A Substation Digital Twin Method and Device Based on Large-Scale Point Clouds

Abstract

The present invention discloses a substation digital twin method and device based on large-scale point clouds. Key device data in the point clouds are effectively extracted through technologies such as point cloud preprocessing, clustering segmentation, and line fitting, facilitating subsequent classification and storage. Through the tile-based inner-layer storage method, point cloud data of different importance levels are stored according to different requirements, reducing the size of the point cloud data volume and facilitating subsequent rendering. Through large-scale point cloud rendering technology, in combination with the actual situation, nodes that need to load and render corresponding levels are found for loading, improving the rendering efficiency of large-scale point clouds. Through technologies such as digital twin virtual-real mapping, data simulation, and real-time synchronization, more intuitive monitoring and analysis of substation equipment and comprehensive management of the entire life cycle are realized. Through robot simulation technology, the robot inspection process is synchronously virtualized to the digital twin virtual model of the substation, making it more convenient to monitor and manage inspection tasks.

Description

Technical Field

[0001] The present invention relates to a substation digital twin method and device based on large-scale point cloud, belonging to the technical field of substations. Background Art

[0002] In the power system, inspecting, maintaining, and promptly identifying equipment problems in substations are among the primary tasks of substation operations and maintenance professionals. Defects, hidden dangers, and abnormal operating conditions in power equipment directly impact the safe operation of power equipment, the reliable power supply of the power grid, and the production and business activities of industrial, agricultural, and commercial users. These defects can even lead to serious economic losses and undermine social harmony and stability. Therefore, substation operators must prioritize the quality of equipment inspections. Promptly identifying equipment problems and improving power supply reliability and service levels have become crucial tools for enhancing the service quality of power supply companies. Therefore, substations must continuously strengthen their inspection capabilities to create greater economic and social benefits for power users. Currently, widespread substation inspections require manual labor from professionals, which is complex, laborious, and highly dangerous. Furthermore, due to physical limitations, inspectors struggle to maintain mental and physical focus for extended periods. Working in the complex substation environment, safety incidents can easily occur due to misconduct or personal errors.

[0003] The shortcomings of existing substation digital twin technology are as follows:

[0004] 1) Substation point cloud data is typically too large to fit into memory. Current point cloud rendering methods consume a lot of memory and CPU resources and fail to account for the differences between different types of objects. For example, the same sampling and rendering strategies are used for both non-critical objects (surroundings) and critical objects (power equipment), resulting in low rendering efficiency.

[0005] 2) Current substation modeling is completed by rendering point cloud data obtained on-site, but after modeling, it can only be used for basic information display and viewing functions. It is impossible to interact with inspection robots in the digital model, nor can it realize substation operation status simulation. Summary of the Invention

[0006] The purpose of the present invention is to overcome the shortcomings of the existing technology and provide a substation digital twin method and device based on large-scale point cloud. Through tile-type hierarchical storage and large-scale point cloud rendering technology, the requirements for memory and CPU are reduced and rendering efficiency is improved; through digital twin virtual-reality mapping, data simulation, real-time synchronization and other technologies, more intuitive monitoring and analysis of substation equipment and comprehensive management of the entire life cycle are achieved; through robot simulation technology, the robot inspection process is synchronized virtual and real to the digital twin virtual model of the substation, making it more convenient to monitor and manage inspection tasks.

[0007] To achieve the above object, the present invention is implemented by the following technical solutions:

[0008] In a first aspect, the present invention provides a substation digital twin method based on large-scale point clouds, including:

[0009] Obtain substation point cloud data;

[0010] Preprocess the point cloud data, downsample the point cloud data using the point cloud voxel filter algorithm, and extract the effective point cloud data;

[0011] Use the Euclidean space clustering segmentation method to divide the extracted effective point cloud data into entities;

[0012] Use the minimum spanning tree and maximum critical path algorithms to extract straight lines from the divided target entities, and fit the straight lines to obtain a straight line model;

[0013] Identify power lines from the straight line model;

