Self-adaptive power supply service grid intelligent generation method, system and device

Abstract

The invention relates to a self-adaptive intelligent generation method for a power supply service grid, and the method comprises the following steps: taking a transformer substation and a power user as samples, and carrying out the clustering of a plurality of samples in a power grid based on geographic position information through employing a hierarchical clustering method, so as to obtain a plurality of clusters; extracting a transformer substation-power consumer topology of the power grid, and respectively corresponding the plurality of clusters to corresponding transformer substations according to the topology; and drawing a clustering boundary line of the clustering cluster and the transformer substation based on the boundary points of the clustering cluster, extracting candidate boundary lines in the reference road network by adopting a Canny algorithm, selecting a boundary line from the candidate boundary lines based on the similarity between the clustering boundary line and the candidate boundary lines, and fitting a power supply service grid based on the selected boundary line. The method is effective and reliable, the power transmission range of the transformer substation is obtained according to the transformer-user topology, the power supply service grid matched with the reference road network is accurately fitted, and accurate basic support is provided for further power marketing business.

Description

Technical Field

[0001] The present invention relates to a field, and more specifically, to a method, system and device for intelligently generating an adaptive power supply service grid. Background Art

[0002] "One Grid Map" is an exclusive map of the power system. It can implement personalized functions and provide convenient and fast guidance according to the needs of different departments, levels, and units in the power system. "One Grid Map" is the foundation of the digital twin power grid and the core application of the digital support platform for equipment management. According to the idea of ​​decoupling the power grid resource business middle platform from the GIS platform, it adopts high-performance technologies such as WebGL graphics acceleration, topological partitioning, dynamic vector slicing, and sub-second topological reconstruction to solve the difficulties of cross-regional power grid splicing, power grid panoramic visualization, and map service sharing, and give full play to the ability of data collection and resource sharing to realize the business application sharing of "One Grid Map" among different departments.

[0003] In the construction of "One Map for the Grid", the existing business scope can cover the full voltage levels from 1000 kV to 400 V, vertically connecting multi-form thematic maps such as grid operation diagrams, GIS grid diagrams, regional system diagrams, thermal maps, etc., superimposed with information such as operation management, grid measurement, and meteorological environment, and combined with the work needs of different professional departments, the graphics startup time is improved to milliseconds, ledgers and functional locations are established in one step, and graphic topology is automatically generated.

[0004] The "One Grid Map" is integrated with the equipment management business, and adopts a graphical layering and grading mechanism to achieve differentiated configuration by position and role. By linking graphics and numbers, the defect list and operation tasks are fully displayed to visualize the "family assets" of the power grid. Using the One Grid Map, through human-machine linkage, the real-time situation on the scene is displayed to provide multi-dimensional information for command decision-making. Using the One Grid Map, all-round tracking of emergency materials, personnel, vehicles and other elements is achieved, which effectively enables the long-term planning of the new generation of emergency command system and supports enterprise-level applications such as safety supervision emergency command.

[0005] In addition, "One Grid Map" also provides open and shared visual map services, and releases standard map SDKs, services and components. It uses map customization tools to support the rapid construction of front-end applications and significantly shortens the map configuration development time.

[0006] However, in terms of the visualization map function, the existing power supply service grid construction has the disadvantages of inaccurate "change-to-household" data, imperfect power supply service topology, invisible and difficult to manage. Existing technologies are all explored at the level of simple data governance and information maintenance, without combining the overall business process for overall research, with limited accuracy, and no way to establish a mature operation mechanism during the application process, resulting in low efficiency.

[0007] In view of the above problems, there is an urgent need for an adaptive power supply service grid intelligent generation method, system and device. Summary of the invention

[0008] In order to solve the deficiencies in the prior art, the present invention provides an adaptive power supply service grid intelligent generation method, system and device, which obtains power users and substations in similar locations through a clustering method, and generates an accurate power supply service grid by fitting the cluster edges and the reference road network edges.

[0009] The present invention adopts the following technical solution.

[0010] The first aspect of the present invention relates to an adaptive power supply service grid intelligent generation method, the method comprising the following steps: taking substations and power users as samples, clustering multiple samples in the power grid based on geographic location information using a hierarchical clustering method to obtain multiple cluster clusters; extracting the substation-power user topology of the power grid, and corresponding multiple cluster clusters to corresponding substations according to the topology; drawing cluster boundary lines of cluster clusters and substations based on the boundary points of the cluster clusters, using the Canny algorithm to extract candidate boundary lines in the reference road network, and selecting boundary lines from the candidate boundary lines based on the similarity between the cluster boundary lines and the candidate boundary lines, and fitting the power supply service grid based on the selected boundary lines.

