Distributed power equipment cooperative maintenance path planning method and system

Abstract

The invention provides a distributed power equipment collaborative maintenance path planning method and system, relates to the technical field of power equipment maintenance, and constructs a multi-dimensional maintenance guide vector to guide a candidate maintenance robot network to carry out collaborative maintenance path planning by constructing a power fault node sensing network and carrying out maintenance guide feature analysis. Obtaining a maintenance path planning first space; loss prediction optimization is carried out according to the first space, and a maintenance path planning second space is obtained; and carrying out maintenance path optimization moderate maximization optimization on the second space, outputting a maintenance path optimization result, and controlling the candidate maintenance robot network to carry out maintenance. The technical problems that in the prior art, due to the fact that the prior art depends on static path planning or manual scheduling and lacks a collaborative scheduling mechanism based on global optimization, response of maintenance path planning is lagged, and resource allocation is unreasonable are solved, and the technical effect of improving the maintenance efficiency of the power equipment and the resource utilization rate is achieved.

Description

Technical Field

[0001] This application relates to the field of power equipment maintenance technology, specifically to a method and system for collaborative maintenance path planning of distributed power equipment. Background Technology

[0002] With the continuous development of new power systems and the widespread application of emerging technologies such as distributed generation and smart grids, the power grid structure is becoming increasingly complex, and equipment distribution is becoming more widespread and dispersed. Against this backdrop, intelligent maintenance of distributed power equipment has become a crucial link in ensuring the safe and stable operation of the power grid. Existing methods for distributed power equipment maintenance path planning mostly employ static path planning strategies or rely on manual scheduling rules for task allocation and path planning. While these methods are applicable to small-scale, relatively stable equipment networks, they are less effective when facing large-scale distributed power equipment clusters and sudden, multi-point fault scenarios. Because they neglect the coupling relationships between multiple factors such as overall task distribution, robot capability matching, and equipment priority, the path planning results are not optimal at the global level. This can easily lead to path failures or task delays, significantly reducing maintenance response speed. It can also cause problems such as multiple robots performing duplicate operations, path conflicts, and insufficient task coverage, affecting the rationality of resource allocation and hindering the reduction of system energy consumption and operation time costs. Summary of the Invention

[0003] This application provides a method and system for collaborative maintenance path planning of distributed power equipment, which solves the technical problems of existing technologies that rely heavily on static path planning or manual scheduling and lack a collaborative scheduling mechanism based on global optimization, resulting in delayed maintenance path planning response and unreasonable resource allocation. It achieves the technical effect of improving the maintenance efficiency and resource utilization of power equipment.

[0004] In view of the above problems, this application provides a method for collaborative maintenance path planning of distributed power equipment. The method includes: constructing a distributed power equipment network based on a set of distributed power equipment in a power grid; performing fault perception based on the distributed power equipment network to establish a power fault node perception network, and performing maintenance guidance feature analysis on the power fault node perception network to construct a multi-dimensional maintenance guidance vector; performing multi-dimensional perception analysis on a maintenance robot library to establish a candidate maintenance robot network; guiding the candidate maintenance robot network to perform collaborative maintenance path planning on the power fault node perception network based on the multi-dimensional maintenance guidance vector to obtain a first space for maintenance path planning; performing loss prediction optimization based on the first space for maintenance path planning to obtain a second space for maintenance path planning; performing maintenance path optimization to maximize the maintenance path optimality in the second space for maintenance path planning based on a maintenance path optimality analysis model, and outputting the maintenance path optimization result; and controlling the candidate maintenance robot network to perform maintenance on the power fault node perception network based on the maintenance path optimization result.

[0005] On the other hand, this application also provides a distributed power equipment collaborative maintenance path planning system, the system comprising: an equipment network construction module, used to construct a distributed power equipment network based on the distributed power equipment set of the power grid; a fault perception module, used to perform fault perception based on the distributed power equipment network, establish a power fault node perception network, and perform maintenance guidance feature analysis on the power fault node perception network to construct a multi-dimensional maintenance guidance vector; a maintenance network construction module, used to perform multi-dimensional perception analysis on a maintenance robot library to establish a candidate maintenance robot network; and a maintenance path planning module, used to plan the path based on the multi-dimensional maintenance... A guiding vector guides the candidate maintenance robot network to perform collaborative maintenance path planning on the power fault node perception network, obtaining a first space for maintenance path planning; a first optimization module is used to perform loss prediction optimization based on the first space for maintenance path planning, obtaining a second space for maintenance path planning; a second optimization module is used to perform maintenance path maximization optimization on the second space for maintenance path planning based on the maintenance path optima analytical model, and output the maintenance path optimization result; an execution module is used to control the candidate maintenance robot network to perform maintenance on the power fault node perception network based on the maintenance path optimization result.

[0006] One or more technical solutions provided in this application have at least the following beneficial effects: By constructing a distributed power equipment network based on the distributed power equipment set of the power grid, the connection relationships and topology between various distributed power equipment in the power grid can be clearly identified. This provides a comprehensive equipment relationship framework for subsequent operations such as fault detection, maintenance guidance, and robot path planning, enabling subsequent operations to proceed in an orderly manner within this framework. Fault detection based on the distributed power equipment network establishes a power fault node perception network, allowing for precise location of faulty nodes. Maintenance guidance feature analysis is performed on these nodes, and a multi-dimensional maintenance guidance vector is constructed. This vector describes the characteristics and maintenance needs of the faulty node from multiple dimensions, providing detailed guidance information for the maintenance robot's path planning. This allows the maintenance robot to plan its path to the faulty node more effectively. Multi-dimensional perception analysis of the maintenance robot library establishes a candidate maintenance robot network. This network comprehensively considers various factors such as the different characteristics, functions, and states of the robots, providing a suitable set of robot resource candidates for subsequent collaborative maintenance path planning, ensuring the selection of the most suitable robot for the maintenance task. The candidate maintenance robot network performs collaborative maintenance path planning on the power fault node perception network based on multi-dimensional maintenance guidance vectors. This achieves initial collaborative matching from robot capabilities to task objectives, forming a coarse-grained path solution and providing a foundation for subsequent optimization. By considering various potential losses during maintenance, the first space of maintenance path planning is optimized to obtain a second space, improving path quality while ensuring feasibility. The second space is then optimized using a maintenance path optimality analysis model to maximize the optimality of the maintenance path, evaluating its suitability from a more comprehensive perspective. This optimization selects the most suitable final path for the current scenario, ensuring maximum satisfaction of maintenance requirements. Based on the optimized path results, the candidate maintenance robot network is controlled to perform maintenance on the power fault node perception network along the optimal path, transforming the planning and optimization results into actual maintenance actions. This achieves closed-loop feedback from path planning to actual operation, improving maintenance response speed and execution accuracy.