[0014] For the identified power lines, adopt a tile-based hierarchical storage method, construct a storage structure based on the octree structure, and save it locally in the form of a database table according to the importance of different point clouds;

[0015] Deploy the tile data, rendering web page, and object database on the server. The Web browser realizes the rendering of large-scale point cloud data by loading the rendering module and tile data;

[0016] Build a data middle platform, and describe all information in the substation environment based on a standard protocol in a unified format and language;

[0017] Adopt the virtual-real mapping technology of substation equipment perception data to establish the correlation between perception data and data, data and equipment or environment, and realize the virtual-real mapping between perception data and equipment;

[0018] Based on comprehensive perception and data governance technology, obtain multi-dimensional and panoramic data, construct a digital twin virtual model of the substation equipment state, and realize the accurate model establishment of the equipment environment;

[0019] Construct a robot simulation environment based on ROS, combine the collected data with the current robot motion state through interactive computer technology and robot-related theories, and synchronously simulate the actual working conditions of the robot in the computer.

[0020] Further, the preprocessing of the point cloud data includes: calculating the average distance of each point to all adjacent points, assuming that the result distribution is a Gaussian distribution with a mean and a standard deviation, and regarding all points whose average distance is outside the interval defined by the global distance mean and standard deviation as outliers and pruning them from the dataset.

[0021] Further, the method of using Euclidean space clustering segmentation to partition the extracted valid point cloud data includes: judging by taking the Euclidean distance between points as a measure function, that is, when the Euclidean distance between a point and any point in a certain class set is less than a threshold, it is classified into the same class for partitioning.

[0022] Further, the method of using the minimum spanning tree and the maximum critical path algorithm to extract straight lines from the partitioned target entities includes: obtaining the minimum spanning tree for each sub-block based on the graph theory algorithm, taking the longest path in the minimum spanning tree as the approximation straight line of the wire, and using the central axis to extract the device central axis.

[0023] Further, the identification of power lines from the straight line model includes: identifying power lines from the straight line model, using the method of calculating the cross-sectional radius, at a certain point on the straight line, taking the range of the coordinates of all points within a sphere with a certain radius around it, and the union of the spreading ranges of all points can be regarded as the wire to which the straight line belongs. If the average distance from all points within the sphere to the straight line is greater than a preset threshold, it is considered that the straight line is not a power line.

[0024] Further, it also includes: digitizing the whole process of the substances and values of substation equipment, and realizing comprehensive management of the entire life cycle from the planning and design of assets, material procurement, project construction, maintenance and operation, inspection and update to retirement and scrapping.

[0025] Further, the construction of the digital twin virtual model of the substation equipment state to achieve the accurate model establishment of the equipment environment also includes: constructing a substation multi-dimensional data space by loading data resources, performing data analysis and simulation calculations on the digital twin virtual model, and real-time feedback to the digital twin virtual model to optimize and evolve the model, forming a self-optimizing intelligent operation mode.

[0026] Further, the tile-based hierarchical storage method includes:

[0027] Step 1: Read the object extraction result from the database, generate a root node, and set it as the current node;

[0028] Step 2: Read a point from the original point cloud file. If the reading is successful, go to Step 3; otherwise, if all the data has been read, go to Step 6;

[0029] Step 3: Add the read points to the current node, and determine whether the octree node being currently operated on is a leaf node. If it is, go to Step 4; otherwise, go to Step 5;

[0030] Step 4: Put the new points into the data cache of the node. When the number of points in the cache exceeds a certain threshold, the leaf node splits into a non-leaf node, and the points in the cache are gradually assigned to the child nodes, then go to Step 3;

[0031] Step 5: Calculate the average distance between this point and its neighbors. If the average distance is greater than the neighbor threshold of the category to which the node belongs, it means the point is relatively sparse, then put the new point into the data cache of the node; otherwise, assign this point to the next-level child node and go to Step 2;

[0032] Step 6: Construct a hierarchical file. The hierarchical file saves the hierarchical relationships between each node in the order of breadth-first traversal. The saved content includes how many points the node itself has and which child nodes the node has;

[0033] Step 7: During the construction of the hierarchical structure, serially store the nodes of the octree at regular intervals, and save the tile files at different levels of detail. The file names of the tile files are named according to their levels.