[0011] Preferably, substations and power users are taken as samples, and a hierarchical clustering method is used to cluster multiple samples within the power grid based on geographic location information, including: taking the topological geographic coordinates of substations and power users as sample information, and implementing clustering based on the distance between the topological geographic coordinates of the samples; clustering takes each sample as an initial cluster, gradually reduces the number of initial clusters by merging similar samples, and updates the clusters until the clustering end conditions are met.

[0012] Preferably, the substation-power user topology of the power grid is extracted, and multiple clusters are respectively corresponded to corresponding substations based on the topology, including: corresponding each sample in a current cluster to the substation-power user topology; extracting the substation that falls into all power users or the most power users, and using the substation as the substation corresponding to the current cluster.

[0013] Preferably, the clustering boundary line of the clustering cluster and the substation is drawn based on the boundary points of the clustering cluster, including: constructing a point set with the geographical location information of all samples in the clustering cluster, analyzing the boundary points in the clustering cluster based on the convex hull algorithm, and obtaining the clustering boundary line of the clustering cluster; if the overlapping area between adjacent clustering clusters exceeds the preset range, it is determined that the adjacent clustering clusters correspond to a substation, and the clustering boundary line of each clustering cluster is generated; the clustering boundary lines of multiple clustering clusters corresponding to a substation are extracted to generate the clustering boundary line of the current substation through the clustering boundary lines of multiple clustering clusters.

[0014] Preferably, the Canny algorithm is used to extract the candidate boundary lines in the reference road network, and the candidate boundary lines are selected based on the similarity between the clustering boundary line and the candidate boundary lines, including: obtaining the extraction result of the Canny algorithm, and using the intersection points and the points with the largest curvature change in the extraction result to segment the extraction result and divide the extraction result into multiple candidate line segments; performing any permutation and combination on the multiple candidate line segments, connecting them into multiple non-overlapping or incompletely overlapping candidate paths, and numbering the multiple candidate paths; setting a starting point for any clustering boundary line, traversing each candidate path corresponding to each number from the starting point, and analyzing the similarity between the clustering boundary line and the candidate path; using the candidate path with the largest similarity as the candidate boundary line segment of the current clustering boundary line until the selection of the candidate boundary lines for all substations and all clustering clusters is completed.

[0015] Preferably, a starting point is set for any clustering boundary line, and each candidate path corresponding to each number is traversed from the starting point to analyze the similarity between the clustering boundary line and the candidate path, including: extracting the sampling point loop of the clustering boundary line, and obtaining the sampling point sequence [(x0,y0),(x1,y1),……,(x n ,y n )] of the clustering boundary line by setting the starting point (x0,y0), where n + 1 is the number of sampling points; extracting the sampling points of the candidate path, and arbitrarily selecting two candidate path sampling points (u a ,v a ) and (u b ,v b ), where a < b and the distance between a and b is fixed, calculating the distance D1 between the starting point (x0,y0) and the sampling point (u a ,v a ), the distance D2 between the starting point (x0,y0) and the sampling point (u b ,v b ), and the distance between the current traversed point (x i ,y i ) on the clustering boundary line and the sampling point (u a ,v a), i is between 1 and n, and the current traversal point (x i ,y i ) and sampling point (u b ,v b ) between the distance D4; if min(D1,D2,D3)=D1, then max(D4,D1) is the current traversal point (x i ,y i ) and the distance between the candidate path, and the next traversal point (x i+1 ,y i+1 ) and candidate path sampling points (u a+1 ,v a+1 ) and (u b+1 ,v b+1 ) and the next starting point (x1, y1); if min(D1, D2, D3) = D2, then max(D4, D2) is used as the current traversal point (x i ,y i ) and the candidate path, and perform distance calculation on the next sampling point i+1 in the sampling point sequence of the cluster boundary line; if min(D1,D2,D3)=D3, then take max(D4,D3) as the current traversal point x i The distance between the candidate path and the next sampling point in the candidate path sampling point sequence is recalculated (u b+1 ,v b+1 ) and the current traversal point (x i ,y i ) to update max(D4,D3); gradually select sampling points along the cluster boundary line and the candidate path to perform calculations until the calculation of the distances between all sampling points in the cluster boundary line and the candidate path is completed, and the sum of the distances between all sampling points in the cluster boundary line and the candidate path is used as the similarity value between the cluster boundary line and the candidate path.

[0016] Preferably, a cluster boundary line with the smallest similarity value to a current cluster boundary line is selected as a candidate path with the greatest similarity.