[0007] In summary, this application constructs a distributed power equipment network and integrates power fault perception, multi-dimensional vector guidance, robot resource analysis, multi-stage path optimization, and optimal drive control to build a comprehensive distributed power equipment maintenance path planning system. This system enables maintenance robots to quickly and accurately reach fault nodes for maintenance, significantly improving the response efficiency of maintenance tasks, the rationality of path execution, and the comprehensive utilization rate of robot resources. It also reduces resource waste and time loss during the maintenance process, achieving efficient, intelligent, and collaborative maintenance of distributed power equipment in complex environments.

[0008] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0009] Figure 1 This is a flowchart illustrating the distributed power equipment collaborative maintenance path planning method provided in an embodiment of this application.

[0010] Figure 2 A flowchart illustrating the process of obtaining the first space for maintenance path planning in the distributed power equipment collaborative maintenance path planning method provided in the embodiments of this application.

[0011] Figure 3 This is a schematic diagram of the structure of the distributed power equipment collaborative maintenance path planning system provided in the embodiments of this application.

[0012] Explanation of reference numerals in the attached diagram: Equipment network construction module 10, fault perception module 20, maintenance network construction module 30, maintenance path planning module 40, first optimization module 50, second optimization module 60, execution module 70. Detailed Implementation

[0013] This application provides a method and system for collaborative maintenance path planning of distributed power equipment, which solves the technical problems of existing technologies that rely heavily on static path planning or manual scheduling and lack a collaborative scheduling mechanism based on global optimization, resulting in delayed maintenance path planning response and unreasonable resource allocation. This achieves the technical effect of improving the maintenance efficiency and resource utilization of power equipment.

[0014] Example 1, as Figure 1 As shown in the figure, this application provides a method for collaborative maintenance path planning of distributed power equipment, the method including: Step S1: Construct a distributed power equipment network based on the distributed power equipment set of the power grid.

[0015] Specifically, distributed power equipment refers to a collection of power infrastructure equipment, including power generation, transmission, distribution, and energy storage devices, deployed across various regions of the power system. Examples include distributed substations, wind turbines, solar photovoltaic arrays, and distribution switches. A unified model is created for the static information (equipment type, geographical coordinates, connectivity) and dynamic information (operating status, load capacity, communication parameters, etc.) of all distributed equipment in the power grid. A distributed power equipment network is then constructed using a graph structure, reflecting the connectivity, power supply paths, geographical distribution, and electrical dependencies between the distributed power equipment. For example, a graph database (such as Neo4j) or a network graph algorithm library (such as NetworkX) can be used to abstract each device as a node in the network, and the connectivity between devices as edges, thus constructing the distributed power equipment network.

[0016] Constructing a distributed power equipment network provides a basic framework for subsequent operations such as fault detection and maintenance planning, making the structure of the entire power system clear and facilitating subsequent operations on specific equipment.

[0017] Step S2: Based on the distributed power equipment network, perform fault perception, establish a power fault node perception network, and perform maintenance guidance feature analysis on the power fault node perception network to construct a multi-dimensional maintenance guidance vector.

[0018] Specifically, based on the existing distributed power equipment network, fault detection sensors (such as current sensors and voltage sensors) are used to monitor the operating status of the equipment in real time. Using sensor data and historical operating data, combined with deep learning models such as Gated Recurrent Neural Networks (GRUs), fault prediction is performed to identify currently faulty or high-risk equipment. These faulty or high-risk equipment are then designated as fault nodes, establishing a power fault node perception network. This network identifies all fault nodes and includes information on their related connected equipment. Then, each fault node is analyzed in detail. Feature engineering methods are used to perform maintenance guidance feature parsing based on the faulty equipment's historical maintenance records and current fault alarm information. Four dimensions of feature vectors are extracted: maintenance urgency (e.g., equipment popularity, risk level), topological dependence (whether the equipment is a hub node), maintenance value (the degree of impact on the overall system power supply), and location association (e.g., the geographical clustering of equipment). These vectors are then concatenated or combined to form the final multi-dimensional maintenance guidance vector. This multi-dimensional maintenance guidance vector is crucial input data for subsequent path planning, guiding the maintenance robot's path planning.

[0019] By extracting multidimensional features from faulty nodes, the goal guidance of path planning and the pertinence of task allocation are effectively improved, enhancing the understanding and response capabilities to complex power grid situations.

[0020] Step S3: Perform multi-dimensional perception analysis on the maintenance robot library and establish a candidate maintenance robot network.

[0021] Specifically, the maintenance robot library is a collection of data related to all schedulable robots during the maintenance of power equipment. The data for each robot in the library is traversed and analyzed to extract current state information, such as the executable time period, geographical location, and energy consumption status, yielding multi-dimensional robot data. A rule engine or fuzzy logic decision model is used to filter this multi-dimensional robot data, selecting robots whose functions and states match the current maintenance task and scenario based on three dimensions: time margin (e.g., whether the task can be completed within the allotted time), robot state (e.g., health status, repair capability), and feature adaptation (e.g., whether the carried tools are suitable for the fault type). A candidate maintenance robot network is then constructed using a graph structure based on these robots and their possible collaborative relationships.

[0022] By conducting multi-dimensional perception analysis on the maintenance robot library, we can comprehensively understand the robot resource situation and establish a candidate maintenance robot network. This enables efficient and refined robot resource screening, avoids blind scheduling and resource waste, and ensures that the scheduling plan is feasible and adaptable.

[0023] Step S4: Based on the multi-dimensional maintenance guidance vector, guide the candidate maintenance robot network to perform collaborative maintenance path planning on the power fault node perception network to obtain the first space of maintenance path planning.

[0024] Specifically, the candidate maintenance robot network includes information such as robots suitable for participating in maintenance tasks and their collaborative relationships. Based on the multi-dimensional maintenance guidance vector obtained in step S2 and the candidate maintenance robot network constructed in step S3, path planning algorithms (such as A-star algorithm, Dijkstra algorithm, etc.) are used to plan a path from its current position to the fault node for each robot, taking into account the collaborative relationships between multiple robots (such as avoiding collisions between robots, and rationally allocating tasks to improve efficiency). These path schemes together constitute the first space of maintenance path planning, providing rich combinations of candidate paths for subsequent optimization, improving the diversity of scheduling results and global optimization potential.