[0034] In a second aspect, the present invention provides a substation digital twin device based on a large-scale point cloud, including:

[0035] An acquisition unit, configured to acquire substation point cloud data;

[0036] A point cloud preprocessing unit, configured to preprocess the point cloud data, downsample the point cloud data using the point cloud voxel filter algorithm, and extract effective point cloud data;

[0037] A clustering and segmentation unit, configured to perform entity division on the extracted effective point cloud data by using the method of Euclidean space clustering and segmentation;

[0038] A line fitting unit, configured to extract lines in the divided target entities by using the minimum spanning tree and maximum critical path algorithms, and fit the lines to obtain a line model;

[0039] A power line recognition unit, configured to recognize power lines from the line model;

[0040] A storage unit, configured to, for the recognized power lines, adopt a tile-based hierarchical storage method, construct a storage structure based on an octree structure, and save them locally in the form of a database table according to the importance of different point clouds;

[0041] The point cloud data rendering unit is used to deploy tile data, render web pages, and object databases on a server. The Web browser loads the rendering module and tile data to achieve the rendering of large-scale point cloud data;

[0042] The data middle platform building unit is used to build a data middle platform to describe all information in the substation environment based on a standard protocol in a unified format and language;

[0043] The virtual-real mapping unit is used to establish the correlation between perception data and data, data and equipment or environment by adopting the virtual-real mapping technology of substation equipment perception data, and realize the virtual-real mapping between perception data and equipment;

[0044] The substation digital twin model building unit is used to obtain multi-dimensional and panoramic data based on comprehensive perception and data governance technology, build a digital twin virtual model of the substation equipment state, and realize the accurate model building of the equipment environment;

[0045] The digital twin virtual robot fusion unit is used to build a robot simulation environment based on ROS, and combine the collected data with the current robot motion state through interactive computer technology and robot-related theories to synchronously simulate the actual working conditions of the robot in the computer.

[0046] Thirdly, the present invention provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the steps of the method described in any one of the above are realized.

[0047] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0048] The substation digital twin method based on large-scale point cloud provided by the present invention effectively extracts key equipment data in the point cloud through technologies such as point cloud preprocessing, clustering segmentation, and line fitting, which is convenient for subsequent classification storage. Through the tile-based inner-layer storage method, point cloud data of different importance levels are stored according to different requirements, reducing the size of the point cloud data volume and facilitating subsequent rendering; through the large-scale point cloud rendering technology, combined with the actual situation, nodes that need to load and render corresponding levels are found for loading, improving the rendering efficiency of large-scale point clouds; through technologies such as digital twin virtual-real mapping, data simulation, and real-time synchronization, more intuitive monitoring and analysis of substation equipment and comprehensive management of the entire life cycle are realized; through the robot simulation technology, the robot inspection process is virtually synchronized to the digital twin virtual model of the substation, making it more convenient to monitor and manage the inspection tasks. Description of the Drawings

[0049] Figure 1 It is a schematic flowchart of a substation digital twin method based on large-scale point cloud provided by an embodiment of the present invention;

[0050] Figure 2 is a schematic diagram for screening the cross-sectional radius provided by an embodiment of the present invention;

[0051] Figure 3 is a point cloud hierarchical storage diagram provided by an embodiment of the present invention;

[0052] Figure 4 is a point cloud storage flow chart provided by an embodiment of the present invention. Detailed implementation manners

[0053] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0054] Embodiment 1

[0055] This embodiment introduces a substation digital twin method based on large-scale point clouds, including:

[0056] Obtain substation point cloud data;

[0057] Perform preprocessing on the point cloud data, downsample the point cloud data using the point cloud voxel filter algorithm, and extract effective point cloud data;

[0058] Adopt the method of Euclidean space clustering segmentation to perform entity division on the extracted effective point cloud data;