[0017] Preferably, fitting a power supply service grid based on the selected boundary line includes: fitting a power supply service grid with candidate paths, and loading the power supply service grid into the reference road network to generate a power supply service map.

[0018] The second aspect of the present invention relates to an adaptive power supply service grid intelligent generation system using the method in the first aspect of the present invention, the system including a clustering module, an extraction module, and a fitting module; wherein the clustering module is used to take substations and power users as samples, and cluster multiple samples in the power grid based on geographic location information using a hierarchical clustering method to obtain multiple cluster clusters; the extraction module is used to extract the substation-power user topology of the power grid, and correspond the multiple cluster clusters to the corresponding substations according to the topology; the fitting module is used to draw cluster boundary lines of cluster clusters and substations based on the boundary points of the cluster clusters, use the Canny algorithm to extract candidate boundary lines in the benchmark road network, and select boundary lines from the candidate boundary lines based on the similarity between the cluster boundary lines and the candidate boundary lines, and fit the power supply service grid based on the selected boundary lines.

[0019] A third aspect of the present invention relates to a terminal, comprising a processor and a storage medium; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the steps of the method in the first aspect of the present invention.

[0020] The beneficial effect of the present invention is that, compared with the prior art, the adaptive power supply service grid intelligent generation method, system and device in the present invention obtains power users and substations at similar locations through a clustering method, and generates an accurate power supply service grid by fitting the cluster edge and the reference network edge. The present invention is effective and reliable, obtains the power transmission range of the substation according to the transformer-user topology, accurately fits the power supply service grid matching the reference network, and provides accurate basic support for further power marketing business.

[0021] The beneficial effects of the present invention also include:

[0022] 1. The transformer-user topology relationship obtained by clustering provides a reliable basis for the division of the power supply service grid, ensuring that the power supply service grid is automatically and efficiently generated in a way that is closest to the actual business topology. The power supply service grid obtained in this way, on the one hand, supports the precise positioning of individual unreasonable factors in the transformer-user topology relationship in operation, thereby ensuring that the marketing business adjusts the complex and changeable low-voltage topology structure at the lowest cost. On the other hand, the power supply service grid ensures that the division results of the power supply grid are fully in line with the actual marketing business, and provides a sufficiently accurate power terminal access method for the power marketing side, the power dispatching side and typical power users without interrupting the power marketing business.

[0023] 2. Based on the logical premise of transformer-user topology and benchmark road network, the present invention conducts a comprehensive study from the perspective of power grid marketing business. The divided power supply service grid can update the scope of the power supply service grid in real time according to the actual access situation of power users, the increase and decrease of power users' capacity and the withdrawal from the network, and accurately analyze the coverage area of ​​the substation, power transmission loss, real-time distribution redundancy and other parameters based on this, to ensure the visualization and real-time nature of the marketing business.

[0024] 3. Under the current situation that all elements of power supply services such as existing power grid equipment, service resources, customer information, work order events, etc. are deeply integrated and shared in real time to "one map of the power grid", the method of the present invention obtains a dynamically dispatchable flexible power supply service grid by combining virtual and real lines. In the case of overlapping subgrids, the method supports the flexible allocation of dispatch work orders, equipment maintenance events, etc. in overlapping areas based on the dynamic parameters of the substation, thereby implementing granular and refined management of the power system at the most basic level of business implementation with the substation as the unit.

[0025] 4. Map matching is performed between the fitted trajectory path and the reference road network to ensure the consistency of the trajectory data with the actual road. Combining the unsimplified global grid and the intelligent sampling area, the optimal path point set is continuously searched in the intelligent sampling area, and the path is used as the guiding path of the algorithm. In this way, the path can be further optimized to make it closer to the actual road. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 A schematic diagram of the steps of a method for intelligently generating an adaptive power supply service grid according to the present invention;

[0027] Figure 2 This is a schematic diagram of the module structure of an adaptive power supply service grid intelligent generation system of the present invention. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical scheme and advantages of the present invention clearer, the technical scheme of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The embodiments described in the present invention are only embodiments of a part of the present invention, not all embodiments. Based on the spirit of the present invention, all other embodiments not recorded in the present invention obtained by ordinary technicians in this field according to the embodiments recorded in the present invention without making creative work should belong to the protection scope of the present invention.

[0029] Figure 1 FIG. 1 is a flow chart of the steps of an adaptive power supply service grid intelligent generation method of the present invention. Figure 1 As shown, the first aspect of the present invention relates to an adaptive power supply service grid intelligent generation method, the method comprising steps 1 to 3.