[0025] Step S5: Based on the first space of the maintenance path planning, perform loss prediction optimization to obtain the second space of the maintenance path planning.

[0026] Specifically, in the simulation environment, each path in the first space of maintenance path planning is executed, obtaining simulated maintenance data for each path. This simulated maintenance data is then input into a pre-built maintenance loss assessment model to evaluate the potential risks or system burdens of executing each path, such as energy consumption, time consumption, and task interruption rate. Based on these path assessment results, paths that meet the expected loss requirements are selected, and these paths collectively form the second space of maintenance path planning. Through loss prediction optimization, high-risk, unreliable, or excessively costly solutions are eliminated, reducing various losses during the maintenance process and making maintenance path planning more economical and efficient.

[0027] Step S6: Based on the maintenance path optimization analysis model, perform maintenance path optimization maximization search in the second space of the maintenance path planning, and output the maintenance path optimization result.

[0028] Specifically, the maintenance path optimization and suitability analytical model is a functional model that comprehensively evaluates maintenance efficiency and path loss, used to measure the merits of maintenance paths. All path schemes in the second space of maintenance path planning are input one by one into the model for optimization and suitability evaluation. Based on the evaluation results of each path, optimization algorithms (such as linear programming algorithms) are used to maximize the optimization of the path schemes in the second space of maintenance path planning, finding the optimal maintenance path scheme (i.e., the one with the highest optimization and suitability) under the comprehensive evaluation, and the output is the maintenance path optimization result.

[0029] By comprehensively evaluating the advantages and disadvantages of maintenance paths through a maintenance path optimization analysis model, and combining it with maximization optimization to obtain the optimal maintenance path scheme, a dynamic trade-off between efficiency and stability is achieved. The final output path scheme has higher cost performance and task execution controllability.

[0030] Step S7: Based on the maintenance path optimization results, control the candidate maintenance robot network to perform maintenance on the power fault node perception network.

[0031] Specifically, based on the maintenance path optimization results obtained in step S6, control commands are sent to each robot in the candidate maintenance robot network through a scheduling system (such as the ROS robot operating system or SCADA interface). These control commands include information such as the robot's specific action path, target equipment, and time window, enabling the robot to perform maintenance on the faulty node in the power fault node perception network according to the optimal path. After receiving the commands, the robot proceeds to the faulty node according to the prescribed path and performs maintenance operations, realizing an automatic execution mechanism from planning to control, improving the automation level of the maintenance process, and reducing manual intervention. At the same time, the robot's local task scheduling module is activated to provide real-time feedback on the execution status to the main control system during the maintenance process, forming a closed-loop control.

[0032] Furthermore, in step S2 of this embodiment, the maintenance guidance feature is analyzed on the power fault node perception network to construct a multi-dimensional maintenance guidance vector, including: Step S24: Evaluate the maintenance urgency of the power fault node sensing network and establish a maintenance urgency guidance vector.

[0033] Step S25: Perform topology dependency analysis on the power fault node sensing network and establish a topology dependency maintenance guidance vector.

[0034] Step S26: Evaluate the maintenance value of the power fault node sensing network and establish a maintenance value guiding vector.

[0035] Step S27: Perform location association analysis based on the power fault node perception network to establish a location association maintenance guidance vector.

[0036] Step S28: Output the maintenance urgency guidance vector, the topology-dependent maintenance guidance vector, the maintenance value guidance vector, and the location-related maintenance guidance vector as the multidimensional maintenance guidance vector.

[0037] Specifically, node data for each node in the power fault node perception network is collected, including load level, impact radius, and historical fault frequency. A fault impact analysis model is used to assess the impact range and cascading reaction probability of the fault based on factors such as the power grid topology and power load distribution. An urgency score for each fault node is calculated. Then, a standardization method (such as Z-score or Min-Max normalization) is used to standardize the urgency score and arrange them according to the fault node order to form a maintenance urgency guidance vector. This vector is used to guide the priority of maintenance work and instruct the robot to respond to high-urgency nodes first.

[0038] A detailed analysis of the topology of the power fault node sensing network is performed using topology analysis tools (such as graph theory-based topology analysis software). The degree centrality and betweenness centrality indices of each fault node are calculated. These two indices are then weighted according to custom weights to obtain the topology dependency analysis results. The topology dependency analysis results of all fault nodes are then constructed into a topology dependency guiding vector. This topology dependency guiding vector reflects the degree of influence of fault nodes on the stability of the network structure.

[0039] Maintenance value refers to the benefits derived from maintaining a specific node, including economic efficiency and contribution to system recovery. Using a cost-benefit analysis model, the maintenance value of each fault node is assessed by considering factors such as equipment cost, restored power capacity, and the cost of avoiding cascading failures. The assessment results are then converted into numerical values ​​or ranges, and the maintenance values ​​of all fault nodes are combined to form a maintenance value guiding vector. This vector contains relevant information about the maintenance value of each fault node, enabling the rational allocation of maintenance resources to more valuable fault nodes when resources are limited.

[0040] Location correlation refers to the spatial proximity of different faulty nodes. Based on the geographical coordinates or relative topological coordinates of the equipment corresponding to each faulty node within the power grid, Euclidean distance or map distance calculation methods are used to determine the node distances between each node. Based on these node distances, K-means clustering or DBSCAN clustering is used to cluster the locations of each faulty node, grouping geographically close equipment into the same maintenance task group for easier unified scheduling and path merging. Based on the clustering results, a location label is assigned to each node, and the average distance from each node to other nodes within its task group is calculated as the node's location correlation degree. The location correlation degrees of all nodes are normalized and arranged into a vector, serving as the location correlation maintenance guidance vector to guide the robot to perform nearby merged maintenance.

[0041] The aforementioned maintenance urgency guidance vector, topology-dependent maintenance guidance vector, maintenance value guidance vector, and location-related maintenance guidance vector are combined in sequence to form a multi-dimensional maintenance guidance vector. This enables a comprehensive quantitative assessment of fault nodes across multiple dimensions, providing a clear guiding foundation for robot collaborative path planning and making maintenance work more scientific, efficient, and reasonable.

