Intelligent inspection method and system based on digital twinborn sum graph database

Abstract

The invention relates to the technical field of intelligent inspection, and discloses an intelligent inspection method and system based on a digital twinborn sum graph database, and the method comprises the steps: carrying out the semantic analysis of an inspection task book, obtaining a structured task table, inputting the equipment name in the structured task table into a graph database, and obtaining an equipment information table; constructing an initial digital twinborn map of the inspection environment, and performing calibration according to the equipment theoretical coordinates in the equipment information table to obtain a digital twinborn map; generating a reachable path by adopting a path planning algorithm based on density clustering, and sampling inspection points to obtain a safe inspection point set; and according to the safety inspection point set and the obstacle coordinate set, adopting a multi-target path planning algorithm to generate an optimal inspection path. According to the invention, through multi-modal data fusion and coordinate calibration, the precision of digital twin mapping is improved, and through a path planning algorithm based on dynamic environment perception, the path planning efficiency and the safety of a routing inspection path are improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent inspection technology, and in particular to an intelligent inspection method and system based on digital twins and a graph database. Background Art

[0002] Currently, the equipment inspection methods for power systems are mainly divided into two types: manual inspection and automated inspection. Manual inspection involves inspectors following a pre-established inspection task book and a fixed route to inspect power equipment one by one. For example, in a power substation inspection scenario, inspectors will divide the area into high-voltage equipment areas, low-voltage equipment areas, transformer rooms, and other areas, and inspect various types of equipment in sequence. The limitation of this method is that different inspectors may have different understandings of the inspection task book, resulting in inconsistent inspection content, which may lead to some equipment being missed or over-inspected. Furthermore, manual inspection routes are usually carried out in a fixed order and lack consideration for real-time environmental changes. When encountering emergencies such as equipment failures or on-site construction, inspectors find it difficult to adjust the route in a timely manner, resulting in low inspection efficiency and potentially increased safety risks. Automated inspections involve inspection robots performing intelligent inspections of power equipment along pre-defined inspection routes. Although this method can overcome the influence of human factors in manual inspections, its inspection routes are still based on static path planning in a fixed inspection environment and cannot respond to moving obstacles and equipment position offsets in real time. Therefore, frequent path failures are prone to occur, making it difficult to balance path safety and operation and maintenance efficiency. Summary of the Invention

[0003] In order to solve the above technical problems, the present invention provides an intelligent inspection method and system based on digital twins and graph databases, which can solve the problem of low inspection efficiency caused by the existing methods due to the disconnection of inspection path planning from the dynamic environment and the inability to dynamically avoid sudden obstacles, thereby achieving the effect of improving the inspection efficiency of power equipment and ensuring the safety of inspection paths.

[0004] In a first aspect, the present invention provides an intelligent inspection method based on digital twins and a graph database, the method comprising: Perform semantic parsing on the inspection task book of the power equipment to obtain a structured task table, and input the equipment names in the structured task table into a preset graph database to obtain an equipment information table; Constructing an initial digital twin map of the inspection environment, and calibrating the device point cloud coordinates in the initial digital twin map according to the device theoretical coordinates in the device information table to obtain a digital twin map; According to the device coordinate set and the obstacle coordinate set in the digital twin map, a density clustering-based path planning algorithm is used to generate a reachable path, and inspection point sampling is performed on the reachable path to obtain a set of safety inspection points; Generate an optimal inspection path using a multi-objective path planning algorithm based on the set of safety inspection points and the set of obstacle coordinates; According to the optimal inspection path and the equipment information table, the inspection robot is controlled to perform equipment inspection, equipment status data is obtained, and based on the equipment status data, it is determined whether there is any abnormality in the power equipment.

[0005] Furthermore, the steps of constructing an initial digital twin map of the inspection environment and calibrating the device point cloud coordinates in the initial digital twin map according to the device theoretical coordinates in the device information table to obtain the digital twin map include: Acquire sensor data of the inspection environment, and perform data fusion and filtering noise reduction on the sensor data to generate an initial digital twin map. The sensor data includes lidar point cloud data, infrared thermal imaging temperature data, and camera visual data. Using the least squares method to align the device point cloud data in the initial digital twin map with the device theoretical coordinates in the device information table and remove noise to obtain calibrated device point cloud data; The initial digital twin map is updated according to the calibrated device point cloud data to obtain a digital twin map.

[0006] Furthermore, the step of generating a reachable path based on the device coordinate set and the obstacle coordinate set in the digital twin map using a density clustering-based path planning algorithm, and sampling inspection points on the reachable path to obtain a set of safety inspection points includes: Using a density clustering algorithm to perform cluster analysis on the device coordinate set in the digital twin map to obtain several device-dense areas; Generate a reachable path in each device-dense area using a path planning algorithm based on the device coordinate set and the obstacle coordinate set in the digital twin map; Inspection points are sampled on the reachable path at fixed intervals to obtain a set of safe inspection points.