[0030] Step 1: Taking substations and power users as samples, a hierarchical clustering method is used to cluster multiple samples in the power grid based on geographic location information to obtain multiple clusters.

[0031] "One Map for the Grid" is committed to comprehensively serving the digital transformation of power marketing, supporting the construction of new power systems, and creating excellent power supply service quality. On the basis of the existing construction results of one map for marketing services, it adheres to the principles of "demand-oriented, capability sharing, practical and efficient", and follows the design concept of "static resource underlying aggregation, dynamic business real-time integration", promoting the deep integration and real-time sharing of all elements of power supply services such as power grid equipment, service resources, customer information, work order events, etc., to realize a visual application that integrates panoramic management of marketing and distribution resources, business monitoring, and statistical analysis overview.

[0032] The "One Grid Map" is implemented based on the main power supply structure. The main power supply structure consists of "station-line-transformer-household". Among them, "station" refers to the power supply station, which is the source of electricity; "line" refers to the transmission line, which is the transmission channel from the power supply station to the destination; "transformer" refers to the substation, which is responsible for transforming high-voltage electricity from the transmission line into low-voltage electricity that users can directly use; "household" is the user, which is the final end of electricity.

[0033] The center of the power supply service grid is the substation. A substation connects multiple users, and these users are uniformly managed and monitored by the substation. Once a power outage or power restoration occurs, monitoring is also based on the substation. Therefore, the core information for building a power supply service grid is the "substation-household" topology. The present invention is also based on the "substation-household" topology. First, an intelligent algorithm is used to discover and process abnormal data in the topology, and then image recognition technology is used to achieve accurate division of the power supply service grid.

[0034] Preferably, substations and power users are taken as samples, and a hierarchical clustering method is used to cluster multiple samples within the power grid based on geographic location information, including: taking the topological geographic coordinates of substations and power users as sample information, and implementing clustering based on the distance between the topological geographic coordinates of the samples; clustering takes each sample as an initial cluster, gradually reduces the number of initial clusters by merging similar samples, and updates the clusters until the clustering end conditions are met.

[0035] The present invention supports combining the topological information and address information between electric power substations and users to form substation-user topological information with geographic coordinates. The above information at least includes the connection relationship and status between substations, lines, various line switches, and users, and the geographic coordinates of each substation and each user.

[0036] Each substation and each user is regarded as a clustering sample by geographic coordinates, and clustering is performed based on geographic coordinates. In the clustering process, each sample is regarded as an initial cluster, and the similarity between samples is calculated according to the distance between samples. In one embodiment, the distance between samples is calculated by Euclidean distance. If the Euclidean distance is smaller, the samples are more similar.

[0037] Merge the two clusters with the smallest distance to update the initial cluster. In this way, the initial cluster is continuously updated iteratively to obtain the final cluster. Starting from the second iteration, the cluster center of each cluster is used as the position of a cluster, and the cluster is re-aggregated with other unmerged samples and unmerged clusters. When the final clustering result reaches the pre-set clustering end condition, clustering is completed.

[0038] In one embodiment, the clustering termination condition may be that a sufficiently small number of clusters is reached, and the cluster number threshold is set in advance, and may be reasonably set according to the number of substations in the region, the number of geographical areas, and the like. In another embodiment, the clustering termination condition is that the sample distance within each sample cluster is small enough, or the inter-cluster distance between sample clusters is large enough. In other embodiments, it is also supported to implement regression on the sample cluster division results in the clustering algorithm of the present invention through actual power distribution business conditions, judge the accuracy and error of the division results, and improve the clustering parameters with the goal of reducing the error. Using similar means in the prior art, the present invention supports tuning the process of iteratively merging sample clusters through other methods, such as a hierarchical clustering algorithm.

[0039] The present invention supports checking the results of the clustering process, such as manual verification. If in the final clustering result, the number of clustered samples in any cluster is less than 5% of the number of all samples in the clustering area, this class can be regarded as an abnormal cluster, and the samples in the abnormal cluster can be checked as suspected abnormal data, such as manual verification or through further clustering process, to split the suspected abnormal samples into other adjacent accurate clusters.

[0040] In the process of implementing pairwise merging of clusters, the method of the present invention supports constructing a distance similarity matrix using the distance between each sample or cluster and other samples or clusters, and selects the number of clusters or samples used for merging in the current iteration process and the merging method through a matrix method.

[0041] It should be noted that in addition to geographic coordinates, other information can also be used to implement the clustering process. For example, although a user is not geographically closest to a substation, based on relevant parameters such as the substation-user topological relationship and the power transmission capacity of the substation, the user is closer to a substation that is farther away from the business logic. When referring to such parameters, more data dimensions can be used as sample parameters to participate in the clustering process.