[0042] Furthermore, such as Figure 2 As shown, step S4 in this embodiment includes: Step S41: Based on each guiding vector in the multi-dimensional maintenance guidance vector, guide the candidate maintenance robot network to perform collaborative maintenance path planning on the power fault node perception network, and obtain the first domain of collaborative maintenance path planning.

[0043] Step S42: Randomly combine and fuse the multidimensional maintenance guidance vectors to obtain multiple fused maintenance guidance vectors.

[0044] Step S43: Based on the multiple fused maintenance guidance vectors, guide the candidate maintenance robot network to perform collaborative maintenance path planning on the power fault node perception network, and obtain the second domain of collaborative maintenance path planning.

[0045] Step S44: Merge the first domain of the collaborative maintenance path planning and the second domain of the collaborative maintenance path planning to generate the first space of the maintenance path planning.

[0046] Specifically, for each guidance vector within the multi-dimensional maintenance guidance vector (maintenance urgency guidance vector, topology-dependent maintenance guidance vector, maintenance value guidance vector, and location-related maintenance guidance vector), this dimension is used as a priority index for path guidance. This guides the candidate maintenance robot network to perform collaborative maintenance path planning on the power fault node perception network, obtaining a set of paths generated independently by each guidance vector. These path sets collectively constitute the first domain of collaborative maintenance path planning. Specifically, for maintenance urgency guidance vectors, nodes with high urgency are prioritized for scheduling. For topology-dependent maintenance guidance vectors, paths are planned based on the topology dependencies of fault nodes, ensuring that other related nodes are considered while repairing the fault node. For maintenance value guidance vectors, fault nodes with high maintenance value are traversed first. For location-related maintenance guidance vectors, paths are planned based on the location relationships of fault nodes, such as prioritizing fault nodes that are close or in the same area for maintenance. By using path planning algorithms (such as A-star algorithm, Dijkstra algorithm, etc.) to plan paths for the robots based on guidance from each guidance vector, the first domain of collaborative maintenance path planning is obtained.

[0047] Multiple vectors are randomly weighted and combined to form multiple fused maintenance guidance vectors by different weighting methods such as balanced fusion or biased fusion. The combination scheme can be a random combination of two, three, or four vectors. Balanced fusion means combining each dimension of the guidance vector with equal weight, while biased fusion means setting a specific weight for each dimension of the guidance vector.

[0048] For each fusion guiding vector, it is used as a task priority guide. Path planning algorithms (such as A-star algorithm, Dijkstra algorithm, etc.) are used to perform collaborative maintenance path planning on the power fault node perception network to obtain a set of paths corresponding to each fusion guiding vector. These path sets together form the second domain of collaborative maintenance path planning.

[0049] By merging all path schemes in the first domain and all path schemes in the second domain of collaborative maintenance path planning, a complete first space for maintenance path planning is formed, resulting in more comprehensive preliminary planning results for maintenance paths. This provides richer basic path schemes for subsequent loss prediction and optimization. This process can be easily achieved by combining path schemes from the two sets into a new set.

[0050] Furthermore, step S5 in this embodiment includes: Step S51: Plan the first space according to the maintenance path and extract the first maintenance path.

[0051] Step S52: Perform loss prediction on the first maintenance path and determine the loss prediction coefficient for the first path.

[0052] Step S53: Determine whether the first path loss prediction coefficient is less than the path loss threshold.

[0053] Step S54: If the first path loss prediction coefficient is less than the path loss threshold, add the first maintenance path to the second space of the maintenance path planning.

[0054] Step S55: If the first path loss prediction coefficient is greater than or equal to the path loss threshold, the first maintenance path is eliminated.

[0055] Step S56: Based on the path loss threshold, continue to perform loss prediction and optimization on the first space of the maintenance path planning to generate the second space of the maintenance path planning.

[0056] Specifically, a planned path is extracted from the first space of the maintenance path planning and used as the first maintenance path. The extraction process can be carried out by selecting paths sequentially according to their storage order, or by selecting them randomly.

[0057] A maintenance loss assessment model is pre-built. Multiple rounds of simulation are performed on the first maintenance path, collecting simulation data such as energy consumption, latency, and task failure rate. This simulation data is then input into the maintenance loss assessment model, which outputs a first-path loss prediction coefficient. This loss prediction coefficient measures the combined losses that may occur during the maintenance process of this path; a higher value indicates a greater loss for the corresponding path.

[0058] The path loss threshold is a pre-set value used as a standard to determine whether a maintenance path is acceptable. The calculated first path loss prediction coefficient is compared with the pre-set path loss threshold to determine whether the first path loss prediction coefficient is less than the path loss threshold.

[0059] When the predicted coefficient of the first path loss is less than the path loss threshold, it means that the first maintenance path is acceptable in terms of loss. At this time, the first maintenance path is directly added to the second space of maintenance path planning.

[0060] When the predicted loss coefficient of the first path is greater than or equal to the path loss threshold, it means that the loss of the first maintenance path is too large and does not meet the requirements. In this case, the first maintenance path should be eliminated from consideration, such as marking the path as eliminated in the first space of maintenance path planning or deleting its related data directly from the data structure of the first space.

[0061] Following the aforementioned path loss prediction and threshold screening method, the other maintenance paths in the first maintenance path planning space are iteratively traversed, continuously extracting the next maintenance path (such as the second maintenance path, the third maintenance path, etc.), calculating its path loss prediction coefficient, and comparing it with the path loss threshold. Based on the comparison result, it is decided whether to add it to the second maintenance path planning space or discard it. By processing all paths in the first maintenance path planning space, the second maintenance path planning space is finally generated.

[0062] Through the above steps, paths with high losses are removed from the first space of maintenance path planning, avoiding excessive energy consumption or failure rate in actual scheduling, and providing a higher quality set of candidate paths for the next stage of optimal optimization.

[0063] Furthermore, step S52 includes: Step S521: Perform multiple simulated maintenance operations according to the first maintenance path to obtain multiple simulated maintenance datasets.

[0064] Step S522: Input the multiple simulated maintenance datasets into the maintenance loss assessment model to obtain multiple maintenance loss assessment coefficients.

[0065] Step S523: Calculate the mean of the multiple maintenance loss assessment coefficients to generate the first path loss prediction coefficient.

