Intelligent substation site selection method and system fusing multi-source geospatial data

Abstract

The invention relates to the technical field of power grid planning and geographic information systems, in particular to an intelligent substation site selection method and system fusing multi-source geographic space data. S1, data preprocessing and effective search area generation; s2, constructing a substation construction cost function; and S3, intelligent addressing based on a particle swarm optimization algorithm. According to the method, a complex site selection decision process is converted into a clearly defined mathematical optimization problem, automatic optimization is carried out through an algorithm, the subjectivity of manual intervention is eliminated, and the efficiency and scientificity of site selection work are remarkably improved. The particle swarm optimization algorithm is adopted, efficient search can be carried out in a wide and complex solution space, and good global optimization capacity is achieved. Meanwhile, through the technical means of geometric object caching, parallel computing and the like, the computing speed during large-scale data processing is effectively increased, and the timeliness requirement of actual engineering application is met.

Description

Technical Field

[0001] This invention relates to the field of power grid planning and geographic information system (GIS) technology, specifically to a method and system for intelligent substation site selection that integrates multi-source geospatial data. Background Technology

[0002] Substations are critical hubs in power systems, and their site selection directly impacts grid construction investment, operational reliability, and power supply economy. Traditional substation site selection relies heavily on the experience of planners, making decisions through manual on-site surveys and drawing comparisons. This approach has the following drawbacks: 1. High subjectivity and low efficiency: The site selection process relies heavily on personal experience and lacks unified, quantifiable evaluation standards, which may result in suboptimal outcomes. Furthermore, manual comparison of solutions is time-consuming and labor-intensive, making it difficult to meet the needs of large-scale power grid planning.

[0003] 2. Limited data processing capabilities: Site selection requires comprehensive consideration of various factors such as topography, geology, transportation, urban planning, and environmental protection. The massive amounts of geospatial data (such as elevation data, remote sensing imagery, and ground feature vector data) are difficult to analyze and utilize comprehensively and in detail manually.

[0004] 3. Disconnection between "substation" and "line" planning: Traditional site selection often determines the substation site first and then plans the line route, failing to coordinate and optimize the construction costs of the substation and the transmission line (i.e., the "incoming and outgoing line corridor") as a whole. This "substation-line separation" model may result in poor overall economic efficiency even if the substation site itself meets the requirements, due to excessively high line construction costs or difficult routes. Summary of the Invention

[0005] The present invention aims to provide a method and system for automatically, accurately and efficiently determining the optimal site of a substation, overcoming the defects of traditional site selection methods such as subjectivity, inefficiency and lack of coordination between station and line planning.

[0006] The technical solution adopted by this invention to solve its technical problem is a substation intelligent site selection method that integrates multi-source geospatial data, including the following steps: S1: Based on the input macroscopic search area, raster elevation data, and feature vector data, the raster elevation data is processed using a connectivity analysis algorithm to generate an elevation restricted area, and the feature vector data is buffered to generate a feature restricted area; the elevation restricted area and the feature restricted area are spatially joined, and then spatially intersected with the macroscopic search area to generate a comprehensive prohibited area; the comprehensive prohibited area is removed from the macroscopic search area to obtain the effective search area; S2: Define multiple types of grid connection points and the corresponding unit distance line construction cost for each type; construct a total investment cost function with the candidate substation location as the input variable, wherein the total investment cost function is the Euclidean distance from the candidate substation location to all the grid connection points, multiplied by the corresponding unit distance line construction cost and summed. S3: Initialize the particle population of the particle swarm optimization algorithm within the effective search area, where the position of each particle represents a candidate substation site; during the algorithm iteration process, calculate the total investment cost function value corresponding to the position of each particle to evaluate its merits, and simultaneously verify the validity of the new particle positions generated after each iteration to ensure that they are within the effective search area; continuously search for the global optimal solution that minimizes the total investment cost through iterative updates, and output the particle position corresponding to the final global optimal solution as the recommended optimal substation site.