[0042] Step 2: extract the substation-power user topology of the power grid, and correspond the multiple clusters to the corresponding substations according to the topology.

[0043] After extracting multiple clusters, the association between each cluster and the substation is preliminarily determined based on the substation in the cluster or the location of the cluster. A substation may correspond to one or more clusters.

[0044] Preferably, cluster boundary lines of clusters and substations are drawn based on the boundary points of clusters, including: constructing a point set with the geographic location information of all samples in the clusters, analyzing the boundary points in the clusters based on the convex hull algorithm, and obtaining the cluster boundary lines of the clusters; if the overlapping area between adjacent clusters exceeds a preset range, it is determined that the adjacent clusters correspond to a substation, and a cluster boundary line for each cluster is generated; the cluster boundary lines of multiple clusters corresponding to a substation are extracted to generate the cluster boundary line of the current substation through the cluster boundary lines of multiple clusters.

[0045] In one embodiment, since the clusters do not actually correspond to a substation, and the dimension parameters of the cluster samples are also angled, there may be grid overlaps between multiple clusters. In this case, the boundaries of each cluster are first analyzed, and then the substation corresponding to the cluster is determined by the location, overlapping relationship, etc. of each cluster.

[0046] When analyzing the boundaries of clusters, the convex hull algorithm is used to obtain the boundary line. First, traverse the edges formed by any two points in the cluster, and for each edge, check whether the remaining points fall on the same side of the edge. If so, the current edge is determined to be the edge of the convex hull. Otherwise, continue to search for the edge of the convex hull until the search is completed, and then connect all the edges end to end to obtain the boundary of the cluster.

[0047] For multiple clusters, if the overlap between two clusters is large, it is close to judging that the two clusters correspond to the same substation. In addition, it is also possible to judge whether a cluster belongs to a substation based on the relationship between the center position of the cluster and the sample position of the substation, topological connection information, etc. Combining the above methods, one or more clusters corresponding to the substation are obtained.

[0048] For a substation, the boundaries of the corresponding multiple clusters can also be further analyzed using the convex hull algorithm to obtain the boundary of the substation, thereby obtaining the cluster boundary line of the substation and the cluster boundary line of the cluster.

[0049] In the subsequent steps, the cluster boundary of a single cluster cluster can be used as a subgrid and displayed in the power grid map in the form of a dotted line, and the cluster boundary of the substation can be used as a power supply service grid and displayed in the power grid map in the form of a solid line. Since the dotted line and the solid line represent the power supply process in different ways and with different business logics, they can also be used as a basis for implementing reasonable business allocation, scheduling and balancing at different levels in other marketing businesses of "One Power Grid Map".

[0050] Step 3: Draw cluster boundary lines of clusters and substations based on the boundary points of clusters, use the Canny algorithm to extract candidate boundary lines in the baseline road network, select boundary lines from the candidate boundary lines based on the similarity between cluster boundary lines and candidate boundary lines, and fit the power supply service grid based on the selected boundary lines.

[0051] After obtaining the clustering clusters and clustering boundary lines of substations, these boundary lines can only characterize the power supply range and status of existing users, and cannot fully characterize the power supply range of substations for users to be expanded in the future. To this end, the present invention further clarifies the influence of the actual geographical division on the power supply range of substations by fitting the reference road network, and ensures the authenticity, effectiveness and rationality of the power supply service grid by fitting the power supply service grid into the power grid map.

[0052] In order to fit the power grid map, the present invention first collects a reference road network image, and uses an edge recognition algorithm to identify roads, boundaries, and landmark boundary lines on the map in the reference road network image. In the boundary recognition process, the present invention uses the Canny algorithm to extract image edges.

[0053] Preferably, the Canny algorithm is used to extract candidate boundary lines in the baseline road network, and the candidate boundary lines are selected based on the similarity between the cluster boundary lines and the candidate boundary lines, including: obtaining the extraction results of the Canny algorithm, and using the intersection points and the maximum curvature points in the extraction results to segment the extraction results, and divide the extraction results into multiple candidate line segments; implementing arbitrary permutations and combinations for the multiple candidate line segments, and connecting them into multiple non-overlapping or incompletely overlapping candidate paths, and numbering the multiple candidate paths; setting a starting point for any cluster boundary line, and traversing each candidate path corresponding to the number from the starting point, and analyzing the similarity between the cluster boundary line and the candidate path; using the candidate path with the greatest similarity as the candidate boundary segment of the current cluster boundary line, until the selection of candidate boundary lines for all substations and all cluster clusters is completed.