[0066] Specifically, simulation software such as MATLAB Simulink and AnyLogic are used to conduct multiple simulated maintenance operations according to the robot's movement route and operation content specified in the first maintenance path. Data from each maintenance process is recorded, such as the time it takes for the robot to reach each fault node from the starting point, the energy consumption in each operation, and the wear values ​​of key robot components, to obtain a simulated maintenance dataset.

[0067] The maintenance loss assessment model is a pre-built model that can be constructed based on methods such as random forests, neural networks, or weighted scoring functions. It can quantify and assess various losses during the maintenance process based on the input simulated maintenance dataset, outputting a comprehensive maintenance loss assessment coefficient. For example, a weighted scoring function can be designed, such as Loss = ω1*T + ω2*E + ω3*F, where Loss is the maintenance loss, T is time consumption, E is energy consumption, F is the number of intermediate failures, and ω1, ω2, and ω3 are the weights of time consumption, energy consumption, and the number of intermediate failures, respectively, reflecting their degree of influence on the maintenance loss, and are determined based on expert experience.

[0068] Multiple sets of simulated maintenance data corresponding to the first maintenance path are input one by one into the maintenance loss assessment model. The model will output the maintenance loss assessment coefficient corresponding to each set of simulated data according to its internal calculation rules. The average value of the multiple maintenance loss assessment coefficients corresponding to the first maintenance path is calculated as the first path loss prediction coefficient.

[0069] Through the above steps, risk prediction and quantification of path execution results are effectively achieved, providing data support for subsequent path selection. Multiple simulations reduce the risk of misjudgment caused by occasional anomalies, ensuring that the path input to the next stage has high stability and execution reliability.

[0070] Furthermore, the maintenance path optimization analytical model includes a maintenance path optimization analytical function, which is: Wherein, MPO represents the optimality of the maintenance path, exp represents the exponential function with the natural constant e as the base, PEW represents the predetermined weight of maintenance efficiency, g(MPE) represents the normalized predicted maintenance efficiency of the path, PLW represents the predetermined weight of path loss, and g(MPL) represents the normalized predicted coefficient of path loss.

[0071] Specifically, the maintenance path optimality analytical function is constructed based on an exponential weighted function. This model comprehensively considers two key dimensions: path efficiency and path loss, setting efficiency weights (PEW) and loss weights (PLW) to balance their impact on optimality. The model first normalizes the predicted path maintenance efficiency (MPE) and the predicted path loss coefficient (MPL) to ensure comparability between the two types of data. Then, it calculates the maintenance path optimality (MPO) using an exponential function. The predicted path maintenance efficiency (MPE) measures the efficiency of a robot's execution path in completing a maintenance task. A higher value indicates more reasonable robot allocation and scheduling, faster task completion, and more optimized resource utilization under that path; it can be calculated based on the aforementioned simulated maintenance dataset. For example, , among which, T avg T represents the average completion time in multiple rounds of simulated maintenance. max R represents the maximum acceptable completion time in historical maintenance tasks, serving as a normalization reference. s T represents the success rate in simulated maintenance (i.e., the number of times the task is completed divided by the number of simulations). wait The average collaborative waiting time is represented by ω1, ω2, and ω3, which are the idle waiting time in the multi-robot interactive scheduling. These are weighting coefficients that can be set according to the actual scheduling objectives.

[0072] This analytical function for determining the optimal maintenance path possesses the ability to adjust its nonlinear sensitivity, enabling it to significantly distinguish the merits of different maintenance paths while maintaining computational efficiency. Based on the calculated optimality of the maintenance path, the path with the highest optimality is selected as the final maintenance path scheme, achieving precise decision-making in path optimization.

[0073] Furthermore, step S3 in this embodiment includes: Step S31: Perform time margin perception optimization based on the maintenance robot library to obtain the first candidate space of maintenance robots that meet the predetermined time margin.

[0074] Step S32: Perform state perception optimization based on the first candidate space of the maintenance robot to obtain the second candidate space of the maintenance robot.

[0075] Step S33: Perform basic feature perception based on the second candidate space of the maintenance robot, and build the candidate maintenance robot network.

[0076] Specifically, time margin refers to the time available for a maintenance robot to perform a new task after completing its current task in a maintenance task scheduling process. The current task and estimated completion time of each robot in the maintenance robot library are collected, and a path planning algorithm is used to calculate the shortest time path from each robot to the target task point (i.e., the fault node). Let T be the required completion time for the task at the target task point. deadline The time required for the robot to travel from its current location to the maintenance point is T. travel The time required to complete the current maintenance task is T. maint Then the total time T total =T travel +T maint Determine the total time T total Is it less than or equal to T? deadline , will satisfy T total ≤T deadline The selected robots are added to the first candidate space of maintenance robots, thus pre-screening the robots under the task time constraint. The first candidate space of maintenance robots contains all maintenance robots in the maintenance robot library that meet the predetermined time margin requirements.

[0077] For each maintenance robot in the first candidate space, status detection is performed. The status information of each robot's sensors and its operation logs are read. Using a status monitoring model or preset status judgment rules, the system identifies whether the robot exhibits operational abnormalities, such as communication interruptions, insufficient power, mechanical failures, or motion deviations. For example, if preset status judgment rules are used, abnormalities can be defined as: power level below 30%, abnormal temperature, and intermittent communication frequency. If a status monitoring model is used, a supervised learning model can be trained based on historical operation data to distinguish between "normal" and "abnormal" labels. If an abnormality is detected in a robot, it is removed from the first candidate space. The remaining robots form the second candidate space, preventing unstable robots from being assigned to critical maintenance tasks and avoiding mid-task failures.

[0078] For each maintenance robot in the second candidate space, its basic characteristic information, including model, specifications, real-time location, and software / hardware compatibility, is extracted by querying the robot's equipment information database. Then, based on this basic characteristic information, the robots are connected and organized to build a candidate maintenance robot network. For example, robots can be classified according to their model, as robots of the same model have similar maintenance capabilities. Combining this with their real-time location, the roles and relationships of different models of robots at different locations within the network are determined, thus constructing a complete candidate maintenance robot network.

[0079] The above steps employ a triple screening mechanism of time margin perception, state anomaly detection, and basic capability matching to accurately construct a candidate maintenance robot network from the robot library that is currently suitable for fault repair tasks. This effectively ensures path availability and maintenance execution success rate, achieving optimal utilization of scheduling resources and a comprehensive improvement in maintenance efficiency.