[0007] Furthermore, the step of using a density clustering algorithm to perform cluster analysis on the device coordinate set in the digital twin map to obtain several device-dense areas includes: According to the device coordinate set, based on the sliding window mechanism, the Euclidean distance between adjacent devices is calculated to obtain a local distance sequence corresponding to each sliding window; According to the local spacing sequence of each sliding window, the local average spacing and the local spacing standard deviation of each sliding window are calculated, and the global average spacing is calculated; Determine whether the ratio of the local spacing standard deviation to the global average spacing of each sliding window is less than a distance threshold; if so, use the sliding window as the first sliding window; otherwise, use the sliding window as the second sliding window; Obtaining, according to the global average spacing, the local average spacing, and a preset scaling factor, a first neighborhood radius of each power device in the first sliding window and a second neighborhood radius of each power device in the second sliding window; According to the first neighborhood radius and the second neighborhood radius, a density-based noise application spatial clustering algorithm is used to perform cluster analysis on the power equipment to obtain a plurality of equipment-intensive areas.

[0008] Furthermore, the step of obtaining a first neighborhood radius of each power device in the first sliding window and a second neighborhood radius of each power device in the second sliding window according to the global average spacing, the local average spacing, and a preset scaling factor includes: Using the product of the global average distance and a preset scaling factor as the first neighborhood radius of each power device in the first sliding window; The product of the local average distance and the scaling factor is used as the second neighborhood radius of each power device in the second sliding window.

[0009] Furthermore, the step of generating a reachable path in each device-dense area using a path planning algorithm based on the device coordinate set and the obstacle coordinate set in the digital twin map includes: Based on the device coordinate set and the obstacle coordinate set in the digital twin map, an improved fast search random tree algorithm is used to perform node search to obtain a reachable path in each device-dense area; The path cost function of the improved fast search random tree algorithm consists of the device path length and the obstacle risk value.

[0010] Furthermore, the improved fast search random tree algorithm adopts a density-guided sampling strategy to perform random point sampling.

[0011] Furthermore, the step of generating an optimal inspection path using a multi-objective path planning algorithm based on the set of safety inspection points and the set of obstacle coordinates includes: Calculating an obstacle risk value for each safety inspection point based on the safety inspection point set and the obstacle coordinate set; The objective function is to minimize the path length and the sum of the obstacle risk values ​​of each safety inspection point, and a multi-objective path planning algorithm is used to solve the problem to obtain the optimal inspection path.

[0012] Furthermore, the multi-objective path planning algorithm is any one of a genetic algorithm, a particle swarm optimization algorithm and an ant colony algorithm.

[0013] In a second aspect, the present invention provides an intelligent inspection system based on digital twins and a graph database, the system comprising: An information query module is used to perform semantic analysis on the inspection task book of the power equipment to obtain a structured task table, and input the equipment names in the structured task table into a preset graph database to obtain an equipment information table; A digital twin module is used to construct an initial digital twin map of the inspection environment and calibrate the device point cloud coordinates in the initial digital twin map according to the device theoretical coordinates in the device information table to obtain a digital twin map; A checkpoint generation module is used to generate a reachable path based on the device coordinate set and the obstacle coordinate set in the digital twin map using a density clustering-based path planning algorithm, and to sample checkpoints on the reachable path to obtain a set of safe checkpoints; A path planning module, configured to generate an optimal inspection path using a multi-objective path planning algorithm based on the set of safety inspection points and the set of obstacle coordinates; The intelligent inspection module is used to control the inspection robot to perform equipment inspection according to the optimal inspection path and the equipment information table, obtain equipment status data, and determine whether there is any abnormality in the power equipment based on the equipment status data.

[0014] This invention provides an intelligent inspection method and system based on digital twins and a graph database. Through multimodal data fusion and coordinate calibration, the invention can improve the accuracy of digital twin mapping. A path planning algorithm based on dynamic environmental perception can improve path planning efficiency and inspection path safety. Intelligent data analysis can improve the accuracy of fault analysis, thereby improving the operation and maintenance efficiency of power equipment and ensuring the safety and stability of power equipment operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of an intelligent inspection method based on digital twins and a graph database in an embodiment of the present invention; Figure 2 It is a structural diagram of the intelligent inspection system based on digital twins and graph database in an embodiment of the present invention. DETAILED DESCRIPTION