[0054] First, perform Gaussian filtering on the image to smooth the image and reduce the impact of noise on edge detection. The formula for Gaussian filtering is:

[0055]

[0056] Where G(·) is a Gaussian function, x and y are the horizontal and vertical coordinates of the image pixels, respectively, and σ is the standard deviation constant, which determines the size and smoothness of the filter.

[0057] The filtering operation is done through convolution, the formula is as follows:

[0058] I smoo (x,y)=I*G(x,y)

[0059] In the formula, I smoo (·) is the image smoothed by Gaussian filtering.

[0060] Secondly, the Sobel operator is used to calculate the gradient strength and direction of each pixel in the image. The Sobel operator determines the gradient strength and direction of the pixel by calculating the gradient value of the pixel in the image in the x and y directions. The Sobel operator is divided into horizontal gradient and vertical gradient, and the gradient is:

[0061]

[0062] Thus, we can get

[0063] I x (x,y)=I smoot *G x (x,y)I y (x,y)=I smooth *G y (x,y)

[0064] In the formula, I x with I y They are the horizontal and vertical coordinates after Sobel transformation.

[0065] In summary, the amplitude and direction of the image are:

[0066]

[0067] Where M(x,y) and θ(x,y) are the gradient magnitude and direction of pixel x and y on the image after Sobel transform.

[0068] Non-maximum suppression is to perform non-maximum suppression on the gradient magnitude in the gradient direction, retaining the pixels with local maxima and suppressing the non-edge pixels. The specific operation is that for each pixel point, according to its gradient direction, the gradient magnitude of this point is compared with the gradient magnitudes of two adjacent points along the gradient direction. If the gradient magnitude of this point is not the maximum among the three, its gradient magnitude is set to 0. Hysteresis thresholding starts from strong edges and recursively labels the weak edges connected to the strong edges as edges; the weak edges not connected to the strong edges are suppressed. Finally, the edges detected by the Canny edge detection algorithm are obtained as candidate boundary lines.

[0069] In order to implement precise fitting within the power supply grid range of the substation, the present invention can obtain multiple line segments by segmenting all the extracted edges. The present invention also supports identifying and numbering all the edges in the edge map by other means. In one embodiment, extraction is performed through intersection points, points with the largest curvature change, or points with a curvature change exceeding a preset value, and all the edge curves are divided by these nodes, thereby obtaining multiple lines with endpoints at both ends. Any permutation and combination of the above lines are performed, and multiple closed candidate lines are obtained by connecting the endpoints end to end. Numbering each candidate line enables the sequential extraction of any candidate line during the calculation process to perform similarity calculation.

[0070] Preferably, a starting point is set for any one clustering boundary line, and starting from the starting point, each candidate path corresponding to each number is traversed, and the similarity between the clustering boundary line and the candidate path is analyzed, including: extracting the sampling point loop of the clustering boundary line, and obtaining the sampling point sequence [(x0,y0),(x1,y1),……,(x n ,y n )] of the clustering boundary line through the set starting point (x0,y0), where n + 1 is the number of sampling points; extracting the sampling points of the candidate path, and arbitrarily selecting two candidate path sampling points (u a ,v a ) and (u b ,v b ), where a < b and the distance between a and b is fixed, calculating the distance D1 between the starting point (x0,y0) and the sampling point (u a ,v a ), the distance D2 between the starting point (x0,y0) and the sampling point (u b ,v b ), the distance D3 between the current traversed point (x i ,y i ) on the clustering boundary line and the sampling point (u a ,v a ), where i is between 1 and n, and the current traversed point (x i ,y i ) and the sampling point (ub ,v b ) between the distance D4; if min(D1,D2,D3)=D1, then max(D4,D1) is the current traversal point (x i ,y i ) and the distance between the candidate path, and the next traversal point (x i+1 ,y i+1 ) and candidate path sampling points (u a+1 ,v a+1 ) and (u b+1 ,v b+1 ) and the next starting point (x1, y1); if min(D1, D2, D3) = D2, then max(D4, D2) is used as the current traversal point (x i ,y i ) and the candidate path, and perform distance calculation on the next sampling point i+1 in the sampling point sequence of the cluster boundary line; if min(D1,D2,D3)=D3, then max(D4,D3) is used as the current traversal point x i The distance between the candidate path and the next sampling point in the candidate path sampling point sequence is recalculated (u b+1 ,v b+1 ) and the current traversal point (x i ,y i ) to update max(D4,D3); gradually select sampling points along the cluster boundary line and the candidate path to perform calculations until the calculation of the distances between all sampling points in the cluster boundary line and the candidate path is completed, and the sum of the distances between all sampling points in the cluster boundary line and the candidate path is used as the similarity value between the cluster boundary line and the candidate path.