[0080] Furthermore, in step S2 of this embodiment, fault perception is performed based on the distributed power equipment network to establish a power fault node perception network, including: Step S21: Perform fault prediction based on the distributed power equipment network and determine the fault prediction coefficient for each device.

[0081] Step S22: Determine whether the fault prediction coefficient of each device is greater than or equal to the fault threshold of each device, and obtain the fault judgment result of each device.

[0082] Step S23: Based on the fault judgment results of each device, sort out the fault nodes of the distributed power equipment network and generate the power fault node perception network.

[0083] Specifically, operational data from each device in the distributed power grid network is collected, including historical and real-time operational data. Supervised learning is then applied to this operational data to train a fault prediction model. Based on the operational data, the fault prediction model can predict fault prediction coefficients for the power devices. These coefficients are a quantitative representation of the probability of each device in the distributed power grid network failing; the higher the coefficient, the greater the probability of the corresponding power device failing.

[0084] Equipment fault thresholds are pre-set for each power device to determine whether the device is in a high-risk fault state. The fault prediction coefficient for each device is compared with its corresponding fault threshold. If the fault prediction coefficient is greater than or equal to the corresponding fault threshold, the fault judgment result is "equipment fault"; if the fault prediction coefficient is less than the corresponding fault threshold, the fault judgment result is "equipment normal".

[0085] Based on the fault judgment results of each device, in the distributed power equipment network, those devices whose fault judgment results are "device fault" (i.e., the equipment fault prediction coefficient is greater than or equal to the equipment fault threshold) are screened out and marked as power fault nodes. These power fault nodes are then combined according to the device connection relationship to construct a power fault node perception network.

[0086] The above steps, through fault prediction, threshold judgment, and fault node identification, establish a power fault node perception network, which can monitor and identify faulty equipment in the network in real time. This provides clear objectives and a foundation for subsequent maintenance path planning, significantly improving the efficiency and pertinence of maintenance work.

[0087] Furthermore, step S21 includes: Step S211: Perform real-time monitoring of the distributed power equipment network to obtain monitoring data for each power equipment.

[0088] Step S212: Based on the monitoring records and fault detection records of each power device in the distributed power device network, supervised learning is performed on the gated loop unit to generate fault prediction models for each power device.

[0089] Step S213: Input the monitoring data of each power equipment into the fault prediction model of each power equipment, and output the fault prediction coefficient of each equipment.

[0090] Specifically, a sensor detection network is established for various types of power equipment (such as distributed transformers, circuit breakers, busbars, cable nodes, etc.) to collect real-time operating data of each power equipment, including but not limited to current, voltage, temperature, operating load, communication signal quality, packet loss rate, etc., to obtain monitoring data of each power equipment.

[0091] The system extracts monitoring and fault detection records from the operation and maintenance records of each power equipment. This includes monitoring data for each equipment over a past period, details of past faults, fault types, and occurrence times. These records are then organized and preprocessed to suit as input data for a gated recurrent unit (GRU). For example, the time-series data in the monitoring records is divided into time intervals, and fault detection records are converted into corresponding fault labels (e.g., 0 for no fault, 1 for a fault). Using the preprocessed monitoring records as input and the fault detection records as output labels, a GRU neural network is used for supervised learning of both records. During the learning process, the parameters of the GRU (such as weights and biases) are adjusted to ensure the model's predicted output closely approximates the actual fault detection records. After multiple iterations of training, when the model's accuracy reaches the expected standard, the fault prediction model for each power equipment is output.

[0092] The obtained monitoring data of each power equipment is input into the corresponding power equipment fault prediction model. The neurons inside the model will calculate based on the pre-learned weights and biases, and output a value representing the probability of the equipment failure, namely the equipment failure prediction coefficient.

[0093] The above steps, through real-time monitoring, GRU model training and prediction, can accurately capture the characteristic patterns of equipment failures, realize the failure prediction of distributed power equipment, and provide a scientific basis for subsequent failure handling.

[0094] In summary, the distributed power equipment collaborative maintenance path planning method provided in this application has the following beneficial effects: By constructing a distributed power equipment network, a system modeling and structured representation of distributed equipment in the power grid is achieved, providing a networked framework for subsequent fault identification and path planning. This serves as the starting point and data foundation for the entire process. Fault perception is performed based on the distributed power equipment network, establishing a power fault node perception network. Maintenance guidance feature analysis is then performed on this network to construct a multi-dimensional maintenance guidance vector. This enables intelligent identification and multi-dimensional feature extraction of power equipment faults, transforming fault nodes into vectorized expressions that can be used to guide path planning, providing guidance for subsequent decision-making. Multi-dimensional perception analysis is performed on the maintenance robot library to establish a schedulable robot resource candidate set, ensuring the accuracy of task allocation and path matching. Based on the multi-dimensional maintenance guidance vector, the candidate maintenance robot network is guided to perform collaborative maintenance path planning with the power fault node perception network, obtaining a first space for maintenance path planning. This achieves initial collaborative matching from robot capabilities to task objectives, forming a coarse-grained path solution. Loss prediction and optimization are then performed based on the first space for maintenance path planning, further optimizing the initial path to obtain a second space for maintenance path planning. This improves path quality while ensuring feasibility. A optimum-appropriate evaluation model is introduced to further filter the second space of maintenance path planning. Based on overall efficiency, the most suitable final path for the current scenario is selected to ensure optimal execution results. The optimization results are then used for robot control execution, achieving closed-loop feedback from path planning to actual operation, thereby improving maintenance response speed and execution accuracy.

[0095] Overall, the embodiments of this application significantly improve the response efficiency of maintenance tasks, the rationality of path execution, and the comprehensive utilization rate of robot resources through multi-dimensional information fusion and adaptive path optimization, realizing efficient, intelligent, and collaborative maintenance of distributed power equipment in complex environments.

[0096] Example 2, as Figure 3 As shown, based on the same inventive concept as in Embodiment 1 above, this application provides a distributed power equipment collaborative maintenance path planning system, the system comprising: The device network construction module 10 is used to construct a distributed power device network based on the distributed power device set of the power grid.

[0097] The fault perception module 20 is used to perceive faults based on the distributed power equipment network, establish a power fault node perception network, and perform maintenance guidance feature analysis on the power fault node perception network to construct a multi-dimensional maintenance guidance vector.

[0098] The maintenance network construction module 30 is used to perform multi-dimensional perception analysis on the maintenance robot library and establish a candidate maintenance robot network.