[0007] Furthermore, in step S1, the elevation restricted area processing includes inputting raster elevation data, setting the allowed elevation range, identifying raster points that exceed the range, clustering the raster points that exceed the range into connected regions through a connectivity analysis algorithm, and converting them into vector polygons as elevation restricted areas. The processing of restricted areas includes buffering point and linear features to generate area-based restricted areas, performing a union operation on the restricted areas and elevation restricted areas, and then performing an intersection operation with the macro search area to generate a comprehensive prohibited area.

[0008] Furthermore, the connectivity analysis algorithm employs a disjoint-set data structure to cluster spatially continuous out-of-bounds grid points into connected regions.

[0009] Furthermore, in step S2, the type of power grid connection point includes at least one of power line T-junction, existing substation, and urban load center, and the unit distance line construction cost is set differently according to the type of connection point.

[0010] Furthermore, in step S3, the addressing based on the particle swarm optimization algorithm specifically includes: S31: Set the number of particles, maximum number of iterations, inertia weight w, cognitive coefficient c1, and social coefficient c2; S32: Particle initialization, randomly generate the initial position of the particles within the effective search area, calculate the initial total investment cost, and randomly generate the initial velocity vector of the particles; S33: Iterative optimization. For each particle, update the velocity vector based on the current velocity vector, the individual optimal position, and the global optimal position. Update the particle position based on the updated velocity. Check if the new position is within the valid search area. If valid, calculate the total investment cost and update the individual optimal position and the global optimal position. The formula for updating the velocity vector is as follows: vnew =w×v old +c1×r1×(pbest-pos)+c2×r2×(gbest-pos); where pbest is the individual's optimal position, gbest is the global optimal position, w is the inertia weight, c1 is the cognitive coefficient, c2 is the social coefficient, and v old Here, r1 is the current velocity vector, pos is the current position of the particle, and r2 are random numbers between [0,1]. The formula for updating the particle position is as follows: pos new =pos+v new ; where v new This is the updated velocity vector, where pos is the particle's current position; S34: Termination and Output. When the maximum number of iterations is reached or the global optimal position is stable, output the global optimal position as the optimal substation site.

[0011] Furthermore, during the iterative optimization process, if a particle's new position falls into a prohibited area or flies out of the macroscopic search area, it is determined to be an invalid position, and its total investment cost is set to infinity.

[0012] A substation intelligent site selection system integrating multi-source geospatial data includes: The data preprocessing module is used to perform multi-source data preprocessing and constraint modeling, and to generate comprehensive prohibited areas and effective search areas based on macro search areas and multi-source geospatial data; The cost model building module is used to define the connection point type and the construction cost per unit distance of the line, and to build a cost calculation function with the candidate substation location as a variable; The intelligent addressing module is used to initialize the particle swarm within the effective search area based on the particle swarm optimization algorithm, and find the optimal substation site through iterative optimization.

[0013] Furthermore, the data preprocessing module is specifically used for: Process raster elevation data and generate elevation restricted areas through connectivity analysis; Process ground feature vector data and generate ground feature restricted areas through buffering operations; Spatial calculations are performed on restricted areas for ground features and elevation restrictions to generate comprehensive prohibited areas and effective search areas.

[0014] Furthermore, the cost model construction module is specifically used to calculate the sum of the Euclidean distance from the candidate location to each connection point multiplied by the corresponding unit distance line construction cost.

[0015] The beneficial effects of this invention are: 1. This method transforms the complex site selection decision-making process into a well-defined mathematical optimization problem. By automatically finding the optimal solution through algorithms, the subjectivity of human intervention is eliminated, which significantly improves the efficiency and scientific nature of site selection.

[0016] 2. By performing raster analysis and connected component extraction on elevation data, as well as buffering and spatial calculations on ground features, it is possible to accurately model irregular and diverse geographical restricted areas, ensuring the feasibility and safety of site selection results in physical space.

[0017] 3. The cost model of this invention tightly couples site selection with the construction cost of multiple access lines, and performs global optimization with the goal of minimizing total investment. This avoids suboptimal decisions that may be caused by the traditional "site first, line later" model, and effectively reduces the overall cost of power grid projects.