[0016] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0017] See also Figure 1 The first embodiment of the present invention proposes an intelligent inspection method based on digital twins and a graph database, including steps S10 to S50: Step S10: semantically analyze the inspection task book of the power equipment to obtain a structured task table, and input the equipment names in the structured task table into a preset graph database to obtain an equipment information table; Step S20: construct an initial digital twin map of the inspection environment, and calibrate the device point cloud coordinates in the initial digital twin map according to the device theoretical coordinates in the device information table to obtain a digital twin map; Step S30: Generate a reachable path using a density clustering-based path planning algorithm based on the device coordinate set and the obstacle coordinate set in the digital twin map, and sample inspection points on the reachable path to obtain a set of safety inspection points; Step S40, generating an optimal inspection path using a multi-objective path planning algorithm based on the set of safety inspection points and the set of obstacle coordinates; Step S50 , controlling the inspection robot to perform equipment inspection according to the optimal inspection path and the equipment information table, obtaining equipment status data, and determining whether there is any abnormality in the power equipment according to the equipment status data.

[0018] In this method, the inspection task list for power equipment, described in natural language, is first semantically parsed and structured. Specifically, a pre-trained language model (BERT model) is used to perform semantic analysis on the inspection task list. The name of the power equipment to be inspected (e.g., transformer), inspection items (e.g., oil level testing), and inspection period (e.g., daily) are extracted from the inspection task list. For missing periods (e.g., if a piece of equipment in the inspection task list requires regular inspections), regular expressions and contextual inference are used to supplement them with specific periods (e.g., monthly). The extracted equipment names, inspection items, and inspection periods are then structured to produce a structured task list.

[0019] Because power equipment typically has multidimensional relationships such as spatial location and functional dependencies, traditional databases require multiple table joins to perform queries. Consequently, performance degrades dramatically when data volumes are large. Furthermore, when relationships between devices change dynamically (such as when new equipment is added or when locations are adjusted), relational databases require frequent table modifications, resulting in high maintenance costs. Therefore, in this embodiment, a graph database is used for device management. Graph databases are capable of handling complex relationships. In a graph database, entities such as equipment, locations, and maintenance records are represented as nodes, and relationships are represented as edges. These relationships can be traversed directly without the need for multiple table joins. When a device moves or a new dependency is added, only the edge attributes need to be modified, without the need to restructure the table.

[0020] For the device names in the structured task table, they are input into the pre-established graph database for associated retrieval to obtain the device type, spatial coordinates, and functional description. The functional description is then segmented and vectorized to verify its consistency with the device type. After the consistency verification is passed, the device information table is output. The device information table contains the device name, device type, device theoretical coordinates, and device functional description.

[0021] In the power inspection environment, there are dynamic obstacles such as material handling. The inspection path generated only based on the static coordinates of the equipment will cause the inspection robot to frequently collide or stagnate during movement. Therefore, in this embodiment, a digital twin map of the inspection environment is established through real-time data fusion to ensure that the inspection path always conforms to the actual scenario. The specific generation steps include: Acquire sensor data of the inspection environment, and perform data fusion and filtering noise reduction on the sensor data to generate an initial digital twin map. The sensor data includes lidar point cloud data, infrared thermal imaging temperature data, and camera visual data. Using the least squares method to align the device point cloud data in the initial digital twin map with the device theoretical coordinates in the device information table and remove noise to obtain calibrated device point cloud data; The initial digital twin map is updated according to the calibrated device point cloud data to obtain a digital twin map.

[0022] In this embodiment, various types of sensor data about the inspection environment are first obtained, including lidar point cloud data, infrared thermal imaging temperature data, and camera vision data, and then multi-sensor data fusion is performed. Specifically, the data acquisition timestamps of the lidar, camera, and temperature sensor are aligned through hardware trigger signals to ensure spatiotemporal consistency. The RGB-D image and the lidar point cloud are aligned using the ICP (Iterative Closest Point) algorithm to generate a color point cloud. The infrared thermal imaging data is then associated with the point cloud coordinates to generate a three-dimensional point cloud with temperature attributes. The three-dimensional point cloud is then subjected to Kalman filtering for denoising to obtain the denoised three-dimensional point cloud, thereby forming an initial digital twin map. The initial digital twin map contains at least the point cloud data of the power equipment and obstacles around the equipment.

[0023] The least squares method is then used to align the device point cloud data in the initial digital twin map with the theoretical device coordinates in the device information table, generating aligned device point cloud data. This aligned device point cloud data is then fed into a trained SVM classification model to classify device points and background noise. Points classified as background noise are then removed to generate calibrated device point cloud data. Finally, the calibrated device point cloud data and real-time sensor data are used to reconstruct the digital twin map.

[0024] After the digital twin map is constructed, a density-based clustering-based path planning algorithm is used to generate a reachable path based on the device coordinate set and obstacle coordinate set in the map. Inspection points are sampled on the reachable path to obtain a set of safety inspection points. The specific steps include: Using a density clustering algorithm to perform cluster analysis on the device coordinate set in the digital twin map to obtain several device-dense areas; Generate a reachable path in each device-dense area using a path planning algorithm based on the device coordinate set and the obstacle coordinate set in the digital twin map; Inspection points are sampled on the reachable path at fixed intervals to obtain a set of safe inspection points.