[0099] The maintenance path planning module 40 is used to guide the candidate maintenance robot network to perform collaborative maintenance path planning on the power fault node perception network according to the multi-dimensional maintenance guidance vector, so as to obtain the first space of maintenance path planning.

[0100] The first optimization module 50 is used to perform loss prediction optimization based on the first space of the maintenance path planning to obtain the second space of the maintenance path planning.

[0101] The second optimization module 60 is used to maximize the optimization of the maintenance path in the second space of the maintenance path planning according to the maintenance path optimization and appropriateness analytical model, and output the maintenance path optimization result.

[0102] The execution module 70 is used to control the candidate maintenance robot network to perform maintenance on the power fault node perception network based on the maintenance path optimization result.

[0103] Furthermore, in this embodiment of the application, the fault perception module 20 is also used to perform the following steps: The maintenance urgency of the power fault node sensing network is evaluated, and a maintenance urgency guidance vector is established. Topology dependency analysis is performed on the power fault node sensing network, and a topology dependency maintenance guidance vector is established. Maintenance value is evaluated on the power fault node sensing network, and a maintenance value guidance vector is established. Location association analysis is performed on the power fault node sensing network, and a location association maintenance guidance vector is established. The maintenance urgency guidance vector, the topology dependency maintenance guidance vector, the maintenance value guidance vector, and the location association maintenance guidance vector are output as the multidimensional maintenance guidance vector.

[0104] Furthermore, in this embodiment of the application, the maintenance path planning module 40 is also used to perform the following steps: Based on each guiding vector within the multidimensional maintenance guidance vector, the candidate maintenance robot network is guided to perform collaborative maintenance path planning on the power fault node perception network, obtaining a first domain of collaborative maintenance path planning; the multidimensional maintenance guidance vector is randomly combined and fused to obtain multiple fused maintenance guidance vectors; based on the multiple fused maintenance guidance vectors, the candidate maintenance robot network is guided to perform collaborative maintenance path planning on the power fault node perception network, obtaining a second domain of collaborative maintenance path planning; the first domain of collaborative maintenance path planning and the second domain of collaborative maintenance path planning are merged to generate a first space of maintenance path planning.

[0105] Furthermore, in this embodiment of the application, the first optimization module 50 is also used to perform the following steps: Based on the maintenance path planning first space, a first maintenance path is extracted; loss prediction is performed on the first maintenance path to determine the first path loss prediction coefficient; it is determined whether the first path loss prediction coefficient is less than the path loss threshold; if the first path loss prediction coefficient is less than the path loss threshold, the first maintenance path is added to the maintenance path planning second space; if the first path loss prediction coefficient is greater than or equal to the path loss threshold, the first maintenance path is eliminated; based on the path loss threshold, loss prediction optimization is continued in the maintenance path planning first space to generate the maintenance path planning second space.

[0106] Furthermore, in this embodiment of the application, the first optimization module 50 is also used to perform the following steps: Multiple simulated maintenance operations are performed based on the first maintenance path to obtain multiple simulated maintenance datasets; the multiple simulated maintenance datasets are input into the maintenance loss assessment model to obtain multiple maintenance loss assessment coefficients; the average of the multiple maintenance loss assessment coefficients is calculated to generate the first path loss prediction coefficient.

[0107] Furthermore, the maintenance path optimization analytical model includes a maintenance path optimization analytical function, which is: Wherein, MPO represents the optimality of the maintenance path, exp represents the exponential function with the natural constant e as the base, PEW represents the predetermined weight of maintenance efficiency, g(MPE) represents the normalized predicted maintenance efficiency of the path, PLW represents the predetermined weight of path loss, and g(MPL) represents the normalized predicted coefficient of path loss.

[0108] Furthermore, in this embodiment of the application, the maintenance network construction module 30 is also used to perform the following steps: Based on the maintenance robot library, time margin perception optimization is performed to obtain a first candidate space of maintenance robots that meet the predetermined time margin; based on the first candidate space of maintenance robots, state perception optimization is performed to obtain a second candidate space of maintenance robots; based on the second candidate space of maintenance robots, basic feature perception is performed to build the candidate maintenance robot network.

[0109] Furthermore, in this embodiment of the application, the fault perception module 20 is also used to perform the following steps: Fault prediction is performed based on the distributed power equipment network to determine the fault prediction coefficient of each device; it is determined whether the fault prediction coefficient of each device is greater than or equal to the fault threshold of each device to obtain the fault judgment result of each device; based on the fault judgment result of each device, the fault nodes of the distributed power equipment network are sorted out to generate the power fault node perception network.

[0110] Furthermore, in this embodiment of the application, the fault perception module 20 is also used to perform the following steps: The distributed power equipment network is monitored in real time to obtain monitoring data of each power equipment; the gated loop unit is supervised learning based on the monitoring records and fault detection records of each power equipment in the distributed power equipment network to generate fault prediction models for each power equipment; the monitoring data of each power equipment is input into the fault prediction models of each power equipment, and the fault prediction coefficients of each equipment are output.

[0111] Through the foregoing detailed description of the distributed power equipment collaborative maintenance path planning method, those skilled in the art can clearly understand that the distributed power equipment collaborative maintenance path planning system in this embodiment, as opposed to the system disclosed in Embodiment 2, has corresponding functional modules and beneficial effects as it corresponds to the method disclosed in Embodiment 1. For relevant details, please refer to the method section.

[0112] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for collaborative maintenance path planning of distributed power equipment, characterized in that, include: Based on the distributed power equipment set of the power grid, construct a distributed power equipment network; Based on the distributed power equipment network, fault perception is performed, a power fault node perception network is established, and maintenance guidance feature analysis is performed on the power fault node perception network to construct a multi-dimensional maintenance guidance vector. Perform multi-dimensional perception analysis on the maintenance robot library to establish a candidate maintenance robot network; Based on the multi-dimensional maintenance guidance vector, the candidate maintenance robot network is guided to perform collaborative maintenance path planning on the power fault node perception network to obtain the first space of maintenance path planning. Based on the first space of the maintenance path planning, loss prediction and optimization are performed to obtain the second space of the maintenance path planning. Based on the maintenance path optimization and optimization analytical model, the second space of the maintenance path planning is optimized to maximize the maintenance path optimization and optimization, and the maintenance path optimization result is output. Based on the maintenance path optimization results, the candidate maintenance robot network is controlled to perform maintenance on the power fault node perception network.