[0018] 4. Employing the particle swarm optimization algorithm, it can efficiently search within a vast and complex solution space, demonstrating excellent global optimization capabilities. Furthermore, through techniques such as geometric object caching and parallel computing, it effectively improves the computational speed when processing large-scale data, meeting the timeliness requirements of practical engineering applications. Attached Figure Description

[0019] Figure 1 This is a schematic diagram illustrating the generation of the effective search region in this invention; Figure 2 This is a schematic diagram illustrating the principle of constructing the substation construction cost function according to the present invention; Figure 3 This is a schematic diagram of the iterative optimization process of the particle swarm optimization algorithm of this invention; Detailed Implementation

[0020] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0021] like Figures 1-3 As shown, the present invention provides a method for intelligent substation site selection that integrates multi-source geospatial data, comprising the following steps: S1: Based on the input macroscopic search area, raster elevation data, and feature vector data, the raster elevation data is processed using a connectivity analysis algorithm to generate an elevation restricted area, and the feature vector data is buffered to generate a feature restricted area; the elevation restricted area and the feature restricted area are spatially joined, and then spatially intersected with the macroscopic search area to generate a comprehensive prohibited area; the comprehensive prohibited area is removed from the macroscopic search area to obtain the effective search area; S2: Define multiple types of grid connection points and the corresponding unit distance line construction cost for each type; construct a total investment cost function with the candidate substation location as the input variable, wherein the total investment cost function is the Euclidean distance from the candidate substation location to all the grid connection points, multiplied by the corresponding unit distance line construction cost and summed. S3: Initialize the particle population of the particle swarm optimization algorithm within the effective search area, where the position of each particle represents a candidate substation site; during the algorithm iteration process, calculate the total investment cost function value corresponding to the position of each particle to evaluate its merits, and simultaneously verify the validity of the new particle positions generated after each iteration to ensure that they are within the effective search area; continuously search for the global optimal solution that minimizes the total investment cost through iterative updates, and output the particle position corresponding to the final global optimal solution as the recommended optimal substation site.

[0022] Furthermore, in step S1, the elevation restricted area processing includes inputting raster elevation data, setting the allowed elevation range, identifying raster points that exceed the range, clustering the raster points that exceed the range into connected regions through a connectivity analysis algorithm, and converting them into vector polygons as elevation restricted areas. The processing of restricted areas includes buffering point and linear features to generate area-based restricted areas, performing a union operation on the restricted areas and elevation restricted areas, and then performing an intersection operation with the macro search area to generate a comprehensive prohibited area.

[0023] Furthermore, the connectivity analysis algorithm employs a disjoint-set data structure to cluster spatially continuous out-of-bounds grid points into connected regions.

[0024] Furthermore, in step S2, the type of power grid connection point includes at least one of power line T-junction, existing substation, and urban load center, and the unit distance line construction cost is set differently according to the type of connection point.

[0025] Further, see Figure 3 In step S3, the addressing based on the particle swarm optimization algorithm specifically includes: S31: Set the number of particles, maximum number of iterations, inertia weight w, cognitive coefficient c1, and social coefficient c2; It's important to note that, for example, with 150 particles and a maximum of 250 iterations, these 100 points will "move" on the map and the cost will be recalculated 250 times. This is a termination condition to prevent the program from calculating indefinitely, balancing computation time and location accuracy.

[0026] Inertia weight w: The degree to which the current search trend is maintained. The formula for w is: the cost of the current particle after the Kth iteration / the cost of all particles after the Kth iteration. If w is large, it means that the particle tends to continue rushing in the original direction; if w is small, it means that the particle tends to make minor adjustments near the original position.

[0027] Cognitive coefficient (c1): The tendency to move closer to its historical best position. It represents how much a candidate point is willing to "look back". If the point finds its current position too expensive, it will tend to move towards the cheaper position (pbest) that it remembers it has visited before.

[0028] Social coefficient (c2): The tendency to move towards the optimal position in the group. It represents how much a candidate point is willing to "follow the crowd". If another point in the group finds an excellent position (gbest), this parameter determines whether the current point should quickly move towards that good position.