[0059] Adopt the minimum spanning tree and maximum critical path algorithms to extract straight lines in the divided target entities, and fit the straight lines to obtain a straight line model;

[0060] Identify power lines from the straight line model;

[0061] For the identified power lines, adopt a tiled hierarchical storage method, construct a storage structure based on the octree structure, and save them locally in the form of a database table according to the importance of different point clouds;

[0062] Deploy the tile data, rendering web page, and object database on the server, and the Web browser realizes the rendering of large-scale point cloud data by loading the rendering module and tile data;

[0063] Build a data middle platform, and describe all the information in the substation environment based on a standard protocol and in a unified format and language;

[0064] Adopt the virtual-real mapping technology of substation equipment perception data to establish the correlation between perception data and data, data and equipment or environment, and realize the virtual-real mapping between perception data and equipment;

[0065] Obtain multi-dimensional and panoramic data based on comprehensive perception and data governance technologies, construct a digital twin virtual model of the substation equipment status, and achieve the accurate model establishment of the equipment environment;

[0066] Construct a robot simulation environment based on ROS. Through interactive computer technology and robot-related theories, combine the collected data with the current robot motion state, and synchronously simulate the actual working conditions of the robot in the computer.

[0067] As Figures 1-4 shown, the substation digital twin method and device based on large-scale point cloud provided in this embodiment specifically involve the following steps in the application process:

[0068] Step 1: Use existing technologies to obtain substation point cloud data through lidar scanning;

[0069] Step 2: Point cloud preprocessing. Calculate the average distance from each point to all adjacent points. Assume that the result distribution is a Gaussian distribution with a mean and a standard deviation. Consider all points whose average distance is outside the interval defined by the global distance mean and standard deviation as outliers, and prune them from the dataset. Then, use the point cloud voxel filter algorithm to downsample the point cloud and extract the effective point cloud.

[0070] Step 3: Clustering segmentation. The extracted point cloud is independently distributed in space. According to this spatial distribution feature, the Euclidean space clustering segmentation method can be used to divide the target entities. This method uses the Euclidean distance between points as the measure function for judgment, that is, when the Euclidean distance between a point and any point in a certain class set is less than the threshold, it is classified into the same class;

[0071] Step 4: Line fitting. Use the minimum spanning tree and maximum critical path algorithms to extract lines from the divided target entities. Obtain the minimum spanning tree for each sub-block based on the graph theory algorithm, take the longest path in the minimum spanning tree as the approximation line of the wire, and use the central axis to extract the equipment central axis;

[0072] Step 5: Identify power lines. Identify power lines from the line model. Adopt the method of calculating the cross-sectional radius. At a certain point on the line, take the range of the coordinates of all points within a sphere with a certain radius (such as 1 meter) around it. The union of the spreading ranges of all points can be regarded as the wire to which the line belongs. If the average distance from all points within the sphere to the line is greater than a certain threshold (such as 20 centimeters), then it is considered that the line is not a power line;

[0073] Step 6: Store the recognition results in the database. For the results after object recognition, adopt the tile-based hierarchical storage method, construct a storage structure based on the octree structure, and save them in the local area in the form of a database table according to the importance of different point clouds;

[0074] Step 7: Render the point cloud data. Deploy the tile data, rendering web page, and object database on the server. The Web browser loads the rendering module and tile data to achieve the rendering of large-scale point cloud data;

[0075] Step 8: Build a data middle platform. Describe all the information in the substation environment based on a standard protocol, using a unified format and language, so that the information can flow and be shared without obstacles in the power transmission and transformation environment, realizing data collection, aggregation, unified management, and use, and providing data for the construction of a digital twin virtual model for intelligent operation;

[0076] Step 9: Lifecycle planning and management. Digitally manage the whole process of the material and value of substation equipment, and achieve comprehensive management of the entire life cycle from the planning and design of assets, material procurement, engineering construction, maintenance and operation, inspection and update to retirement and scrapping;

[0077] Step 10: Virtual-real mapping. Adopt the virtual-real mapping technology of substation equipment perception data to establish the correlation between perception data and data, data and equipment or environment, realize the virtual-real mapping between perception data and equipment, and make the analysis of data more intuitive.

