Distribution network unmanned aerial vehicle inspection route planning method and system based on greedy algorithm

Abstract

The embodiment of the invention provides a distribution network unmanned aerial vehicle inspection route planning method and system based on a greedy algorithm. The method is applied to the technical field of unmanned aerial vehicles and comprises the following steps: establishing a distribution network unmanned aerial vehicle inspection nest optimization model; determining a nest optimization result; acquiring current environment data, and using a greedy search algorithm to construct an initial inspection route according to the inspection sub-region information, the nest optimization result, the distribution network unmanned aerial vehicle voyage data and the current environment data; and generating a route search graph according to the initial routing inspection route, obtaining real-time state information of the distribution network unmanned aerial vehicle, and determining an optimized routing inspection route according to the route search graph and the real-time state information of the distribution network unmanned aerial vehicle. According to the scheme, the inspection efficiency is improved, the energy consumption is reduced, and the safety is enhanced. The nest optimization model reasonably plans the position of the nest and reduces the flight distance. The greedy search algorithm quickly deals with emergency situations and optimizes route selection. The route search graph dynamically adjusts the inspection route to ensure full coverage of task points.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of unmanned aerial vehicles, in particular to a power distribution network unmanned aerial vehicle inspection route planning method and system based on a greedy algorithm. BACKGROUND

[0002] Power distribution network unmanned aerial vehicle inspection route planning is a key technology in the field of power inspection. It ensures the safe and efficient completion of inspection tasks by scientific and reasonable planning, fully covers each point in the power grid, and reduces energy consumption while improving inspection quality and efficiency.

[0003] Nowadays, the power distribution network unmanned aerial vehicle inspection route is planned manually. The operator first needs to outline the flight route of the unmanned aerial vehicle on the map according to the inspection target and the layout of the power grid, combined with his own professional knowledge and field investigation. In this process, they need to carefully consider the take-off point, inspection point (i.e. the power grid equipment or area that needs to be checked in detail), and possible landing point of the unmanned aerial vehicle, to ensure that these points are accurately marked and reasonably linked to form a safe and efficient flight route.

[0004] However, manual marking of points is easily affected by subjective factors of the operator, and different personnel's marking standards may differ, resulting in unreasonable route planning and failure to fully cover all points that need to be inspected. In addition, manual route creation consumes a lot of time and effort, and it is difficult to respond quickly in the face of emergency inspection tasks. At the same time, manual planning may not accurately calculate the optimal flight route of the unmanned aerial vehicle, resulting in decreased inspection efficiency and quality. SUMMARY

[0005] In order to solve the technical problems of the prior art, the present disclosure provides a power distribution network unmanned aerial vehicle inspection route planning method and system based on a greedy algorithm. The present disclosure solves the technical problems that manual marking of points is easily affected by subjective factors of the operator, different personnel's marking standards may differ, resulting in unreasonable route planning and failure to fully cover all points that need to be inspected. In addition, manual route creation consumes a lot of time and effort, and it is difficult to respond quickly in the face of emergency inspection tasks. At the same time, manual planning may not accurately calculate the optimal flight route of the unmanned aerial vehicle, resulting in decreased inspection efficiency and quality.

[0006] According to a first aspect of the present disclosure, a power distribution network unmanned aerial vehicle inspection route planning method based on a greedy algorithm is provided, comprising: obtaining an inspection network graph, inspection sub-region information, and power distribution network unmanned aerial vehicle range data, and establishing a power distribution network unmanned aerial vehicle inspection nest optimization model according to the inspection network graph, the inspection sub-region information, and the power distribution network unmanned aerial vehicle range data;

[0007] Obtaining a nest optimization target, inputting the nest optimization target into the power distribution network unmanned aerial vehicle inspection nest optimization model, and obtaining a nest optimization result;

[0008] acquire current environment data, and construct an initial inspection route using a greedy search algorithm according to the inspection sub-region information, the nest optimization result, the distribution network unmanned aerial vehicle range data, and the current environment data;

[0009] According to the initial inspection route, a route search graph is generated, real-time state information of the distribution network unmanned aerial vehicle is acquired, and an optimized inspection route is determined according to the route search graph and the real-time state information of the distribution network unmanned aerial vehicle.

[0010] According to a second aspect of the present disclosure, a distribution network unmanned aerial vehicle inspection route planning system based on a greedy algorithm is provided for performing the method as described in the first aspect, comprising: a model establishing module for acquiring an inspection network graph, inspection sub-region information, and distribution network unmanned aerial vehicle range data, and establishing a distribution network unmanned aerial vehicle inspection nest optimization model according to the inspection network graph, the inspection sub-region information, and the distribution network unmanned aerial vehicle range data;

[0011] a nest optimization module for acquiring a nest optimization target, inputting the nest optimization target into the distribution network unmanned aerial vehicle inspection nest optimization model, and obtaining a nest optimization result;

[0012] an inspection route constructing module for acquiring current environment data, and constructing an initial inspection route using a greedy search algorithm according to the inspection sub-region information, the nest optimization result, the distribution network unmanned aerial vehicle range data, and the current environment data;

[0013] an inspection route optimizing module for generating a route search graph according to the initial inspection route, acquiring real-time state information of the distribution network unmanned aerial vehicle, and determining an optimized inspection route according to the route search graph and the real-time state information of the distribution network unmanned aerial vehicle.

[0014] According to a third aspect of the present disclosure, an electronic device is provided, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the method as described above when executing the program.

[0015] In the method, system, and device provided above, the embodiments of the present disclosure improve inspection efficiency, reduce energy consumption, and enhance safety. The nest optimization model reasonably plans nest positions and reduces flight distances. The greedy search algorithm quickly responds to unexpected situations and optimizes route selection. The route search graph dynamically adjusts inspection routes to ensure full coverage of task points. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present disclosure, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0017] Figure 1 A flowchart of a power distribution network unmanned aerial vehicle inspection route planning method based on a greedy algorithm is shown according to an embodiment of the present disclosure.

[0018] Figure 2 A flowchart of a power distribution network unmanned aerial vehicle inspection route planning method based on a greedy algorithm is shown according to an embodiment of the present disclosure.

[0019] Figure 3 A schematic block diagram of a power distribution network unmanned aerial vehicle inspection route planning system based on a greedy algorithm is shown according to an embodiment of the present disclosure.

[0020] Figure 4 A block diagram of an exemplary electronic device according to an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0021] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. Note that the relative arrangement, numerical expressions, and numerical values of the components and steps set forth in these embodiments are not limiting to the scope of the present disclosure unless otherwise specifically stated.