[0029] S32: Particle initialization, randomly generate the initial position of the particles within the effective search area, calculate the initial total investment cost, and randomly generate the initial velocity vector of the particles; S33: Iterative optimization. For each particle, update the velocity vector based on the current velocity vector, the individual optimal position, and the global optimal position. Update the particle position based on the updated velocity. Check if the new position is within the valid search area. If valid, calculate the total investment cost and update the individual optimal position and the global optimal position. The formula for updating the velocity vector is as follows: v new =w×v old +c1×r1×(pbest-pos)+c2×r2×(gbest-pos); where pbest is the individual's optimal position, gbest is the global optimal position, w is the inertia weight, c1 is the cognitive coefficient, c2 is the social coefficient, and v old The current velocity vector is w×v, where pos is the particle's current position, and r1 and r2 are random numbers between [0,1]. old The vector is the inertia vector; c1×r1×(pbest-pos) is the cognitive vector, and c2×r2×(gbest-pos) is the social vector; the random numbers (r1,r2) represent random perturbations during the search process. This simulates uncertainty in the real-world environment, preventing all points from rigidly moving in straight lines and increasing the possibility of escaping local traps.

[0030] The formula for updating the particle position is as follows: pos new =pos+v new ; where v new The updated velocity vector is the new velocity vector, and pos is the particle's current position. S34: Termination and Output. When the maximum number of iterations is reached or the global optimal position is stable, output the global optimal position as the optimal substation site.

[0031] Furthermore, during the iterative optimization process, if a particle's new position falls into a prohibited area or flies out of the macroscopic search area, it is determined to be an invalid position, and its total investment cost is set to infinity.

[0032] Example 1

[0033] 1. Data preparation and parameter settings

[0034] The macroscopic search area is specifically a vector polygon file covering an area of ​​approximately 100 square kilometers.

[0035] The raster elevation data is specifically a 30-meter resolution ASTERGDEM dataset, covering the entire macroscopic search area.

[0036] The vector data of ground features includes the following: linear features, primary highways; point features, and settlements.

[0037] Connection points and costs: Connection point 1 (T-connector, type A): coordinates (X1, Y1), unit cost 800,000 yuan / km.

[0038] Connection point 2 (T-connector, type A): coordinates (X2, Y2), unit cost 800,000 yuan / km.

[0039] Connection point 3 (existing 220kV substation, type B): coordinates (X3, Y3), unit cost 1 million yuan / km.

[0040] Connection point 4 (urban load center, type C): coordinates (X4, Y4), unit cost 1.2 million yuan / km.

[0041] Particle Swarm Optimization Algorithm Parameters: Number of particles = 150. Maximum number of iterations = 250.

[0042] 2. Execution process

[0043] S1: Data Preprocessing and Effective Search Region Generation

[0044] (1) Load all input vector data and macroscopic search area boundaries into Shapely geometry objects and store them in the cache.

[0045] (2) Handling elevation restricted areas: The allowed altitude range is set to 50 to 200 meters. The system traverses the DEM raster, marking all raster points with elevation values ​​<50 meters or >200 meters as exceeding the limit. A disjoint-set data structure algorithm is used to perform four-neighborhood connectivity analysis on these exceeding points, generating a single large connected region (a high hill). The outer envelope of this region is extracted to generate a vector polygon, which serves as the elevation restricted area.

[0046] (2) Treatment and integration of restricted areas

[0047] For "Class I highways," a 50-meter buffer analysis is performed based on the linear data of the road centerline, generating a 100-meter-wide areal restricted area. For "residential areas," a 300-meter buffer analysis is performed on the buildings. All the aforementioned restricted areas are spatially joined with the elevation restricted areas to obtain an initial integrated region. This initial integrated region is then spatially intersected with the macro-search region, and invalid portions outside the boundaries are removed to obtain the final integrated prohibited region. The integrated prohibited region is then removed from the macro-search region to obtain the effective search region. S2: Constructing the substation construction cost function The cost function takes a candidate substation location P(x,y) as input. Its calculation logic is as follows: Total cost = Σ(distance from candidate position P to the i-th connection point × unit cost of the i-th connection point) For example, total cost = (d1×80) + (d2×80) + (d3×100) + (d4×120).

[0048] Wherein, the distance d is calculated using the two-dimensional Euclidean distance formula. Calculate the total cost value, which is the fitness value of the candidate point. The smaller the value, the better the location.

[0049] S3: Intelligent Addressing Based on Particle Swarm Optimization Algorithm

[0050] Within the effective search area, 150 effective particle positions are randomly generated, the initial cost of each particle is calculated, and an initial velocity vector is randomly set for each particle.