[0078] Step 11: Establish a substation digital twin model. Obtain multi-dimensional and panoramic data based on comprehensive perception and data governance technology, construct a digital twin virtual model of the substation equipment state, and realize the accurate model establishment of the equipment environment. Build a substation multi-dimensional data space by loading data resources, conduct data analysis and simulation calculations on the digital twin virtual model, and feedback to the digital twin virtual model in real time to optimize and evolve the model, forming a self-optimizing intelligent operation mode;

[0079] Step 12: Integrate digital twin virtual robots. Build a robot simulation environment based on ROS, including robot kinematic analysis, motion path planning, dynamic control, robot-environment interaction, etc. Combine the collected data with the current robot motion state through interactive computer technology and relevant robot theories, and synchronously simulate the actual working conditions of the robot in the computer to facilitate the observation of the entire inspection process.

[0080] Specifically, for the given connected undirected graph G=(V, E) of the minimum spanning tree algorithm, if there is an acyclic subset A in E that contains all vertices and the sum of the weights of all edges is the smallest, then we call this subset the minimum spanning tree. The present invention uses the Kruskal algorithm to find the minimum spanning tree and uses the union-find set to determine whether the newly added vertex forms a loop in the current path;

[0081] Specifically, for the extraction of the central axis, the point cloud on the minimum spanning tree path extracted is smoothed to obtain the approximate central axis of the current straight line point cloud, and then a straight line fitting method of RANSAC is used to find a straight line that conforms to this path;

[0082] Specifically, for the RANSAC algorithm, the core is to divide the points into "inliers" and "outliers". A matrix is obtained by random sampling and solving, and then it is verified whether other points conform to the model. Then the conforming points become "inliers", and the non-conforming points become "outliers". Next, points are still sampled from the "new inlier set" to construct a new matrix, and the error is recalculated. Finally, the one with the smallest error and the largest number of points is the final model;

[0083] Specifically, for the tile-based hierarchical storage method, initially, the points read from the original point cloud file are added to the root node. If the total number of points cached in the node is lower than the threshold, the point will be assigned to this node and stored in the cache. Once enough points that potentially fall into a new child node have accumulated, a new child node is created. When a newly read point is added to a child node, the average distance between this point and its neighbors needs to be judged, and based on the distance, it is decided which layer of the octree this point should be stored in. If the current average neighbor distance is greater than the neighbor distance threshold corresponding to this point, the point is directly stored in the current child node; otherwise, the point is stored in the next-level child node with a higher level of the current child node, and so on. It can be seen that the smaller the current neighbor distance, the denser the point is, and the higher its level in the octree, which can realize the function of tile-level storage; for the points in non-critical objects, if the assigned level is greater than the maximum level allowed by the system, the point is discarded, which can realize the function of downsampling non-critical objects and reduce the point cloud storage space; the smaller the neighbor distance threshold is set, the greater the possibility that the point is placed in the low-level nodes of the octree, and it can be preferentially displayed during breadth-first traversal rendering, so that the function of improving its rendering priority can be realized;

[0084] Specifically, for the calculation of the average distance between a point and its neighbors, a hash table is used to cache the neighbor distances. With the three-dimensional coordinate combination as the index, the retrieved hash value is a cell address. Inside the cell, there are the coordinates of multiple points in this cell and the addresses of multiple neighbor cells, which allows multiple points to share a cell, thus solving the hash conflict. Given the three-dimensional coordinates of a certain point, the average distance between it and multiple points in this cell and all points in each neighbor can be calculated;

[0085] Specifically, for the tile-based hierarchical storage method, the method includes the following steps:

[0086] Step 1: Read the object extraction result from the database, generate the root node, and set it as the current node;

[0087] Step 2: Read a point from the original point cloud file. If the reading is successful, go to Step 3; otherwise, if all the data has been read, go to Step 6.