[0071] In another embodiment, the distance is calculated by the following formula, and the calculated similarity value between the cluster boundary line P and the candidate path Q is the distance between the two curves. The distance calculation formula is:

[0072]

[0073] Where i and j are the sampling point numbers on the cluster boundary line P and the candidate path Q, respectively. (x, y) are the horizontal and vertical coordinates of the sampling point on the cluster boundary line, and (u, v) are the horizontal and vertical coordinates of the sampling point on the candidate path.

[0074] Finally, all processed trajectory data and path analysis results are combined to generate a map fitting result, which should accurately reflect the relationship between the polygon boundaries and the nearest roads.

[0075] The present invention proposes for the first time to combine the topological information of the transformer and the geographical location clustering algorithm, and accurately detect the abnormal values ​​of the electricity address data through the clustering algorithm combined with knowledge verification. In addition, the present invention combines image recognition intelligent algorithms such as boundary recognition and road recognition to improve the map understanding ability and accurately locate the grid boundary lines in complex geographical data. The adaptive boundary line fitting algorithm can accurately define the scope of the power supply service grid, overcome the problem of grid division overlap, and also make the grid boundary reasonable and the regional division clear.

[0076] Figure 2 The figure is a schematic diagram of the module structure of an adaptive power supply service grid intelligent generation system of the present invention. Figure 2 As shown, the second aspect of the present invention relates to an adaptive power supply service grid intelligent generation system using the method in the first aspect of the present invention, the system includes a clustering module, an extraction module, and a fitting module; wherein the clustering module is used to take substations and power users as samples, and adopt a hierarchical clustering method to cluster multiple samples in the power grid based on geographic location information to obtain multiple cluster clusters; the extraction module is used to extract the substation-power user topology of the power grid, and correspond the multiple cluster clusters to corresponding substations according to the topology; the fitting module is used to draw cluster boundary lines of cluster clusters and substations based on the boundary points of the cluster clusters, adopt the Canny algorithm to extract candidate boundary lines in the benchmark road network, and select boundary lines from the candidate boundary lines based on the similarity between the cluster boundary lines and the candidate boundary lines, and fit the power supply service grid based on the selected boundary lines.

[0077] A third aspect of the present invention relates to a terminal, comprising a processor and a storage medium; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the steps of the method in the first aspect of the present invention.

[0078] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. An adaptive power supply service grid intelligent generation method, characterized in that: The method comprises the following steps: Taking substations and power users as samples, a hierarchical clustering method is used to cluster multiple samples in the power grid based on geographical location information to obtain multiple clusters. Extracting a substation-power user topology of a power grid, and corresponding the plurality of clusters to corresponding substations according to the topology; Based on the boundary points of the cluster clusters, cluster boundary lines of cluster clusters and substations are drawn, and the Canny algorithm is used to extract candidate boundary lines in the benchmark road network. Based on the similarity between the cluster boundary lines and the candidate boundary lines, boundary lines are selected from the candidate boundary lines, and the power supply service grid is fitted based on the selected boundary lines.

2. The method for intelligently generating an adaptive power supply service grid according to claim 1, characterized in that: The substation and the power user are taken as samples, and a hierarchical clustering method is used to cluster multiple samples in the power grid based on geographical location information, including: The topological geographic coordinates of substations and power users are used as sample information, and clustering is performed based on the distance between the topological geographic coordinates of the samples; The clustering takes each sample as an initial cluster, gradually reduces the number of initial clusters by merging similar samples, and updates the clusters until the clustering end condition is met.

3. The method for intelligently generating an adaptive power supply service grid according to claim 2, characterized in that: The extracting the substation-power user topology of the power grid and corresponding the plurality of clusters to corresponding substations according to the topology includes: Corresponding each sample in the current cluster to the substation-power user topology; The substation that falls within all power users or the most power users is extracted, and the substation is used as the substation corresponding to the current cluster.

4. The method for intelligently generating an adaptive power supply service grid according to claim 3, characterized in that: The step of drawing the cluster boundary lines of the clusters and the substations based on the cluster boundary points includes: Constructing a point set with the geographical location information of all samples in the cluster, analyzing the boundary points in the cluster based on the convex hull algorithm, and obtaining the cluster boundary line of the cluster; If the overlapping area between adjacent clusters exceeds a preset range, it is determined that the adjacent clusters correspond to one substation, and a cluster boundary line of each cluster is generated; Cluster boundary lines of a plurality of cluster clusters corresponding to one substation are extracted to generate a cluster boundary line of a current substation through the cluster boundary lines of the plurality of cluster clusters.