[0051] Iterative optimization (example of the k-th iteration): For particle i, a linearly decreasing weight strategy is adopted. In the current k-th iteration, the calculated inertia weight w = 0.05. The cognitive coefficient c1 = 1.8 and the social coefficient c2 = 1.8 are set. Its current position is (123456.7, 456789.0), the current velocity vector is (25.0, -15.0), the pbest individual optimal position is (123450.0, 456800.0), and the current gbest global optimal position is (123500.0, 456500.0).

[0052] Speed ​​update: Generate two random numbers r1=0.3, r2=0.7. Calculate the new speed according to the formula: v new =0.05×(25.0,-15.0)+1.8×0.3×((123450.0,456800.0)-(123456.7,456789.0))+1.8×0.7×((123500.0,456500.0)-(123456.7,456789.0)) The calculated result is v new =(52.19,-358.95).

[0053] New location pos new =(123456.7,456789.0)+(52.19,-358.95)=(123508.89,456430.05).

[0054] Cost calculation and update: The cost of the new position is calculated to be 78.5 million yuan. Compared with the cost of particle pbest (79 million yuan), the new cost is lower, so pbest is updated to the new position. Then, compared with the cost of gbest (78.2 million yuan), it is not better, so gbest is not updated.

[0055] Invalid position handling: If another particle j's updated position falls into the synthesis forbidden region, it is deemed invalid. The system sets its cost to infinity, resets its position to its previous generation's pbest, and randomly perturbs a new position around pbest. If it is still invalid, the system continues to reset until a valid position is found.

[0056] When the iteration reached the 215th iteration, the global optimal solution gbest remained unchanged for 30 consecutive iterations, and the algorithm terminated prematurely.

[0057] 3. Results Output

[0058] The system outputs the final result: Recommended optimal site coordinates (123508.89, 456430.05). The estimated minimum total investment cost is 78.15 million yuan, with a total of 215 iterations.

[0059] A substation intelligent site selection system integrating multi-source geospatial data includes: The data preprocessing module is used to perform multi-source data preprocessing and constraint modeling, and to generate comprehensive prohibited areas and effective search areas based on macro search areas and multi-source geospatial data; The cost model building module is used to define the connection point type and the construction cost per unit distance of the line, and to build a cost calculation function with the candidate substation location as a variable; The intelligent addressing module is used to initialize the particle swarm within the effective search area based on the particle swarm optimization algorithm, and find the optimal substation site through iterative optimization.

[0060] Furthermore, the data preprocessing module is specifically used for: Process raster elevation data and generate elevation restricted areas through connectivity analysis; Process ground feature vector data and generate ground feature restricted areas through buffering operations; Spatial calculations are performed on restricted areas for ground features and elevation restrictions to generate comprehensive prohibited areas and effective search areas.

[0061] Furthermore, the cost model construction module is specifically used to calculate the sum of the Euclidean distance from the candidate location to each connection point multiplied by the corresponding unit distance line construction cost.

[0062] The embodiments described herein are preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape, and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A substation intelligent siting method of fusing multi-source geospatial data, characterized in that, The method comprises the following steps: S1: based on the input macro search area, grid elevation data and surface vector data, the grid elevation data is processed by a connectivity analysis algorithm to generate an elevation forbidden area, and the surface vector data is processed by a buffer to generate a surface forbidden area; the elevation forbidden area and the surface forbidden area are subjected to a spatial union operation, and the macro search area is subjected to a spatial intersection operation to generate a comprehensive forbidden area; the comprehensive forbidden area is removed from the macro search area to obtain an effective search area; S2: defining multiple types of power grid connection points and corresponding unit distance line construction costs; constructing a total investment cost function with the candidate substation location as the input variable, which is the Euclidean distance from the candidate substation location to all power grid connection points, multiplied by the corresponding unit distance line construction cost and accumulated; S3: initializing the particle population of the particle swarm optimization algorithm in the effective search area, and the position of each particle represents a candidate substation site; In the iteration process of the algorithm, the total investment cost function value corresponding to the position of each particle is calculated to evaluate its advantages and disadvantages, and the new particle position generated after each iteration is subjected to effectiveness verification to ensure that it is located in the effective search area; the global optimal solution that minimizes the total investment cost is constantly searched through iteration update, and the particle position corresponding to the final global optimal solution is output as the recommended optimal substation site.