[0088] Step 3: Add the read point to the current node. Determine whether the octree node being currently operated on is a leaf node. If it is, go to Step 4; otherwise, go to Step 5.

[0089] Step 4: Put the new point into the data cache of the node. When the number of points in the cache exceeds a certain threshold, the leaf node splits into a non-leaf node, and the points in the cache are hierarchically assigned to the child nodes, then go to Step 3.

[0090] Step 5: Calculate the average distance between this point and its neighbors. If the average distance is greater than the neighbor threshold of the category to which this node belongs, it means this point is sparse, then put the new point into the data cache of the node; otherwise, assign this point to the next-level child node. Then go to Step 2.

[0091] Step 6: Construct a hierarchical file. The hierarchical file saves the hierarchical relationships between various nodes in the order of breadth-first traversal. The saved content includes how many points the node itself has, which child nodes the node has, etc.

[0092] Step 7: During the construction of the hierarchical structure, serialize and store the octree nodes regularly, and save the tile files at different levels of detail. The file names of the tile files are named according to their levels.

[0093] Specifically, for the rendering of the point cloud data, when rendering the web page, it is necessary to combine the current viewing angle and the viewing range, apply the multi-thread dynamic scheduling technology and the local rendering technology, calculate the rendering resolutions of different regions within the viewing range, so as to find the nodes at the corresponding levels that need to be loaded and rendered for loading, or eliminate the nodes at the corresponding levels according to the viewing distance.

[0094] Embodiment 2

[0095] This embodiment provides a substation digital twin device based on large-scale point clouds, including:

[0096] An acquisition unit, used to acquire substation point cloud data;

[0097] A point cloud preprocessing unit, used to preprocess the point cloud data, downsample the point cloud data using the point cloud voxel filter algorithm, and extract effective point cloud data;

[0098] A clustering and segmentation unit, used to perform entity division on the extracted effective point cloud data by using the method of Euclidean space clustering and segmentation;

[0099] A straight line fitting unit, which is used to extract straight lines from the divided target entities by using the minimum spanning tree and maximum critical path algorithms, and fit the straight lines to obtain a straight line model;

[0100] A power line recognition unit, which is used to recognize power lines from the straight line model;

[0101] A storage unit, which is used to adopt a tiled hierarchical storage method for the recognized power lines, construct a storage structure based on an octree structure, and save them locally in the form of a database table according to the importance of different point clouds;

[0102] A point cloud data rendering unit, which is used to deploy tile data, a rendering web page, and an object database on a server. The Web browser loads the rendering module and tile data to realize the rendering of large-scale point cloud data;

[0103] A data middle platform building unit, which is used to build a data middle platform and describe all information in the substation environment based on a standard protocol in a unified format and language;

[0104] A virtual-real mapping unit, which is used to adopt the virtual-real mapping technology of substation equipment perception data to establish the correlation between perception data and data, data and equipment or environment, and realize the virtual-real mapping between perception data and equipment;

[0105] A substation digital twin model building unit, which is used to obtain multi-dimensional and panoramic data based on comprehensive perception and data governance technology, construct a digital twin virtual model of the substation equipment state, and realize the accurate model building of the equipment environment;

[0106] A digital twin virtual robot fusion unit, which is used to build a robot simulation environment based on ROS, and combine the collected data with the current robot motion state through interactive computer technology and robot-related theories to synchronously simulate the actual working conditions of the robot in the computer.

[0107] Embodiment 3

[0108] This embodiment provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, it realizes the steps of any one of the following methods:

[0109] Obtain substation point cloud data;

[0110] Preprocess the point cloud data, downsample the point cloud data by using the point cloud voxel filter algorithm, and extract the effective point cloud data;

[0111] Adopt the method of Euclidean space clustering segmentation to conduct entity division on the extracted effective point cloud data;

[0112] Extract straight lines from the divided target entities using the minimum spanning tree and maximum critical path algorithms, and fit the straight lines to obtain a straight line model;

[0113] Identify power lines from the straight line model;