[0022] Those skilled in the art can understand that the terms "first", "second", and the like in the embodiments of the present disclosure are only used to distinguish different steps, devices, or modules, and do not represent any specific technical meaning, nor do they represent the inevitable logical sequence between them. It should also be understood that in the embodiments of the present disclosure, "multiple" can mean two or more, and "at least one" can mean one, two, or more. It should also be understood that for any component, data, or structure mentioned in the embodiments of the present disclosure, unless specifically limited or given the opposite implication by the context, it can generally be understood as one or more. In addition, the term "and / or" in the present disclosure is only a description of the association relationship between the associated objects, which means that there can be three relationships, for example, A and / or B can represent the existence of A alone, the existence of A and B together, and the existence of B alone. In addition, the character " / " in the present disclosure generally represents an "or" relationship between the front and rear associated objects. It should also be understood that the description of each embodiment of the present disclosure emphasizes the differences between each embodiment, and the same or similar parts can be referred to each other, and for the sake of brevity, they will not be repeated.

[0023] It is also to be understood that the drawings are not necessarily drawn to scale, that the dimensions of the various parts can be exaggerated for the purpose of explanation, and that the technology disclosed herein is by no means limited to the specific examples represented therein. It should be noted that like reference numerals and letters refer to like items in the several views, and thus, no further discussion regarding such items can be needed in the following description.

[0024] In order to make the objects, technical solutions and advantages of the embodiments of the present disclosure clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present disclosure with reference to the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all the embodiments of the present disclosure. Based on the embodiments in the present disclosure, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present disclosure.

[0025] Figure 1 A flowchart of a power distribution network unmanned aerial vehicle inspection route planning method based on a greedy algorithm is provided in the embodiments of the present disclosure. The method in the embodiments of the present disclosure aims to realize accurate detection of large targets and small targets in pictures.

[0026] In S101, the inspection network graph, the inspection sub-region information and the power distribution network unmanned aerial vehicle flight range data are acquired, and a power distribution network unmanned aerial vehicle inspection nest optimization model is established according to the inspection network graph, the inspection sub-region information and the power distribution network unmanned aerial vehicle flight range data.

[0027] The inspection network graph can be geographical information representation of the power distribution network unmanned aerial vehicle inspection area, usually represented in a graph structure, in which nodes represent key task points (such as power poles, transformers, device terminals, etc.). Edges represent passable routes between task points, and the weights of edges are usually determined by distance (such as meters, kilometers), flight energy consumption (calculated according to wind speed, terrain conditions, etc.), and flight time (combined with factors such as unmanned aerial vehicle speed and terrain).

[0028] The inspection sub-region information can be a further division of the entire inspection area, and the task points in each sub-region are relatively dense, which is suitable for local optimization. Each sub-region can be defined as a cluster area, which is beneficial to unmanned aerial vehicle scheduling.

[0029] The distribution network unmanned aerial vehicle can be a unmanned aerial vehicle specially used for power grid inspection, and is designed for inspection of power distribution network equipment (such as transformers, line towers, etc.). Equipped with a camera, an infrared thermal imager, a laser radar and other devices, it is convenient to detect equipment failure, line heating, foreign matter intrusion and other problems. It has the functions of automatic inspection, route tracking, obstacle avoidance, etc., and supports remote control and automatic return.

[0030] The distribution network unmanned aerial vehicle range data can be the flight distance, task time and return distance parameters that the distribution network unmanned aerial vehicle can perform in a single flight, and can include maximum range (km): the theoretical maximum flight distance under full power. Flight speed (m / s): the average cruising speed of the unmanned aerial vehicle. Unit energy consumption (Wh / m): the average energy consumption of the unmanned aerial vehicle per 1m flight. Load capacity (kg): the load capacity when carrying inspection equipment. Environmental factor tolerance (such as maximum wind speed bearing capacity, temperature adaptation range, etc.).

[0031] The distribution network unmanned aerial vehicle inspection nest optimization model can be used to determine the optimal nest arrangement scheme to maximize the inspection efficiency and meet the range and task point coverage requirements of the unmanned aerial vehicle.

[0032] GIS data, satellite maps, equipment layout maps, etc. can be collected. Specifically, GIS (Geographic Information System) data: obtain the geographic coordinates of power grid towers, transformers, terminal equipment, etc. Power line layout map: provide the layout of power equipment and the flight route. Satellite map / aerial photography data: used to supplement the geographic information of complex areas. Unmanned aerial vehicle LiDAR (Light Detection and Ranging) scanning: more accurate for mountainous areas, dense equipment areas, etc. Then convert the positions of towers, transformers and terminal equipment into graph nodes. According to the GIS data or layout map, draw the connected flight routes between towers as the edges of the graph. Assign weights (such as flight distance, energy consumption, flight time) to each edge. Use the graph structure to represent the inspection network graph:

[0033] G = (V, E);

[0034] Wherein, V is a set of task points (such as towers, equipment); E is the connected flight route between task points;

[0035] Using the task point coordinates in the inspection network graph as input data, using the K-Means clustering algorithm with task point coordinates as input, automatically clustering into several sub-regions, each clustering center as a nest candidate point, obtaining the task point coordinates in each sub-region, and the boundary coordinates of each sub-region. Determine the distribution network unmanned aerial vehicle range data through the specification data of the unmanned aerial vehicle manufacturer.

[0036] Then the inspection network diagram is taken as the basic structure of the inspection area, indicating the task point coordinates, task type, priority and other information. The inspection sub-area information is introduced to optimize the task point allocation and route planning. The UAV flight range data is imported as the constraint condition for route planning and nest deployment. According to the task point density, UAV flight range and other parameters, the candidate nest positions are preliminarily determined, which can be arranged in the task point dense area, the terrain open area, the traffic convenient area and the like. The geographic coordinates and the distance to the surrounding task points are recorded for each candidate point. A matrix is set to record the shortest route from each task point to each nest candidate point. The matrix serves as the basis for route planning and nest allocation in the model. The nest constraint condition is set, and each task point needs to be within the flight range of at least one UAV. After completing the inspection of each task point, the UAV must be able to return to its own nest. Considering the importance of different task points, a backup nest can be set in the key equipment area. After the UAV departs from the nest and completes the task point inspection, it must ensure sufficient power to return. The core of the model is the optimization of the nest position and the task point allocation strategy. In the initial stage, several task point dense areas are selected as candidate nest points. The Voronoi diagram or K-Means clustering algorithm is used to divide the inspection coverage range of each nest, ensuring that the task points in each sub-area are within the flight range of the corresponding nest. According to the task point information determined by the inspection network diagram, each task point is allocated to the nest covering it. If multiple nests can cover it, the nest with the shortest inspection route is selected. If there are task points that cannot be covered by the nest due to flight range limitations, a nest candidate point is added near the optimal point of the route until all task points are covered. After design and construction, the distribution network UAV inspection nest optimization model is completed. The model is based on the inspection network diagram, integrates task point, sub-area information and UAV flight range data, and ensures the efficient execution of the inspection work and the safe operation of the power grid through scientific planning of the nest position and task point allocation strategy.

