demand are subjected to multi-scale decomposition, and in the decomposition process, the weight of feature extraction is dynamically adjusted according to the distribution difference of adjacent grid units to extract initial features; The initial features are aligned to fuse the spatio-temporal correlation of adjacent units; On the aligned sequence, the dynamic coupling characteristics representing the coupling relationship between power and load are extracted.

5. The method of claim 4, wherein, Based on the power-load time sequence feature pair, the spatial mismatch of power output and load demand is analyzed, specifically: According to the power-load time sequence feature pair, the unit-level mismatch index vector of each grid unit is calculated; Based on the spatial proximity relationship, the difference degree operation is performed on the mismatch index vectors of adjacent units to generate a grid coupling deviation matrix.

6. The method of claim 5, wherein the method further comprises: The spatial distribution difference mode of power and load is constructed, specifically: Based on the coupling deviation matrix, a weighted graph model with topological constraints is constructed; In a sliding time window, the weighted graph model is subjected to dynamic spatial clustering to obtain a time-evolving spatial cluster division result; The unit-level mismatch indexes of all grid units in each spatial cluster are weighted and fused to generate a representative feature vector of each spatial cluster; The representative eigenvector of the aggregated dynamic cluster is formed to represent the spatial-temporal coupling difference pattern of the power supply and the load.

7. The method of claim 6, wherein the method further comprises: The dynamic spatial clustering is performed on the weighted graph model in the sliding time window to obtain a spatial cluster division result evolving over time, and the method comprises the following steps: initial clustering of the grid cells based on the edge weight of the weighted graph in the current time window; merging adjacent cells according to a preset intra-cluster consistency condition to form an optimized spatial cluster division; applying a temporal stability constraint to the spatial cluster division result of the adjacent time window to minimize the change of the cluster structure over time; updating the representative eigenvector of the cluster based on the spatial cluster division in the current time window and the mismatch index of each cell in the cluster; outputting the spatial cluster division and the representative eigenvector of the cluster in each time window to form the spatial cluster division result evolving over time.

8. The method of claim 7, wherein the method further comprises: The reachability and matching degree of the power supply and the load are calculated based on the power grid topology and the power transmission constraint to obtain an achievable distribution pattern, and the method comprises the following steps: determining a reachability matrix between the power supply nodes and the load nodes considering the power transmission constraint based on the spatial distribution difference pattern and the power grid topology structure; for the connected power supply-load nodes in the reachability matrix, calculating the matching degree thereof based on the spatial distribution difference pattern to generate a matching degree matrix; based on the reachability matrix and the matching degree matrix, screening power supply-load combinations satisfying a preset matching degree requirement to form the achievable distribution pattern.

9. The method of claim 8, wherein, Based on the distribution pattern, the coupling features of the power supply structure and the load distribution are extracted, and a regional distribution feature analysis result is outputted, and the method comprises the following steps: based on the achievable distribution pattern, obtaining an effective supply-demand pair formed by the power supply nodes and the corresponding load nodes under the power transmission constraint; for each effective supply-demand pair, calculating a temporal coupling index according to the temporal characteristics thereof, and generating a spatial coupling degree of the supply-demand pair in combination with the matching degree; based on the spatial coupling degrees of the effective supply-demand pairs, aggregating and calculating a regional-level temporal complementarity and a power self-consumption potential; outputting the regional distribution feature analysis result containing the spatial coupling degree, the temporal complementarity and the self-consumption potential.

Images