2. A substation intelligent siting method for fusing multi-source geospatial data as claimed in claim 1, wherein, In step S1, the elevation forbidden area processing includes inputting the grid elevation data, setting the allowed elevation range, identifying the out-of-range grid points, clustering the out-of-range grid points into connected regions through a connectivity analysis algorithm, and converting them into vector polygons as elevation forbidden areas; The surface forbidden area processing includes buffering point and line features to generate surface forbidden areas, performing a set operation on the surface forbidden areas and the elevation forbidden areas, and performing an intersection operation on the macro search area to generate a comprehensive forbidden area.

3. A substation intelligent siting method for fusing multi-source geospatial data as claimed in claim 2, wherein, The connectivity analysis algorithm uses a union-find set algorithm to cluster spatially continuous out-of-range grid points into connected regions.

4. A substation intelligent siting method for fusing multi-source geospatial data as claimed in claim 1, wherein, In step S2, the power grid connection point types include at least one of a power line T joint, an existing substation and a city load center, and the unit distance line construction cost is differentiated according to the connection point type.

5. A substation intelligent siting method for fusing multi-source geospatial data as claimed in claim 1, wherein, In step S3, the addressing based on the particle swarm optimization algorithm specifically includes: S31: setting the number of particles, the maximum number of iterations, the inertia weight w, the cognitive coefficient c1 and the social coefficient c2; S32: particle initialization, randomly generating the initial position of the particle in the effective search area, calculating the initial total investment cost, and randomly generating the initial velocity vector of the particle; S33: iterative optimization, for each particle, updating the velocity vector according to the current velocity vector, the individual optimal position and the global optimal position, updating the particle position according to the updated velocity, checking whether the new position is located in the effective search area, if yes, calculating the total investment cost, and updating the individual optimal position and the global optimal position; wherein the formula of velocity vector update is as follows: v new = w x v old + c1 x r1 x (pbest - pos) + c2 x r2 x (gbest - pos); wherein pbest is the individual optimal position, gbest is the global optimal position, w is the inertia weight, c1 is the cognitive coefficient, c2 is the social coefficient, v old is the current velocity vector, pos is the current position of the particle, and r1 and r2 are random numbers between 0 and 1; The update formula of the particle position is as follows new =pos+v new ; wherein, v new is an updated velocity vector, and pos is the current position of the particle; S34: Terminate and output, when the maximum number of iterations is reached or the global optimal position is stable, output the global optimal position as the optimal substation site.

6. A substation intelligent siting method for fusing multi-source geospatial data as claimed in claim 5, wherein, In the iterative optimization process, if the new position of the particle falls into the forbidden area or flies out of the macro search area, it is determined as an invalid position, and the total investment cost value is set to infinity.

7. A substation intelligent siting system that fuses multi-source geospatial data, characterized by, The method comprises the following steps: a data preprocessing module for performing multi-source data preprocessing and constraint modeling, generating a comprehensive forbidden area and an effective search area based on the macro search area and multi-source geospatial data; a cost model construction module for defining connection point types and unit distance line construction costs, and constructing a cost calculation function with the candidate substation position as the variable; an intelligent addressing module for initializing a particle swarm in the effective search area based on the particle swarm optimization algorithm, and finding the optimal substation site through iterative optimization.

8. A substation intelligent siting system that fuses multi-source geospatial data as claimed in claim 7, wherein, The data preprocessing module is specifically used for: processing raster elevation data to generate an elevation forbidden area through connectivity analysis; processing terrain vector data to generate a terrain forbidden area through buffer operation; performing spatial operation on the terrain forbidden area and the elevation forbidden area to generate a comprehensive forbidden area and an effective search area.

9. A substation intelligent siting system that fuses multi-source geospatial data as claimed in claim 7, wherein, The cost model construction module is specifically used for calculating the sum of the Euclidean distance from the candidate position to each connection point multiplied by the corresponding unit distance line construction cost.

Images