[0037] In S102, the nest optimization target is obtained, and the nest optimization target is input into the distribution network UAV inspection nest optimization model to obtain a nest optimization result.

[0038] The nest optimization target can be the core target of the optimization model in decision-making, usually involving key indicators such as inspection efficiency, UAV resource allocation, coverage range, cost, etc. Specifically, it can include minimizing the total energy consumption of UAV inspection (reducing power consumption and improving endurance), minimizing the total distance of UAV inspection (shortening the route and reducing flight time), minimizing the number of nests (reducing the cost of nest arrangement), maximizing the inspection coverage rate (ensuring that all task points are within the coverage range), maximizing the UAV schedulability (optimizing the arrangement of backup nests and enhancing the scheduling flexibility), and meeting the inspection cycle requirements (ensuring that the inspection task is completed according to the plan).

[0039] The nest optimization result can be the output result after the model is run, and can specifically include the nest location and number (clearly the geographic coordinates of each nest), the task point allocation of each nest (i.e., which nest each task point belongs to), and the backup nest arrangement scheme (to ensure an alternative scheme in case of an emergency).

[0040] The nest optimization target can be sent by the control center. If the nest optimization target is received, these targets need to be converted into specific mathematical models or parameters and input into the distribution network unmanned aerial vehicle nest optimization model. Specifically, the specific values or ranges of the nest optimization target can be input into the model as the target or constraint condition during solving. According to the complexity and size of the model, a suitable optimization algorithm is selected for solving. Common algorithms include linear programming, integer programming, heuristic algorithms, etc. Through the selected optimization algorithm, the model will calculate according to the input target and constraint condition, and finally obtain the optimization result. The solving process can involve iteration and adjustment until the nest optimization result is obtained.

[0041] In S103, the current environment data is obtained, and an initial inspection route is constructed using a greedy search algorithm according to the inspection sub-region information, the nest optimization result, the distribution network unmanned aerial vehicle range data, and the current environment data.

[0042] The current environment data can be real-time or immediate environment information related to the inspection task, which will affect the inspection route and task execution of the unmanned aerial vehicle. The environment data can include weather data: wind speed, air temperature, precipitation, and other weather factors, which will affect the flight stability and battery consumption of the unmanned aerial vehicle. Geographic data: such as terrain, obstacles, air traffic conditions, etc., which affect the flight route of the unmanned aerial vehicle. Real-time traffic information: ground traffic conditions or other aviation activities that can affect the flight of the unmanned aerial vehicle. Power grid load data: real-time load, fault information, etc. of the power grid to determine the priority of the inspection task.

[0043] The greedy search algorithm can be an algorithm that makes the locally optimal choice at each step. Specifically, the core idea of the greedy algorithm is local optimization: make the choice that looks best at the moment, without considering the global optimum. No backtracking: once a choice is made, it is no longer adjusted. In route planning, the greedy algorithm can be used to select the optimal route step by step according to a certain criterion (such as shortest distance, least energy consumption, etc.). Although this method does not guarantee a globally optimal solution, it often finds a good enough solution under limited computing resources.

[0044] The initial inspection route can be a preliminary route of the UAV inspection based on inspection task points, nest optimization results, UAV range data, and other information. This route is the starting route to ensure that the UAV can effectively and efficiently complete the inspection task. Specifically, it can include the order of task points: determine the order in which the UAV needs to inspect each task point. The allocation of the nest: determine the nest to which each task point belongs to ensure that the UAV can complete the task and return to the nest. Route planning: use a route planning algorithm (such as a greedy algorithm) to calculate the inspection route to ensure that the UAV's power, time, and other resources can meet the inspection requirements.

[0045] Current environmental data can be obtained in real time from multiple data sources. These data can be obtained through sensors, weather forecasts, sensors on the UAV, etc. Specifically, real-time weather information (such as wind speed, temperature, etc.) can be obtained. Battery status and power information can be obtained. Real-time geographic data (such as obstacles, flight restricted areas, etc.) can be obtained. The current state of the power grid, especially the inspection tasks of important equipment, can be obtained. Based on the inspection sub-area information and the nest optimization results, set the initial inspection route. These routes take into account the range of the UAV and the priority of each task point, and select the starting point from the task points or the nest. Based on the current environmental data (such as weather, terrain, obstacles, etc.), at each selection, use a greedy algorithm to select the next task point that is the shortest, most power-saving, or most suitable for the current environment. That is, the nearest and most suitable task point near the current task point is selected for inspection each time. Ensure that each task point is within the range of the corresponding nest, and that the UAV can return to its own nest after completing the inspection. Continuously adjust the route and task point allocation until all task points are inspected. Finally, after optimization by the greedy algorithm, the inspection route obtained is preliminary, but has taken into account current environmental data, task point allocation, nest deployment, and other factors. This route can provide a reference for the actual inspection task execution, and can be further optimized and adjusted according to specific circumstances. To ensure flight safety, a battery capacity threshold can also be introduced to determine whether to allow the next hop flight task to be executed, for example, the battery capacity threshold = total battery capacity x 80%, if the current power is lower than this threshold, the UAV needs to return to the nest or enter the charging scheduling process. And the threshold can be dynamically corrected according to the battery health state (SOH state):

[0046] C th-adj = C total x r base x (1 - a · (1 - SOH));

[0047] where C th-adj is the dynamically adjusted UAV battery capacity threshold; C total is the rated total capacity of the UAV battery; r baseis the initial UAV battery capacity threshold proportion; a is a preset aging correction coefficient; SOH is the battery health state;

[0048] wherein a can be used to regress the relationship between the probability of insufficient power leading to task interruption / return and SOH using a large amount of flight log data, so as to derive the optimal a value.

[0049] In order to highlight the advantages of the greedy search algorithm in the construction of the distribution network UAV inspection route, a comparative experiment can be carried out with other commonly used route planning algorithms (such as A* algorithm, genetic algorithm, ant colony algorithm, simulated annealing algorithm), and the performance of the route length, calculation time, feasibility ratio, etc. The following is a schematic comparative experiment result and an explanation of why the greedy algorithm is selected:

[0050] Input parameters: inspection sub-region information (location and task points of each region), nest optimization results (starting point and deployment location), distribution network UAV range data (range limit, such as maximum 20 km), current environment data (weather conditions, wind speed, obstacle distribution, etc.)