[0025] In this embodiment, a density clustering algorithm, such as the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) or the OPTICS (Ordering Points To Identify the Clustering Structure) algorithm, is first used to identify densely distributed areas of equipment. Then, within each densely populated area, a path planning algorithm, such as the RRT (Rapidly Exploring Random Tree) or the PRM (Probabilistic Roadmap) algorithm, is used to generate reachable paths within the area. Finally, uniform sampling is performed along the reachable paths to obtain safety inspection points. These safety inspection points in each area are then combined into a safe and reachable inspection point set, thus forming a safety inspection point set.

[0026] This embodiment identifies equipment-dense areas and generates inspection points in each area, ensuring that the inspection points cover all key equipment. It filters points blocked by obstacles through accessible paths, improves the safety of the inspection points, and transforms the full-space search problem into a local path optimization problem, reducing the complexity of the inspection point generation calculation.

[0027] In a preferred embodiment, the present invention uses an improved DBSCAN algorithm to perform cluster analysis, and the specific steps include: According to the device coordinate set, based on the sliding window mechanism, the Euclidean distance between adjacent devices is calculated to obtain a local distance sequence corresponding to each sliding window; According to the local spacing sequence of each sliding window, the local average spacing and the local spacing standard deviation of each sliding window are calculated, and the global average spacing is calculated; Determine whether the ratio of the local spacing standard deviation to the global average spacing of each sliding window is less than a distance threshold; if so, use the sliding window as the first sliding window; otherwise, use the sliding window as the second sliding window; Obtaining, according to the global average spacing, the local average spacing, and a preset scaling factor, a first neighborhood radius of each power device in the first sliding window and a second neighborhood radius of each power device in the second sliding window; According to the first neighborhood radius and the second neighborhood radius, a density-based noise application spatial clustering algorithm is used to perform cluster analysis on the power equipment to obtain a plurality of equipment-intensive areas.

[0028] In this embodiment, the DBSCAN algorithm (also known as the density-based spatial clustering of noise applications) is a density-based clustering algorithm that can discover clusters of arbitrary shapes and effectively process noisy data. Its core idea is to define density by neighborhood radius and minimum number of points (min_samples) and divide data points into core points, boundary points, and noise points.

[0029] A core point is a sample whose neighborhood contains at least min_samples points; a boundary point is a point within the neighborhood of a core point but does not meet the core point criteria; a noise point is a sample that is neither a core point nor a boundary point; and density-reachable and density-connected clusters are formed by recursively expanding the neighborhood of a core point. The algorithm involves the following steps: traversing all unvisited points to determine if they are core points; if so, recursively expanding their neighborhood to group density-reachable points into the same cluster; and marking non-core points that cannot be expanded as noise.

[0030] The difference between the improved DBSCAN algorithm in this embodiment and the conventional DBSCAN algorithm lies in the definition of the neighborhood radius. In the conventional DBSCAN algorithm, the neighborhood radius is a fixed value pre-set by the user, that is, a global fixed value. In the actual scenario of power equipment inspection, the equipment distribution is uneven, and the global fixed value cannot adapt to the changes in data density, resulting in over-segmentation or under-segmentation. Over-segmentation means that the dense area is divided into multiple small clusters, and under-segmentation means that the equipment in the sparse area is not effectively clustered.

[0031] To solve the problem of uneven segmentation, this embodiment adopts a dynamic neighborhood radius calculation method. By judging the device spacing differences in the local area, the neighborhood radius is dynamically adjusted. Specifically, a sliding window is first defined. Assume that the window size is k and the moving step is l. That is, each time the spacing of k consecutive devices is analyzed, the window is sequentially slid in the device sequence, and the position of l devices is moved each time.

[0032] Based on the sliding window mechanism, the Euclidean distance between adjacent devices is calculated to obtain the local spacing sequence within each sliding window. Based on the local spacing sequences within all sliding windows, the global average spacing between all devices, as well as the local average spacing and local spacing standard deviation corresponding to each sliding window are calculated. Based on the ratio of the local spacing standard deviation to the global average spacing, it is judged whether the spacing difference in the local area is too large. If the ratio is greater than the spacing threshold, it is considered that the spacing difference is significant, and the neighborhood radius of all device points in the window needs to be dynamically adjusted. Otherwise, the neighborhood radius does not need to be dynamically adjusted. The sliding window that does not need to be adjusted is defined as the first sliding window, and the sliding window that needs to be adjusted is defined as the second sliding window. Different methods are used to determine the neighborhood radius of device points in different sliding windows. The specific steps include: Using the product of the global average distance and a preset scaling factor as the first neighborhood radius of each power device in the first sliding window; The product of the local average distance and the scaling factor is used as the second neighborhood radius of each power device in the second sliding window.