5. The method for intelligently generating an adaptive power supply service grid according to claim 4, characterized in that: The method of extracting candidate boundary lines from the reference road network by using the Canny algorithm and selecting the candidate boundary lines based on the similarity between the cluster boundary lines and the candidate boundary lines includes: Obtaining the extraction result of the Canny algorithm, and using the intersection points and the points with the maximum curvature change in the extraction result to segment the extraction result, and dividing the extraction result into a plurality of candidate line segments; Performing arbitrary permutations and combinations on a plurality of candidate line segments, and connecting them into a plurality of non-overlapping or incompletely overlapping candidate paths, and numbering the plurality of candidate paths; Setting a starting point for any cluster boundary line, and traversing each candidate path corresponding to a number from the starting point, and analyzing the similarity between the cluster boundary line and the candidate path; The candidate path with the greatest similarity is used as a candidate boundary line segment of the current cluster boundary line until the candidate boundary lines of all substations and all clusters are selected.

6. The method for intelligently generating an adaptive power supply service grid according to claim 5, characterized in that: The step of setting a starting point for any cluster boundary line, traversing each candidate path corresponding to a number from the starting point, and analyzing the similarity between the cluster boundary line and the candidate path includes: Extract the sampling point cycle of the cluster boundary line, and obtain the sampling point sequence of the cluster boundary line [(x0, y0), (x1, y1), ..., (x n ,y n )], n+1 is the number of sampling points; Extract the sampling points of the candidate path and select any two candidate path sampling points (u a , v a ) and (u b , v b ), a<b and the distance between a and b is fixed, calculate the starting point (x0, y0) and the sampling point (u a , v a ), the distance D1 between the starting point (x0, y0) and the sampling point (u b , v b ) between the current traversal point (x i ,y i ) and sampling point (u a , v a ), i is between 1 and n, and the current traversal point (x i ,y i ) and sampling point (u b , v b ) between the distance D4; If min(D1, D2, D3) = D1, then max(D4, D1) is the current traversal point (x i ,y i ) and the distance between the candidate path, and the next traversal point (x i+1 ,y i+1 ) and candidate path sampling points (u a+1 , v a+1 ) and (u b+1 , v b+1 ), the distance between the next starting point (x1, y1) is calculated; If min(D1, D2, D3) = D2, then max(D4, D2) is the current traversal point (x i ,y i ) and the candidate path, and performing distance calculation on the next sampling point i+1 in the sampling point sequence of the cluster boundary line; If min(D1, D2, D3) = D3, then max(D4, D3) is the current traversal point x. i The distance between the candidate path and the candidate path is calculated, and the next sampling point in the sampling point sequence of the candidate path is recalculated (u b+1 , v b+1 ) and the current traversal point (x i ,y i ) to update max(D4, D3); Sampling points are gradually selected along the cluster boundary line and the candidate path to perform calculations until the calculation of the distances between all sampling points in the cluster boundary line and the candidate path is completed, and the sum of the distances between all sampling points in the cluster boundary line and the candidate path is used as the similarity value between the cluster boundary line and the candidate path.

7. The method for intelligently generating an adaptive power supply service grid according to claim 6, characterized in that: The cluster boundary line with the smallest similarity value to the current cluster boundary line is selected as the candidate path with the largest similarity.

8. The method for intelligently generating an adaptive power supply service grid according to claim 7, characterized in that: The step of fitting a power supply service grid based on the selected boundary line includes: The power supply service grid is fitted with the candidate paths, and the power supply service grid is loaded into the reference road network to generate a power supply service map.

9. An adaptive power supply service grid intelligent generation system using the method according to any one of claims 1 to 8, characterized in that: The system includes a clustering module, an extraction module, and a fitting module; wherein, The clustering module is used to take substations and power users as samples, and cluster multiple samples in the power grid based on geographical location information using a hierarchical clustering method to obtain multiple clusters; The extraction module is used to extract the substation-power user topology of the power grid, and respectively correspond the multiple clusters to corresponding substations according to the topology; The fitting module is used to draw cluster boundary lines of cluster clusters and substations based on the boundary points of cluster clusters, use the Canny algorithm to extract candidate boundary lines in the benchmark road network, and select boundary lines from the candidate boundary lines based on the similarity between the cluster boundary lines and the candidate boundary lines, and fit the power supply service grid based on the selected boundary lines.

10. A terminal comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the steps of the method according to any one of claims 1-8.

Images