[0051] Evaluation indicators: calculation time (ms): response speed, route length (km): total flight distance, feasibility (%): success rate of generating executable route

[0052] The average calculation time of the greedy algorithm is 45 milliseconds, the average route length is 18.3 kilometers, and the feasible route generation rate is 100%; the average calculation time of the A* algorithm is 210 milliseconds, the average route length is 17.5 kilometers, and the feasible route generation rate is 100%; the average calculation time of the genetic algorithm is 480 milliseconds, the average route length is 17.9 kilometers, and the feasible route generation rate is 93%; the average calculation time of the ant colony algorithm is 620 milliseconds, the average route length is 18.0 kilometers, and the feasible route generation rate is 90%; the average calculation time of the simulated annealing algorithm is 340 milliseconds, the average route length is 18.1 kilometers, and the feasible route generation rate is 94%.

[0053] The reason for choosing the greedy algorithm is that in route planning, the nearest next target point is selected each time, the calculation is simple and direct, and only tens of milliseconds are needed, which is especially suitable for dynamic inspection scenarios affected by environmental data. Compared with other algorithms, the greedy algorithm has the shortest calculation time, only 1 / 5 of the A* algorithm. The route generated by the greedy algorithm under the condition of limiting the range always falls within the effective range and does not fail due to excessive optimization, so it has high stability. Experiments show that the success rate of the route under different random sub-region layouts is 100%. The initial route generated by the greedy algorithm can be used as a good initial solution for secondary optimization algorithms such as genetic algorithms and simulated annealing, providing a high-quality starting point and improving overall efficiency. After adding constraints such as current environmental data, obstacles, and wind speed, the greedy algorithm can quickly exclude infeasible points, keep the decision-making process simple and stable, and avoid falling into local complex search.

[0054] S104, generating a route search graph according to the initial inspection route, obtaining real-time state information of the distribution network unmanned aerial vehicle, and determining an optimized inspection route according to the route search graph and the real-time state information of the distribution network unmanned aerial vehicle.

[0055] The route search graph can be a graph structure that describes all possible choices of task points, nests, and unmanned aerial vehicle inspection routes. It is usually composed of nodes and edges: nodes represent inspection task points, nest locations, or other related locations. Edges represent the connection relationship between task points or nests, and the weight of the edge can represent the flight distance, power consumption, time consumption, etc. The route search graph can be used to represent all possible routes for the unmanned aerial vehicle to choose the optimal inspection route. The weight of each edge is usually affected by factors such as the priority of the task point, the range of the unmanned aerial vehicle, environmental conditions, etc.

[0056] The real-time state information can refer to the instantaneous state and environmental data of the distribution network unmanned aerial vehicle during task execution, which can include the current battery state (current power of the unmanned aerial vehicle battery, charging state, etc.), current flight speed, flight height, remaining range, current position information (GPS coordinates) of the unmanned aerial vehicle, current flight direction and track, whether the unmanned aerial vehicle has a fault or abnormal state.

[0057] The optimized inspection route can be obtained by adjusting and optimizing the existing route to make it more suitable for current environmental conditions, unmanned aerial vehicle state, and task requirements, thereby improving inspection efficiency, reducing energy consumption and time consumption, and ensuring the safety of the unmanned aerial vehicle during inspection.

[0058] A route search graph can be generated based on the initial inspection route. The nodes in the graph are task points, nests or other key locations, and the weights of the edges represent the cost of the UAV from one point to another (such as flight distance, energy consumption, etc.). In the graph, information such as the priority of the task point, geographic coordinates, etc. is labeled for each node. The weight of each edge is set, taking into account the density of the task points, environmental factors (weather, obstacles, etc.), battery power and flight distance, etc. The UAV collects real-time state information by integrating various sensors (such as battery management systems, pitot tubes, barometric altimeters, GPS receivers, inertial navigation systems, magnetometers, etc.) and monitoring systems, and updates the route search graph according to the real-time state information of the distribution UAV. For example, if the battery power of some task points is insufficient to complete the task, the route search graph will consider re-planning the route at a new nest point. The route graph is dynamically adjusted according to changes in weather or other environmental factors, for example, if the wind speed is too high, the route may need to avoid open areas. A suitable route optimization algorithm (such as Dijkstra's algorithm, A* algorithm, genetic algorithm, etc.) is used to optimize the route. In the graph, the optimal route from the starting point to the end point is calculated, taking into account the constraints of each node and the optimization goals (such as the shortest route, minimum energy consumption, etc.), and finally an optimized inspection route is obtained, for example, Dijkstra's algorithm: used to find the shortest route from the starting point to each node. A* algorithm: through heuristic search, combining distance and current state information, to find the optimal route. Genetic algorithm: simulates natural selection to optimize the route, suitable for complex and highly constrained scenarios.

[0059] In the embodiments of the present application, the inspection network graph, the inspection sub-region information, and the distribution UAV range data are obtained, and a distribution UAV inspection nest optimization model is established according to the inspection network graph, the inspection sub-region information, and the distribution UAV range data; the nest optimization target is obtained, the nest optimization target is input into the distribution UAV inspection nest optimization model, and a nest optimization result is obtained; the current environment data is obtained, and an initial inspection route is constructed according to the inspection sub-region information, the nest optimization result, the distribution UAV range data, and the current environment data using a greedy search algorithm; a route search graph is generated according to the initial inspection route, real-time state information of the distribution UAV is obtained, and an optimized inspection route is determined according to the route search graph and the real-time state information of the distribution UAV. Through the above-mentioned distribution UAV inspection route planning method based on the greedy algorithm, the inspection efficiency is improved, the energy consumption is reduced, and the safety is enhanced. The nest optimization model reasonably plans the nest position, and reduces the flight distance. The greedy search algorithm quickly responds to unexpected situations and optimizes route selection. The route search graph dynamically adjusts the inspection route to ensure full coverage of the task points.

[0060] On the basis of the above technical solutions, after the optimized inspection route is determined according to the route search graph and the real-time state information of the distribution UAV, the method further includes:

[0061] send the optimized inspection route to the distribution network UAV, and if it is identified that the distribution network UAV starts to inspect according to the optimized inspection route, determine in real time whether the current environment data has changed;

[0062] reacquire the current environment data each time the current environment data has changed, acquire real-time position information of the distribution network UAV, and update the optimized inspection route according to the real-time position information and the reacquired current environment data;

[0063] send the updated optimized inspection route to the distribution network UAV, so that the distribution network UAV inspects according to the updated optimized inspection route until the distribution network UAV completes the inspection.

[0064] In the scheme, the real-time position information of the UAV can be data about the current position of the UAV acquired and transmitted in real time by specific positioning technology and equipment during flight of the UAV. These data are crucial for safe flight, route planning, task execution and remote monitoring of the UAV.