[0033] In this embodiment, for the first sliding window, the first neighborhood radius of the device point in the window is calculated based on the global average spacing: Where, represents the first neighborhood radius, represents the global average spacing, and α represents the scaling factor; preferably, α=0.5.

[0034] For the second sliding window, the second neighborhood radius of the device point within the window is calculated based on the local average spacing: Where, represents the second neighborhood radius, represents the local average spacing.

[0035] Finally, we perform cluster analysis using the DBSCAN algorithm based on the neighborhood radius of each sliding window (i.e., the first neighborhood radius of the device points in the first sliding window and the second neighborhood radius of the device points in the second sliding window), thereby obtaining multiple device-dense areas. The specific clustering steps can be referred to the conventional steps of the DBSCAN algorithm and will not be detailed here.

[0036] In this embodiment, local statistics are calculated by sliding the window and the neighborhood radius is adaptively adjusted, which can reduce the radius in dense areas and expand the radius in sparse areas, thereby avoiding over-segmentation and under-segmentation and improving the rationality of region division.

[0037] In another preferred embodiment, the present invention uses an improved RRT* algorithm for reachable path planning. The RRT* (Rapidly-exploring Random Tree Star) algorithm is an improved version of the RRT (also known as the Rapidly Exploring Random Tree Star) algorithm. By introducing reconnection and optimization steps, the generated path is gradually optimized, making the quality of the final path higher. The reachable path generation steps of the conventional RRT* algorithm include: structuring the tree structure, starting from the root node, and defining the search space; generating random points by random sampling in the search space; finding the node closest to the random point in the tree, expanding from the nearest node to the random point, generating candidate nodes, and checking whether the straight path from the nearest node to the candidate node intersects with the obstacle. If so, re-random sampling is performed. If not, a dynamic neighborhood radius is defined according to the tree node, and all neighboring nodes of the candidate node are found within the neighborhood radius. The node with the smallest path cost is selected from the neighboring nodes as the parent node of the candidate node. For each neighboring node, check whether the path to the neighboring node through the candidate node is better. If so, the parent node of the neighboring node is changed to the candidate node and its path cost is updated. If the candidate node enters the target area, the complete path is returned.

[0038] The difference between the improved RRT* algorithm in this embodiment and the conventional RRT* algorithm is that the path cost function in the conventional algorithm is represented by the total path length consisting of the cumulative path length from the starting point to the parent node and the Euclidean distance from the parent node to the candidate node. The optimization goal of the conventional path cost function is to minimize the total path length, and it does not take into account the importance of different devices in the actual inspection environment and the risk impact of dynamic obstacles around the devices. Therefore, in this embodiment, on the basis of the conventional path cost function, an obstacle risk value based on the device weight is added to each candidate node, where the obstacle risk value can be expressed as: Where Risk(p) represents the obstacle risk value of node p, The obstacle density around node p is obtained by dividing the number of obstacles within a preset radius around node p by the area of ​​the range. The weight of the device closest to the node p is a preset value and is positively correlated with the historical failure rate of the device. a represents the weight coefficient, and b represents the density coefficient.

[0039] By modifying the path cost function according to the obstacle risk value, we can obtain: Where Cost(p) represents the path cost of node p, Cost(p f ) represents the parent node p fThe cumulative path cost, λ represents the risk weight coefficient, and Risk(p) represents the obstacle risk value of node p.

[0040] In the above path cost function, Cost(p f ) is the historical cumulative cost from the starting point to the parent node. Using the parent node cost as the basis can ensure that the path cost of the new node, that is, the candidate node p, inherits the global optimality, avoids isolated optimization of path fragments, and ensures that the overall path cost from the starting point to the end point is minimized. Since the parent node Cost(p f ) has integrated all the target parameters from the starting point to the parent node, and the cost of the new node only needs to accumulate incremental items on this basis. Under the premise of ensuring global optimality, repeated calculations are reduced and the efficiency of the algorithm is improved. In the parent node cost, the parent node path has given priority to covering high-weight devices. In the incremental item, the physical distance is accumulated through the Euclidean distance, and the obstacle density and device status of the new node are quantified in real time through the risk value to avoid the risk accumulation of fixed paths. The path cost function provided in this embodiment can ensure that global optimality and dynamic environment adaptability are taken into account when the new node is expanded. It should be noted here that the expression of the risk value and the expression of the path cost function will normalize the parameters in the expression during the actual calculation, convert them into dimensionless parameters, and then calculate each dimensionless parameter. Of course, conventional path planning algorithms can also be used for the planning of reachable paths, such as the fast search random tree algorithm, the ‌A* algorithm‌, the Dijkstra algorithm, etc. The above-mentioned improved fast random tree algorithm is only a preferred method and not a specific limitation.