[0114] For the identified power lines, adopt a tile-based hierarchical storage method, construct a storage structure based on the octree structure, and save them locally in the form of a database table according to the importance of different point clouds;

[0115] Deploy tile data, rendering web pages, and object databases on the server. The Web browser realizes the rendering of large-scale point cloud data by loading the rendering module and tile data;

[0116] Build a data middle platform to describe all information in the substation environment based on a standard protocol using a unified format and language;

[0117] Adopt the virtual-real mapping technology of substation equipment perception data to establish the correlation between perception data and data, data and equipment or environment, and realize the virtual-real mapping between perception data and equipment;

[0118] Based on comprehensive perception and data governance technology, obtain multi-dimensional and panoramic data, construct a digital twin virtual model of the substation equipment state, and realize the accurate model establishment of the equipment environment;

[0119] Construct a robot simulation environment based on ROS. Combine the collected data with the current robot motion state through interactive computer technology and robot-related theories to synchronously simulate the actual working conditions of the robot in the computer.

[0120] The above are only the preferred embodiments of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the technical principle of the present invention, several improvements and deformations can be made, and these improvements and deformations should also be regarded as the protection scope of the present invention.

Claims

1. A substation digital twin method based on large-scale point clouds, characterized in that, Including: Obtain the substation point cloud data; Preprocess the point cloud data, downsample the point cloud data using the point cloud voxel filter algorithm, and extract the effective point cloud data; Use the Euclidean space clustering segmentation method to perform entity division on the extracted effective point cloud data; Use the minimum spanning tree and maximum critical path algorithms to extract straight lines from the divided target entities, and fit the straight lines to obtain a straight line model; Identify the power lines from the straight line model; For the identified power lines, adopt a tiled hierarchical storage method, construct a storage structure based on the octree structure, and save them in a database table locally according to the importance of different point clouds; Deploy the tile data, rendering web page, and object database on the server. The Web browser realizes the rendering of large-scale point cloud data by loading the rendering module and tile data; Build a data middle platform to describe all the information in the substation environment based on a standard protocol and in a unified format and language; Adopt the virtual-real mapping technology of substation equipment perception data to establish the correlation between perception data and data, data and equipment or environment, and realize the virtual-real mapping between perception data and equipment; Based on comprehensive perception and data governance technology, obtain multi-dimensional and panoramic data, construct a digital twin virtual model of the substation equipment status, and realize the accurate model establishment of the equipment environment; Build a robot simulation environment based on ROS. Through interactive computer technology and robot-related theories, combine the collected data with the current robot motion state, and synchronously simulate the actual working conditions of the robot in the computer.

2. The method and device for digital twin of a substation based on large-scale point clouds according to claim 1, wherein: The preprocessing of the point cloud data includes: calculating the average distance from each point to all adjacent points. Assuming that the result distribution is a Gaussian distribution with a mean and standard deviation, regard all points whose average distance is outside the interval defined by the global distance mean and standard deviation as outliers, and prune them from the data set.

3. The method and device for substation digital twin based on large-scale point cloud according to claim 1, wherein: The method of using the Euclidean space clustering segmentation to perform entity division on the extracted effective point cloud data includes: using the Euclidean distance between points as a measure function for judgment, that is, when the Euclidean distance between this point and any point in a certain class set is less than the threshold, it is classified into the same class for division.

4. The method and device for substation digital twin based on large-scale point cloud according to claim 1, characterized in that: The method of using the minimum spanning tree and maximum critical path algorithms to extract straight lines from the divided target entities includes: obtaining the minimum spanning tree for each sub-block based on the graph theory algorithm, taking the longest path in the minimum spanning tree as the approximation straight line of the wire, and using the central axis to extract the equipment central axis.

5. The method and device for substation digital twin based on large-scale point cloud according to claim 1, characterized in that: The method of identifying the power lines from the straight line model includes: identifying the power lines from the straight line model. Adopt the method of calculating the cross-sectional radius. At a certain point on the straight line, take the range of the coordinates of all points within a sphere with a certain radius around it. The union of the spreading ranges of all points can be regarded as the wire to which this straight line belongs. If the average distance from all points within the sphere to the straight line is greater than the preset threshold, it is considered that this straight line is not a power line.