[0065] The determined optimized inspection route can be sent to the distribution network UAV, and it is ensured that the UAV successfully receives the route instruction. Then, the takeoff signal and task execution state of the UAV are monitored. If it is identified that the UAV has started to inspect according to the optimized route, an environment data monitoring mechanism is started. Specifically, the current environment data needs to be monitored regularly or continuously. If the environment data changes significantly (such as an increase in wind speed or the appearance of temporary obstacles), a route updating mechanism is triggered. Through positioning means such as GPS and inertial navigation, real-time position information of the UAV is acquired. According to the latest environment data and the current position of the UAV, the inspection route is reoptimized using a route planning algorithm (such as A* algorithm, Dijkstra algorithm, etc.). When the route is adjusted, it should be ensured that all task points are covered, the UAV has enough power to return safely, and the identified risk areas are avoided. The updated optimized inspection route is sent to the UAV, so that it switches to the latest route and continues to inspect.

[0066] In the scheme, by updating the inspection route in real time, the distribution network UAV can fly more accurately according to the current environment, avoiding unnecessary flight time and energy consumption. By optimizing the inspection route and reducing unnecessary flight time, the operation and maintenance cost of the UAV can be reduced.

[0067] On the basis of the above technical scheme, optionally, after determining in real time whether the current environment data has changed, the method further includes:

[0068] If there is no change in the current environmental data in the inspection process of the distribution network UAV, the hovering power consumption, the residence time and the actual flight speed of the distribution network UAV at each task point in the optimized inspection route are determined, and the average flight speed, the total flight power consumption, the windward area, the preset resistance coefficient of the distribution network UAV and the flight distance between each task point of the distribution network UAV are obtained, and the unit energy consumption data of the distribution network UAV is calculated according to the total flight power consumption and the average flight speed;

[0069] The air density is obtained, and the air resistance coefficient of the distribution network UAV in the optimized inspection route is calculated according to the average flight speed, the total flight power consumption, the windward area, the preset resistance coefficient of the distribution network UAV, the air density and a preset air resistance coefficient calculation formula; wherein the preset air resistance coefficient calculation formula is:

[0070]

[0071] Wherein, k is the air resistance coefficient; P drag is the total flight power consumption; C d is the preset resistance coefficient of the distribution network UAV; p is the air density; A is the windward area; v avg is the average flight speed;

[0072] The total number of task points is obtained, and the total energy consumption data of the distribution network UAV is calculated according to the total number of task points, the hovering power consumption, the residence time, the actual flight speed, the average flight speed, the unit energy consumption data, the flight distance between each task point of the distribution network UAV, the air resistance coefficient and a preset total energy consumption calculation formula of the distribution network UAV;

[0073] If the total energy consumption data is lower than the preset energy consumption threshold, the optimized inspection route is sent to the distribution network UAV after a preset inspection time interval, so that the distribution network UAV completes the next inspection according to the optimized inspection route:

[0074] Correspondingly, after the optimized inspection route is sent to the distribution network UAV after a preset inspection time interval, the method further comprises:

[0075] If it is identified that the distribution network UAV starts to inspect according to the optimized inspection route, it is determined in real time whether there is a change in the current environmental data;

[0076] The current environmental data is reacquired each time there is a change in the current environmental data, the real-time position information of the distribution network UAV is obtained, and the optimized inspection route is updated according to the real-time position information and the reacquired current environmental data;

[0077] The updated optimized inspection route is sent to the distribution network UAV, so that the distribution network UAV inspects according to the updated optimized inspection route until the distribution network UAV completes the inspection.

[0078] In the present solution, the hovering power consumption can be the energy consumption of the power distribution network unmanned aerial vehicle when hovering above each task point.

[0079] The residence time can be the time required for the power distribution network unmanned aerial vehicle to perform inspection tasks (such as shooting, scanning, signal testing, etc.) at each task point.

[0080] The actual flight speed can be the real-time flight speed of the power distribution network unmanned aerial vehicle in a non-hovering state, affected by wind speed, load, flight mode, etc.

[0081] The average flight speed can be the average flight speed of the unmanned aerial vehicle during the entire inspection process.

[0082] The total flight power consumption can be the total energy consumed by the unmanned aerial vehicle during flight, including power consumption during flight, hovering, etc.

[0083] The total energy consumption data can be the total energy consumed by the unmanned aerial vehicle to complete all task point inspections, including energy consumed during flight and energy consumed during hovering at task points.

[0084] The windward area can be the cross-sectional area of the unmanned aerial vehicle facing the airflow during flight, affecting the size of air resistance.

[0085] The preset power distribution network unmanned aerial vehicle drag coefficient refers to a dimensionless parameter that is preset or expected during the design and manufacture of the power distribution network unmanned aerial vehicle, based on factors such as the layout of the unmanned aerial vehicle, the shape of the wings, the shape of the fuselage, the shape of the tail, etc. The ratio of the resistance of the unmanned aerial vehicle during flight to the dynamic pressure of the airflow and the reference area.

[0086] The flight distance can be the straight-line flight distance of the unmanned aerial vehicle between each task point, affecting the total flight power consumption and time.

[0087] The unit energy consumption data can refer to the energy consumption index of the unmanned aerial vehicle per unit of flight distance or time.

[0088] Air density can be the mass of a unit volume of air, usually expressed in kg / m 3 . Air density fluctuates with changes in temperature, air pressure, and humidity.

[0089] The air resistance coefficient can be a dimensionless coefficient that describes the resistance experienced by an object moving in air, indicating the degree of air resistance affecting the object.

[0090] The total number of task points can be the total number of task points that need to be completed in the optimized inspection route.

[0091] The preset energy consumption threshold can be the maximum energy consumption value that the unmanned aerial vehicle can withstand, exceeding which can result in insufficient power and task failure.

[0092] The hovering power consumption can be monitored by a current sensor during hovering, and the power consumption can be calculated by multiplying the current and voltage. The residence time can be extracted from the task log. The actual flight speed can be obtained in real time by using a GPS sensor or an inertial navigation system (INS). The average flight speed can be obtained by calculating the ratio of the total flight distance to the total flight time of the UAV on the entire inspection route. The total flight power consumption can be recorded by a current sensor or a power gauge of the UAV power module. The windward area can be measured according to the appearance size of the UAV. The flight distance between each task point of the UAV can be obtained by measurement or estimation according to map data. The air density can be measured and calculated by using environmental monitoring sensors such as temperature and humidity sensors and barometers. The total number of task points can be read from the inspection network diagram or the task plan data. The preset energy consumption threshold and the preset distribution network UAV drag coefficient can be read from the database. The unit energy consumption data can be obtained by dividing the total flight power consumption by the average flight speed, which represents the energy required by the UAV to fly a certain distance. The average flight speed, the total flight power consumption, the windward area, the preset distribution network UAV drag coefficient, and the air density are substituted into the preset air resistance coefficient calculation formula to calculate the air resistance coefficient experienced by the distribution network UAV in the optimized inspection route. Then, the total number of task points, the hovering power consumption, the residence time, the actual flight speed, the average flight speed, the unit energy consumption data, the flight distance between the distribution network UAV at each task point, and the air resistance coefficient are substituted into the preset total energy consumption calculation formula of the distribution network UAV to calculate the total energy consumption data of the distribution network UAV. If the calculated total energy consumption data is lower than the preset energy consumption threshold, it is considered that the optimized inspection route is feasible, and the optimized inspection route can be sent to the distribution network UAV after a preset inspection time interval.