[0041] In addition, another difference between the improved RRT* algorithm in this embodiment and the conventional RRT* algorithm is that the conventional RRT* algorithm uses random sampling. However, due to the uneven distribution of equipment in actual inspection environments, completely random sampling may lead to circuitous paths or ignore key areas. Therefore, in this embodiment, the improved RRT* algorithm uses density-guided sampling, combined with the clustering results of the DBSCAN algorithm, to select different sampling areas with a certain probability, so as to prioritize the generation of paths in equipment-dense areas. Specifically, a uniformly distributed random number c∈[0,1] is generated at each sampling. If c≤c1, assuming c1=0.8, a cluster is randomly selected from the clustering results, that is, a device-dense area is randomly selected, and a random point is generated within its bounding box. If c>c1, random points are randomly generated in the free space of the entire map. The value of c1 can be adjusted according to actual needs to adapt to different scenarios. For example, in high-security scenarios, increasing c1 can increase the sampling probability within the area and reduce the risk of the robot entering unknown areas. In highly dynamic scenarios, decreasing c1 can reduce the sampling probability within the area and enhance global exploration capabilities. The density-guided sampling strategy of this embodiment can ensure that areas with dense equipment are covered first and the detection frequency of key equipment is increased, while retaining the ability to discover new paths and being able to cope with dynamic obstacles and complex environments.

[0042] For each accessible path within the device density area, inspection points are sampled along the accessible path at fixed intervals to obtain a set of safe inspection points. Then, based on the set of safe inspection points and the set of obstacle coordinates, a dynamic multi-objective path planning algorithm is used to generate the optimal inspection path. The specific steps include: Calculating an obstacle risk value for each safety inspection point based on the safety inspection point set and the obstacle coordinate set; The objective function is to minimize the path length and the sum of the obstacle risk values ​​of each safety inspection point, and a multi-objective path planning algorithm is used to solve the problem to obtain the optimal inspection path.

[0043] In this embodiment, the obstacle risk value Risk(s) of each safety inspection point is first considered. i ), the calculation formula is the same as the above obstacle risk value calculation formula, and then the multi-objective optimization function is established by minimizing the path length and the sum of the obstacle risk values: Where m represents the total number of safety inspection points, Indicates safety inspection point s i+1 and safety inspection points i The Euclidean distance between Risk(s i ) represents the safety inspection point s iIt should be noted that if the above-mentioned path planning algorithm adopts a conventional path planning algorithm and does not adopt the improved fast search random tree algorithm, the multi-objective optimization function here can still be constructed based on the obstacle risk value calculation formula in the above embodiment.

[0044] For the above objective function, conventional multi-objective path planning algorithms can be used to solve it, such as genetic algorithms, particle swarm optimization algorithms, or ant colony algorithms. Taking the NSGA-II algorithm (also known as the non-dominated sorting genetic algorithm) as an example, its algorithm steps include: initializing the population and randomly generating a path sequence; non-dominated sorting, grading according to the Pareto front; crossover mutation, exchanging path segments and randomly perturbing them; iterative optimization until convergence, thereby obtaining the optimal inspection path. The specific solution steps can refer to the conventional solution steps of the NSGA-II algorithm and will not be repeated here.

[0045] After generating the optimal inspection path, a control strategy for the inspection robot is generated based on the optimal inspection path and the device information table, so that the inspection robot inspects the power equipment in the device information table according to the optimal inspection path, thereby obtaining the device status data. The inspection robot transmits this device status data to an external server, where the device status data is analyzed to determine whether there are any abnormalities in the power equipment during the current inspection phase. When analyzing the device status data, a data analysis model based on a neural network model can be used, such as a sampling long short-term memory neural network or a deep convolutional neural network. Specifically, one or more corresponding neural network models can be selected for data analysis based on the type of device status data, and no further restrictions are imposed here.

[0046] In this embodiment, the digital twin map will be dynamically updated based on the collected real-time sensor data to ensure the timeliness of the digital twin map, so that the inspection path of each inspection can respond to the dynamic changes of the inspection environment in a timely manner, thereby improving the inspection efficiency of power equipment.

[0047] This embodiment provides an intelligent inspection method based on digital twins and a graph database. The present invention improves the accuracy of digital twin mapping through multimodal data fusion and coordinate calibration, improves the path planning efficiency and the safety of the inspection path through a path planning algorithm based on dynamic environment perception, and improves the accuracy of fault analysis through intelligent data analysis, thereby improving the operation and maintenance efficiency of power equipment and ensuring the safety and stability of power equipment operation.