2. The method for collaborative maintenance path planning of distributed power equipment as described in claim 1, characterized in that, The maintenance guidance features of the power fault node perception network are analyzed to construct a multi-dimensional maintenance guidance vector, including: The maintenance urgency of the power fault node sensing network is evaluated, and a maintenance urgency guidance vector is established. Perform topology dependency analysis on the power fault node sensing network and establish a topology dependency maintenance guidance vector; The maintenance value of the power fault node sensing network is evaluated, and a maintenance value guidance vector is established. Based on the power fault node perception network, a location correlation analysis is performed to establish a location correlation maintenance guidance vector. The maintenance urgency guidance vector, the topology-dependent maintenance guidance vector, the maintenance value guidance vector, and the location-related maintenance guidance vector are output as the multidimensional maintenance guidance vector.

3. The distributed power equipment collaborative maintenance path planning method as described in claim 1, characterized in that, Based on the multi-dimensional maintenance guidance vector, the candidate maintenance robot network is guided to perform collaborative maintenance path planning on the power fault node perception network, obtaining a first space for maintenance path planning, including: Based on each guiding vector in the multidimensional maintenance guiding vector, the candidate maintenance robot network is guided to perform collaborative maintenance path planning on the power fault node perception network to obtain the first domain of collaborative maintenance path planning. Multiple fused maintenance guidance vectors are obtained by randomly combining and fusing the multidimensional maintenance guidance vectors. Based on the multiple fused maintenance guidance vectors, the candidate maintenance robot network is guided to perform collaborative maintenance path planning on the power fault node perception network to obtain the second domain of collaborative maintenance path planning. The first domain of the collaborative maintenance path planning and the second domain of the collaborative maintenance path planning are merged to generate the first space of the maintenance path planning.

4. The method for collaborative maintenance path planning of distributed power equipment as described in claim 1, characterized in that, Based on the first space of the maintenance path planning, loss prediction and optimization are performed to obtain the second space of the maintenance path planning, including: Plan the first space according to the maintenance path and extract the first maintenance path; Loss prediction is performed on the first maintenance path to determine the loss prediction coefficient for the first path. Determine whether the predicted coefficient of the first path loss is less than the path loss threshold; If the first path loss prediction coefficient is less than the path loss threshold, the first maintenance path is added to the second space of the maintenance path planning. If the first path loss prediction coefficient is greater than or equal to the path loss threshold, the first maintenance path is eliminated. Based on the path loss threshold, the first space of the maintenance path planning is further optimized by loss prediction to generate the second space of the maintenance path planning.

5. The distributed power equipment collaborative maintenance path planning method as described in claim 4, characterized in that, Loss prediction is performed on the first maintenance path, and the loss prediction coefficient for the first path is determined, including: Multiple simulated maintenance operations were performed based on the first maintenance path to obtain multiple simulated maintenance datasets. Input the multiple simulated maintenance datasets into the maintenance loss assessment model to obtain multiple maintenance loss assessment coefficients; The average of the multiple maintenance loss assessment coefficients is calculated to generate the first path loss prediction coefficient.

6. The method for collaborative maintenance path planning of distributed power equipment as described in claim 1, characterized in that, The maintenance path optimization analytical model includes a maintenance path optimization analytical function, which is: ； Among them, MPO represents the optimality of the maintenance path, exp represents the exponential function with the natural constant e as the base, PEW represents the predetermined weight of maintenance efficiency, g(MPE) represents the normalized predicted maintenance efficiency of the path, PLW represents the predetermined weight of path loss, and g(MPL) represents the normalized predicted coefficient of path loss.

7. The method for collaborative maintenance path planning of distributed power equipment as described in claim 1, characterized in that, Multidimensional perception analysis is performed on the maintenance robot library to establish a candidate maintenance robot network, including: Based on the maintenance robot library, time margin perception optimization is performed to obtain the first candidate space of maintenance robots that meet the predetermined time margin. Based on the first candidate space of the maintenance robot, state perception optimization is performed to obtain the second candidate space of the maintenance robot; Based on the second candidate space of the maintenance robot, basic feature perception is performed to build the candidate maintenance robot network.

8. The method for collaborative maintenance path planning of distributed power equipment as described in claim 1, characterized in that, Based on the distributed power equipment network, a power fault node sensing network is established, including: Based on the distributed power equipment network, fault prediction is performed, and the fault prediction coefficient of each device is determined. Determine whether the fault prediction coefficient of each device is greater than or equal to the fault threshold of each device, and obtain the fault judgment result of each device. Based on the fault judgment results of each device, the fault nodes of the distributed power equipment network are sorted out to generate the power fault node perception network.

9. The method for collaborative maintenance path planning of distributed power equipment as described in claim 8, characterized in that, Based on the distributed power equipment network, fault prediction is performed to determine the fault prediction coefficients for each device, including: The distributed power equipment network is monitored in real time to obtain monitoring data of each power equipment; Based on the monitoring records and fault detection records of each power device in the distributed power equipment network, the gated loop unit is subjected to supervised learning to generate a fault prediction model for each power device. The monitoring data of each power device is input into the fault prediction model of each power device, and the fault prediction coefficient of each device is output.

10. A distributed power equipment collaborative maintenance path planning system, characterized in that, The system is used to execute the distributed power equipment collaborative maintenance path planning method according to any one of claims 1-9, including: The device network construction module is used to construct a distributed power device network based on the distributed power device set of the power grid; The fault perception module is used to perceive faults based on the distributed power equipment network, establish a power fault node perception network, and perform maintenance guidance feature analysis on the power fault node perception network to construct a multi-dimensional maintenance guidance vector. The maintenance network construction module is used to perform multi-dimensional perception analysis on the maintenance robot library and establish a candidate maintenance robot network. The maintenance path planning module is used to guide the candidate maintenance robot network to perform collaborative maintenance path planning on the power fault node perception network according to the multi-dimensional maintenance guidance vector, so as to obtain the first space of maintenance path planning. The first optimization module is used to perform loss prediction optimization based on the first space of the maintenance path planning to obtain the second space of the maintenance path planning. The second optimization module is used to maximize the maintenance path optimization in the second space of the maintenance path planning according to the maintenance path optimization and appropriateness analytical model, and output the maintenance path optimization result. The execution module is used to control the candidate maintenance robot network to perform maintenance on the power fault node perception network based on the maintenance path optimization result.

Images