[0093] In the scheme, by introducing the total energy consumption threshold judgment mechanism, it is ensured that the UAV performs inspection when the power is sufficient, and the crash or task failure caused by power depletion is avoided.

[0094] On the basis of the above technical scheme, optionally, after calculating the total energy consumption data of the distribution network UAV, the method further comprises:

[0095] If the total energy consumption data is higher than the preset energy consumption threshold, the optimized inspection route, the hovering power consumption, the residence time, the actual flight speed, the average flight speed, the unit energy consumption data, the flight distance between each task point, and the air resistance coefficient are input into the preset route planning model to update the optimized inspection route.

[0096] In the scheme, the preset route planning model can be an optimization model specially designed for UAV inspection tasks, and the purpose is to dynamically adjust and optimize the inspection route under the premise of meeting the energy consumption constraint, so as to ensure the smooth completion of the inspection task.

[0097] If the total energy consumption data is higher than the preset energy consumption threshold, the optimized inspection route, hovering power consumption, dwell time, actual flight speed, average flight speed, unit energy consumption data, flight distance between task points, air resistance coefficient are input into the preset route planning model. The model analyzes the collected data and identifies the following key issues: high energy consumption section identification: by analyzing the unit energy consumption data, determine which sections have abnormal energy consumption, which may be caused by strong winds, steep slopes, or frequent takeoffs and landings. Label these high energy consumption sections as key optimization objects. Range limited area identification: combined with the distance between task points, the distribution of return points, identify areas that exceed the maximum range of the unmanned aerial vehicle. Task point aggregation analysis: determine the task point density, whether there is a high-density area to optimize task point distribution and reduce the number of round trips. According to the analysis results, the model executes the following optimization strategies: route rearrangement (local optimization): for high energy consumption sections caused by unreasonable routes, the model adjusts the order of task points and optimizes flight routes to reduce total energy consumption. Select the route with the lowest unit energy consumption as the preferred option. Task point distribution optimization: if the task point density is high, the model may reassign some task points to more suitable nests to reduce flight distance. Avoid unfavorable environmental areas: combined with air resistance analysis, the model can choose to avoid high wind speed, complex terrain, and other high resistance areas. Add relay points or nests (backup strategy): if the route cannot be optimized to meet the energy consumption threshold, the model suggests adding relay points or temporary charging points in key areas. The model generates multiple candidate routes according to the optimization strategy. Recalculate the total energy consumption for each candidate route and select the route with the lowest total energy consumption as the optimization result. Then send the optimized inspection route to the unmanned aerial vehicle control system for it to perform the inspection task according to the new route.

[0098] The preset route planning model training steps are:

[0099] Collect and organize the data required for model training, ensuring comprehensive coverage of various inspection scenarios. Data types include flight path data: inspection network map (task point location, number, connected flight path), historical inspection flight path, and corresponding total energy consumption. UAV performance data: hovering power consumption, dwell time, actual flight speed, average flight speed, unit energy consumption, total flight power consumption, maximum range. Environmental parameter data: air density, wind speed, temperature and humidity under different weather conditions, air resistance coefficient between typical task points. Task point feature data: task point density, device type (critical equipment, ordinary equipment, etc.), priority, inspection frequency. To ensure the accuracy of model training, the original data needs to be preprocessed: data cleaning: remove abnormal data (such as false energy consumption values or unreasonable flight paths). Data normalization: normalize numerical data such as flight speed, power consumption, and flight distance to avoid dimensional differences affecting the model. Data augmentation: generate simulated data (such as flight parameters under extreme weather and complex terrain conditions) to improve model robustness. Extract core features that have a greater impact on the flight path planning model, including: spatial coordinates (X, Y, Z) of each task point, shortest flight path distance between each task point, unit energy consumption data for each flight path, environmental parameters (air density, resistance coefficient), priority and density of inspection task points. Choose the appropriate model framework and build the flight path planning model, for example, Dijkstra algorithm: suitable for shortest path search, can be used as the initial scheme of flight path. A(A-Star) algorithm: suitable for flight path search in complex environments, combined with heuristic function to reduce search space. Reinforcement learning models (such as DQN, PPO): use deep learning to continuously optimize the flight path in dynamic environments. The model structure is: input layer: receives feature data (flight path information, UAV performance, environmental parameters, etc.), hidden layer: constructs flight path search logic, calculates total energy consumption of each flight path, output layer: outputs the optimal flight path sequence. The training process is: initialize the environment: define the environment state, including task point distribution, flight path map, UAV performance data, set the starting point (nest) and the end point (all task points covered). Define the reward mechanism: reward decisions that have short inspection flight paths and low energy consumption, and punish decisions that exceed the maximum range or energy consumption threshold of the UAV. Strategy update: use the ε-greedy strategy to explore the optimal flight path during training, and continuously adjust the parameters to make the model tend to the flight path with the lowest energy consumption. Model optimization: use gradient descent optimizers (such as Adam, RMSProp) to minimize energy consumption error, set early stopping mechanism to avoid model overfitting. Then use an independent validation set to evaluate the performance of the model, the key indicators include: flight path optimality: whether the inspection flight path is close to the theoretical optimal solution. Energy consumption control: whether the total energy consumption of the model recommended flight path meets the threshold. Robustness: whether the model can still stably output in a variable environment (such as sudden wind speed changes, complex terrain).

[0100] In the scheme, the model preferentially selects the optimal route to ensure the lowest total energy consumption. The model can quickly respond to environmental changes and flexibly adjust the inspection route.

[0101] On the basis of the above technical solutions, optionally, the preset total energy consumption calculation formula of the distribution network unmanned aerial vehicle is:

[0102]

[0103] wherein, E total is the total energy consumption data; N is the total number of task points; i is the index of the current task point; P h is the hovering power consumption; t i is the dwell time; P m is the unit energy consumption data; D i is the flight distance of the distribution network unmanned aerial vehicle between each task point; V is the average flight speed; k is the air resistance coefficient; v i is the actual flight speed; g is the gravitational acceleration.

[0104] Figure 2 is the flowchart of the distribution network unmanned aerial vehicle inspection route planning method based on the greedy algorithm provided by the embodiments of the present disclosure. The method can include the following steps:

[0105] S201, real-time temperature data, real-time amplitude data and fault alarm times of the distribution network unmanned aerial vehicle are obtained, and the fault probability of the distribution network unmanned aerial vehicle is calculated according to the real-time temperature data, the real-time amplitude data, the fault alarm times and a preset fault probability calculation formula.