6. The method and device for digital twin of substation based on large-scale point cloud according to claim 1, wherein: Also including: Digitally manage the whole process of the substation equipment's materials and value, and realize comprehensive management of the entire life cycle of asset planning and design, material procurement, engineering construction, maintenance and operation, inspection and update until retirement and scrapping.

7. The method and device for digital twin of a substation based on large-scale point cloud according to claim 1, characterized in that: The construction of a digital twin virtual model for substation equipment status to achieve an accurate model of the equipment environment further includes: constructing a multi-dimensional data space for the substation by loading data resources, performing data analysis and simulation calculations on the digital twin virtual model, and real-time feedback to the digital twin virtual model for model optimization and evolution, forming a self-optimizing intelligent operation mode.

8. The method and device for digital twin of substation based on large-scale point cloud according to claim 1, characterized in that: The tile-based hierarchical storage method includes: Step 1: Read the object extraction result from the database, generate a root node, and set it as the current node; Step 2: Read a point from the original point cloud file. If the reading is successful, go to Step 3; otherwise, if all the data has been read, go to Step 6; Step 3: Add the read point to the current node. Determine whether the octree node of the current operation is a leaf node. If it is, go to Step 4; otherwise, go to Step 5; Step 4: Put the new point into the data cache of the node. When the number of points in the cache exceeds the set threshold, the leaf node splits into non-leaf nodes, and the points in the cache are gradually assigned to the child nodes, then go to Step 3; Step 5: Calculate the average distance between this point and its neighbors. If the average distance is greater than the preset distance threshold of the category to which the node belongs, it means the point is sparse, then put the new point into the data cache of the node; otherwise, assign the point to the next-level child node and go to Step 2; Step 6: Construct a hierarchical file. The hierarchical file saves the hierarchical relationship between each node in the order of breadth-first traversal. The saved content includes how many points the node itself has and which child nodes the node has; Step 7: During the construction of the hierarchical structure, periodically serialize and store the nodes of the octree, and save the tile files of different levels of detail. The file names of the tile files are named according to their levels.

9. A digital twin device for a substation based on large-scale point clouds, characterized in that, It includes: An acquisition unit for acquiring substation point cloud data; A point cloud preprocessing unit for preprocessing the point cloud data, downsampling the point cloud data using the point cloud voxel filter algorithm, and extracting effective point cloud data; A clustering and segmentation unit for entity division of the extracted effective point cloud data using the Euclidean space clustering and segmentation method; A line fitting unit for extracting lines in the divided target entities using the minimum spanning tree and maximum critical path algorithms, and fitting the lines to obtain a line model; A power line identification unit for identifying power lines from the line model; A storage unit for storing the identified power lines using the tile-based hierarchical storage method, constructing a storage structure based on the octree structure, and saving them locally in the form of a database table according to the importance of different point clouds; A point cloud data rendering unit for deploying tile data, rendering web pages, and object databases on the server. The Web browser realizes the rendering of large-scale point cloud data by loading the rendering module and tile data; A data middle platform building unit for building a data middle platform to describe all information in the substation environment based on a standard protocol in a unified format and language; The virtual-real mapping unit is used to establish the correlation between the sensed data and data, data and equipment or environment by adopting the virtual-real mapping technology of substation equipment sensed data, and realize the virtual-real mapping between the sensed data and equipment; The substation digital twin model establishment unit is used to obtain multi-dimensional and panoramic data based on comprehensive sensing and data governance technology, construct a digital twin virtual model of the substation equipment state, and realize the accurate model establishment of the equipment environment; The digital twin virtual robot fusion unit is used to construct a robot simulation environment based on ROS, and combine the collected data with the current robot motion state through interactive computer technology and robot-related theories to synchronously simulate the actual working conditions of the robot in the computer.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, it realizes the steps of the method described in any one of claims 1 to 8.

Images