[0048] See also Figure 2 Based on the same inventive concept, the second embodiment of the present invention proposes an intelligent inspection system based on digital twins and a graph database, including: The information query module 10 is used to perform semantic analysis on the inspection task book of the power equipment to obtain a structured task table, and input the equipment names in the structured task table into a preset graph database to obtain an equipment information table; The digital twin module 20 is used to construct an initial digital twin map of the inspection environment and calibrate the device point cloud coordinates in the initial digital twin map according to the device theoretical coordinates in the device information table to obtain a digital twin map; A checkpoint generation module 30 is configured to generate a reachable path based on the device coordinate set and the obstacle coordinate set in the digital twin map using a density clustering-based path planning algorithm, and perform checkpoint sampling on the reachable path to obtain a set of safe checkpoints; A path planning module 40 is configured to generate an optimal inspection path using a multi-objective path planning algorithm based on the set of safety inspection points and the set of obstacle coordinates; The intelligent inspection module 50 is used to control the inspection robot to perform equipment inspection according to the optimal inspection path and the equipment information table, obtain equipment status data, and determine whether there is any abnormality in the power equipment based on the equipment status data.

[0049] The technical features and technical effects of the intelligent inspection system based on digital twins and graph databases proposed in the embodiment of the present invention are the same as the method proposed in the embodiment of the present invention, and will not be described in detail here. The various modules in the above-mentioned intelligent inspection system based on digital twins and graph databases can be implemented in whole or in part by software, hardware and their combination. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0050] In summary, an embodiment of the present invention proposes an intelligent inspection method and system based on digital twins and a graph database. The method performs semantic parsing on the inspection task book of the power equipment to obtain a structured task table, and inputs the equipment name in the structured task table into a preset graph database to obtain an equipment information table; constructs an initial digital twin map of the inspection environment, and calibrates the equipment point cloud coordinates in the initial digital twin map according to the theoretical coordinates of the equipment in the equipment information table to obtain a digital twin map; based on the equipment coordinate set and the obstacle coordinate set in the digital twin map, a density clustering-based path planning algorithm is used to generate a reachable path, and inspection points are sampled on the reachable path to obtain a safe inspection point set; based on the safe inspection point set and the obstacle coordinate set, a multi-objective path planning algorithm is used to generate an optimal inspection path; based on the optimal inspection path and the equipment information table, the inspection robot is controlled to perform equipment inspection to obtain equipment status data, and based on the equipment status data, it is judged whether there is an abnormality in the power equipment. The present invention improves the accuracy of digital twin mapping through multimodal data fusion and coordinate calibration, improves the path planning efficiency and the safety of inspection paths through a path planning algorithm based on dynamic environment perception, and improves the accuracy of fault analysis through intelligent data analysis, thereby improving the operation and maintenance efficiency of power equipment and ensuring the safety and stability of power equipment operation.

[0051] Each embodiment in this specification is described in a progressive manner, and the same or similar parts of each embodiment can be directly referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. It should be noted that the various technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the various technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0052] The above-described embodiments merely represent several preferred implementations of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent. It should be noted that a person skilled in the art could make several improvements and substitutions without departing from the technical principles of the present invention, and these improvements and substitutions should also be considered within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be based on the scope of protection of the claims.

Claims

1. An intelligent inspection method based on digital twins and graph database, characterized in that: include: Perform semantic parsing on the inspection task book of the power equipment to obtain a structured task table, and input the equipment names in the structured task table into a preset graph database to obtain an equipment information table; Constructing an initial digital twin map of the inspection environment, and calibrating the device point cloud coordinates in the initial digital twin map according to the device theoretical coordinates in the device information table to obtain a digital twin map; According to the device coordinate set and the obstacle coordinate set in the digital twin map, a density clustering-based path planning algorithm is used to generate a reachable path, and inspection point sampling is performed on the reachable path to obtain a set of safety inspection points; Generate an optimal inspection path using a multi-objective path planning algorithm based on the set of safety inspection points and the set of obstacle coordinates; According to the optimal inspection path and the equipment information table, the inspection robot is controlled to perform equipment inspection, equipment status data is obtained, and based on the equipment status data, it is determined whether there is any abnormality in the power equipment.

2. The intelligent inspection method based on digital twins and graph database according to claim 1 is characterized in that: The steps of constructing an initial digital twin map of the inspection environment and calibrating the device point cloud coordinates in the initial digital twin map according to the device theoretical coordinates in the device information table to obtain the digital twin map include: Acquire sensor data of the inspection environment, and perform data fusion and filtering noise reduction on the sensor data to generate an initial digital twin map. The sensor data includes lidar point cloud data, infrared thermal imaging temperature data, and camera visual data. Using the least squares method to align the device point cloud data in the initial digital twin map with the device theoretical coordinates in the device information table and remove noise to obtain calibrated device point cloud data; The initial digital twin map is updated according to the calibrated device point cloud data to obtain a digital twin map.