[0106] The real-time temperature data can be the temperature information inside and outside the body of the distribution network unmanned aerial vehicle during flight.

[0107] The real-time amplitude data can refer to the body vibration amplitude of the distribution network unmanned aerial vehicle during flight.

[0108] The fault alarm times can be the number of abnormal alarms triggered by the unmanned aerial vehicle in the current inspection period.

[0109] The fault probability can represent the possibility of failure of the unmanned aerial vehicle in the current state, and the value is usually between 0 and 1.

[0110] The temperature sensor (such as PT100, NTC thermistor) equipped inside the UAV can monitor the temperature of the motor, control system, shell and other components in real time. The temperature data is uploaded in real time through the flight controller (FCU) of the UAV or the remote monitoring platform. A sensor capable of measuring amplitude, such as an acceleration sensor or a vibration sensor, is installed on the UAV. Through the flight control system or data acquisition module of the UAV, the amplitude data output by the sensor is read. The UAV control system is provided with a self-checking mechanism, and the number of triggering times of each abnormal event is recorded through the state monitoring module. Each alarm event (such as motor abnormality, battery voltage fluctuation, etc.) is recorded and the count is accumulated. Then the temperature data, amplitude data and alarm frequency are normalized to ensure that the data dimensions are consistent. Remove noise data to avoid calculation deviation caused by sensor error. The real-time temperature data, real-time amplitude data and fault alarm frequency are substituted into the preset fault probability calculation formula to obtain the fault probability of the distribution network UAV.

[0111] In S202, if the fault probability exceeds the preset fault probability threshold, the position data of each abnormal repair point and the current position data of the distribution network UAV are obtained, and the target abnormal repair point is determined according to the position data of each abnormal repair point and the current position data of the distribution network UAV.

[0112] The preset fault probability threshold can refer to a criterion for determining whether the UAV has reached an abnormal state, which is usually a set value.

[0113] The abnormal repair point can refer to a place with fault handling or emergency maintenance function in the UAV inspection area, which can include a UAV docking station (providing battery replacement, equipment maintenance), a maintenance service station (with professional personnel and tools), and a safe landing point (emergency landing position, ensuring safe landing of the UAV).

[0114] The position data can be geographic coordinate data of each abnormal repair point.

[0115] The current position data can be the current GPS positioning data of the UAV, reflecting the real-time position information.

[0116] The target abnormal repair point can refer to the selected optimal repair point among the many abnormal repair points.

[0117] If the failure probability exceeds the preset failure probability threshold, the coordinates of the maintenance stations and stop points previously entered in the system are queried. The real-time coordinates are obtained using the GPS module or the inertial navigation system (INS) of the unmanned aerial vehicle. A weighted graph is constructed, with the current position of the unmanned aerial vehicle and each abnormality repair point as a node in the graph, and the length of the route between each point as the weight. The Dijkstra algorithm or the A* algorithm is used to calculate the shortest route. Specifically, the current position of the unmanned aerial vehicle can be set as the starting point, and each abnormality repair point can be set as the target point. The route length of each node is initialized to infinity, and the route length of the starting point is set to 0. The point with the shortest route is selected and marked as visited. The shortest route of its adjacent nodes is updated in turn until all nodes have been visited. The abnormality repair point with the shortest route or the highest score is selected as the final target.

[0118] In S203, the position data of the target abnormality repair point is sent to the network distribution unmanned aerial vehicle, so that the network distribution unmanned aerial vehicle goes to the target abnormality repair point for abnormality checking according to the position data of the target abnormality repair point.

[0119] According to the communication capability of the unmanned aerial vehicle and the on-site environment, a suitable communication mode is selected, such as 4G / 5G, Wi-Fi, satellite communication, etc. Then, the position data of the target abnormality repair point is sent to the network distribution unmanned aerial vehicle according to the selected communication mode.

[0120] In this embodiment, by obtaining real-time data such as temperature, amplitude, and failure alarm frequency, and combining the calculation of failure probability, self-diagnosis and intelligent patrol route optimization of the unmanned aerial vehicle can be realized. This not only improves the patrol efficiency, but also reduces the maintenance cost, increases the reliability and safety of the patrol, and finally provides a strong guarantee for the stable operation of the network distribution system.

[0121] On the basis of the above technical solutions, optionally, the preset failure probability calculation formula is:

[0122]

[0123] wherein, P f (t) is the failure probability; w1 is the preset temperature weight coefficient; T t is the real-time temperature data; w2 is the preset amplitude weight coefficient; V t is the real-time amplitude data; w3 is the preset failure alarm weight coefficient; A t is the number of failure alarms; and b is the preset bias term.

[0124] In this solution, the bias term b is a model parameter that adjusts the output of the model to ensure that it can accommodate the values of all input features. For most machine learning models, the bias term is learned through the training process, but at the initialization stage of the model, it is usually preset. Bias term presetting can be done in the following ways: zero initialization: the most common approach is to initialize the bias term to zero. The advantage of this is to simplify the initial settings of the model and avoid any biased assumptions.

[0125] Small constant initialization: the bias term is sometimes initialized to a very small constant value (e.g., b = 0.1). This method can ensure that at the beginning of training, the output of the model will not be greatly biased due to initialization problems.

[0126] Random initialization: in some complex models, the bias term may be randomly initialized to a small random value. This can prevent symmetry problems during model training (especially in neural networks).

[0127] Setting by prior knowledge: in some applications, the bias term can be set based on domain knowledge or prior data. For example, if the initial state of the system or the model output in certain situations should have a specific value, the bias term can be set based on this knowledge.

[0128] The weights w1, w2, w3, w4 can be initialized to very small random values, usually randomly drawn from a uniform distribution or a normal distribution.

[0129] On the basis of the above technical solutions, optionally, after calculating the fault probability of the distribution network unmanned aerial vehicle, the method further comprises:

[0130] If the fault probability exceeds the preset critical warning threshold and the fault probability does not exceed the preset critical warning threshold, the real-time temperature data, the real-time amplitude data, and the fault alarm number of the distribution network unmanned aerial vehicle are updated every preset monitoring interval, and the fault probability of the distribution network unmanned aerial vehicle is updated according to the updated real-time temperature data, real-time amplitude data, fault alarm number, and preset fault probability calculation formula;

[0131] When the fault probability exceeds the preset fault probability threshold, the position data of each abnormal repair point and the current position data of the distribution network unmanned aerial vehicle are reacquired, and the target abnormal repair point is re-determined according to the position data of each abnormal repair point and the current position data of the distribution network unmanned aerial vehicle;

[0132] The position data of the re-determined target abnormal repair point is sent to the distribution network unmanned aerial vehicle, so that the distribution network unmanned aerial vehicle goes to the target abnormal repair point for abnormality checking according to the position data of the target abnormal repair point.