3. The intelligent inspection method based on digital twin and graph database according to claim 1 is characterized in that: The steps of generating a reachable path based on the device coordinate set and the obstacle coordinate set in the digital twin map using a density clustering-based path planning algorithm, and sampling inspection points on the reachable path to obtain a set of safety inspection points include: Using a density clustering algorithm to perform cluster analysis on the device coordinate set in the digital twin map to obtain several device-dense areas; Generate a reachable path in each device-dense area using a path planning algorithm based on the device coordinate set and the obstacle coordinate set in the digital twin map; Inspection points are sampled on the reachable path at fixed intervals to obtain a set of safe inspection points.

4. The intelligent inspection method based on digital twins and graph database according to claim 3 is characterized in that: The step of using a density clustering algorithm to perform cluster analysis on the device coordinate set in the digital twin map to obtain several device-dense areas includes: According to the device coordinate set, based on the sliding window mechanism, the Euclidean distance between adjacent devices is calculated to obtain a local distance sequence corresponding to each sliding window; According to the local spacing sequence of each sliding window, the local average spacing and the local spacing standard deviation of each sliding window are calculated, and the global average spacing is calculated; Determine whether the ratio of the local spacing standard deviation to the global average spacing of each sliding window is less than a distance threshold; if so, use the sliding window as the first sliding window; otherwise, use the sliding window as the second sliding window; Obtaining, according to the global average spacing, the local average spacing, and a preset scaling factor, a first neighborhood radius of each power device in the first sliding window and a second neighborhood radius of each power device in the second sliding window; According to the first neighborhood radius and the second neighborhood radius, a density-based noise application spatial clustering algorithm is used to perform cluster analysis on the power equipment to obtain a plurality of equipment-intensive areas.

5. The intelligent inspection method based on digital twin and graph database according to claim 4 is characterized in that: The step of obtaining a first neighborhood radius of each power device in the first sliding window and a second neighborhood radius of each power device in the second sliding window according to the global average spacing, the local average spacing, and a preset scaling factor includes: Using the product of the global average distance and a preset scaling factor as the first neighborhood radius of each power device in the first sliding window; The product of the local average distance and the scaling factor is used as the second neighborhood radius of each power device in the second sliding window.

6. The intelligent inspection method based on digital twin and graph database according to claim 1 is characterized in that: The step of generating a reachable path in each device-dense area using a path planning algorithm based on the device coordinate set and the obstacle coordinate set in the digital twin map includes: Based on the device coordinate set and the obstacle coordinate set in the digital twin map, an improved fast search random tree algorithm is used to perform node search to obtain a reachable path in each device-dense area; The path cost function of the improved fast search random tree algorithm consists of the device path length and the obstacle risk value.

7. The intelligent inspection method based on digital twin and graph database according to claim 6 is characterized in that: The improved fast search random tree algorithm adopts a density-guided sampling strategy to perform random point sampling.

8. The intelligent inspection method based on digital twin and graph database according to claim 7 is characterized in that: The step of generating an optimal inspection path using a multi-objective path planning algorithm based on the set of safety inspection points and the set of obstacle coordinates includes: Calculating an obstacle risk value for each safety inspection point based on the safety inspection point set and the obstacle coordinate set; The objective function is to minimize the path length and the sum of the obstacle risk values ​​of each safety inspection point, and a multi-objective path planning algorithm is used to solve the problem to obtain the optimal inspection path.

9. The intelligent inspection method based on digital twin and graph database according to claim 8 is characterized in that: The multi-objective path planning algorithm is any one of a genetic algorithm, a particle swarm optimization algorithm and an ant colony algorithm.

10. An intelligent inspection system based on digital twins and graph database, characterized in that: include: An information query module is used to perform semantic analysis on the inspection task book of the power equipment to obtain a structured task table, and input the equipment names in the structured task table into a preset graph database to obtain an equipment information table; A digital twin module is used to construct an initial digital twin map of the inspection environment and calibrate the device point cloud coordinates in the initial digital twin map according to the device theoretical coordinates in the device information table to obtain a digital twin map; A checkpoint generation module is used to generate a reachable path based on the device coordinate set and the obstacle coordinate set in the digital twin map using a density clustering-based path planning algorithm, and to sample checkpoints on the reachable path to obtain a set of safe checkpoints; A path planning module, configured to generate an optimal inspection path using a multi-objective path planning algorithm based on the set of safety inspection points and the set of obstacle coordinates; The intelligent inspection module is used to control the inspection robot to perform equipment inspection according to the optimal inspection path and the equipment information table, obtain equipment status data, and determine whether there is any abnormality in the power equipment based on the equipment status data.

Images