[0133] In this scheme, the preset critical early warning threshold can be a fault probability value used to determine whether the distribution network UAV needs to cause fault attention. If the fault probability of the distribution network UAV exceeds this threshold, an emergency response or early warning will be triggered, and necessary repair, inspection or adjustment will be carried out.

[0134] The preset monitoring interval can refer to the time interval between two updates of the distribution network UAV in the monitoring process. For example, it can be set to every 10 minutes, 1 hour or other reasonable time period. At the end of each monitoring interval, the distribution network UAV will reacquire real-time data and update its fault probability.

[0135] The system will periodically acquire new real-time data, which includes real-time temperature data, real-time amplitude data and fault alarm times of the distribution network UAV. Using the preset fault probability calculation formula, the real-time data (temperature, amplitude and alarm times) are input into the formula to calculate the updated fault probability. The fault probability recalculated based on the new input data is compared with the preset fault probability threshold. If the calculated fault probability exceeds the threshold, the subsequent step of re-determining the abnormal repair point is triggered. When the fault probability exceeds the preset fault probability threshold, the position data of each abnormal repair point and the current position data of the distribution network UAV can be acquired from the monitoring system, and the optimal flight route is calculated using a route planning algorithm (such as A* algorithm, Dijkstra algorithm, etc.) according to the current position of the UAV and the position of the abnormal repair point. The abnormal repair point closest to the current UAV and with the most urgent repair demand is selected as the target repair point. The position data of the target repair point is transmitted to the distribution network UAV for it to perform the inspection task according to the new repair point.

[0136] In this scheme, the accuracy of fault prediction is improved through dynamic monitoring and intelligent decision-making, the efficiency and accuracy of the inspection task are optimized, the reliability of the distribution network UAV is improved, the waste of energy and resources is reduced, the maintenance cost is reduced, and the efficient and intelligent operation of the system is ensured.

[0137] Figure 3 A distribution network UAV inspection route planning system based on a greedy algorithm is provided for the embodiments of the present disclosure. The system comprises:

[0138] The model establishment module 301 is configured to acquire the inspection network graph, the inspection sub-region information and the distribution network UAV flight range data, and establish a distribution network UAV inspection nest optimization model according to the inspection network graph, the inspection sub-region information and the distribution network UAV flight range data.

[0139] The nest optimization module 302 is configured to acquire a nest optimization target, input the nest optimization target into the distribution network UAV inspection nest optimization model, and obtain a nest optimization result.

[0140] The inspection route construction module 303 is configured to acquire current environment data, and construct an initial inspection route according to the inspection sub-region information, the nest optimization result, the distribution network UAV range data and the current environment data by using a greedy search algorithm.

[0141] The inspection route optimization module 304 is configured to generate a route search graph according to the initial inspection route, acquire real-time state information of the distribution network UAV, and determine an optimized inspection route according to the route search graph and the real-time state information of the distribution network UAV.

[0142] Figure 4 A schematic block diagram of an electronic device 400 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptops, desktops, tablets, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown here, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the present disclosure described and / or claimed in this document.

[0143] The electronic device 400 includes a computing unit 401 that can perform various appropriate actions and processes according to a computer program stored in a ROM 402 or a computer program loaded into a RAM 403 from a storage unit 408. Various programs and data required for the operation of the electronic device 400 can also be stored in the RAM 403. The computing unit 401, the ROM 402, and the RAM 403 are connected to each other through a bus 404. An I / O interface 405 is also connected to the bus 404.

[0144] Various components in the electronic device 400 are connected to the I / O interface 405, including an input unit 406, such as a keyboard, a mouse, etc., an output unit 407, such as various types of displays, a speaker, etc., a storage unit 408, such as a magnetic disk, an optical disk, etc., and a communication unit 409, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 409 allows the electronic device 400 to exchange information / data with other devices through a computer network, such as the Internet, and / or various telecommunication networks.

[0145] The computing unit 401 can be various general and / or special purpose processing components with processing and computing capabilities. Some examples of the computing unit 401 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, and the like. The computing unit 401 performs various methods and processes described above, such as the power grid inspection route planning method based on a greedy algorithm. For example, in some embodiments, the power grid inspection route planning method based on a greedy algorithm can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 408. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 400 via the ROM 402 and / or the communication unit 409. When the computer program is loaded into the RAM 403 and executed by the computing unit 401, one or more steps of the power grid inspection route planning method based on a greedy algorithm described above can be performed. Alternatively, in other embodiments, the computing unit 401 can be configured to perform the power grid inspection route planning method based on a greedy algorithm by any other appropriate means, such as by means of firmware.

[0146] Various implementations of the systems and techniques described above herein can be realized in digital electronic circuitry, integrated circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a system on a chip (SOC), a programmable logic device (PLD), a computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0147] Program code for carrying out methods of the present disclosure can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the program code, when executed by the processor or controller, produces a means for implementing the functions / operations specified in the flowchart diagrams and / or block diagrams. The program code can be executed entirely on a machine, partially on a machine, partially on a machine as a standalone software package, partially on a machine and partially on a remote machine or entirely on a remote machine or server.

[0148] It should be understood that the various forms of flow shown above can be used with reordering, additions, or deletions of steps. For example, the steps recited in the present disclosure can be performed in parallel, in series, or in a different order, as long as the desired results of the technology disclosed in the present disclosure are achieved, which are not limited herein.

[0149] The specific embodiments discussed above do not limit the scope of the present disclosure. Various modifications, combinations, sub-combinations and alternatives can be apparent to one of ordinary skill in the art and can be made to the disclosure without departing from the spirit and scope of the disclosure. Any modifications, changes, and improvements that have not been described herein are intended to be included within the scope of the present disclosure.

Claims

1. A distribution network UAV inspection route planning method based on a greedy algorithm, characterized by: The method comprises: Obtain the inspection network diagram, inspection sub-area information and distribution network UAV range data, and establish a distribution network UAV inspection nest optimization model based on the inspection network diagram, inspection sub-area information and distribution network UAV range data; obtain the nest optimization target, input the nest optimization target into the distribution network UAV inspection nest optimization model to obtain the nest optimization result; obtain the current environmental data, use the greedy search algorithm to construct the initial inspection route based on the inspection sub-area information, nest optimization results, distribution network UAV range data and current environmental data; generate a route search map based on the initial inspection route, obtain the real-time status information of the distribution network UAV, and determine the optimized inspection route based on the route search map and the real-time status information of the distribution network UAV.

2. The method according to claim 1, characterized in that in, After determining the optimized inspection route based on the route search map and the real-time status information of the network-connected UAV, the method further includes: The optimized inspection route is sent to the network distribution drone. If the network distribution drone is identified, it will start to inspect according to the optimized inspection route and determine in real time whether there are any changes in the current environmental data. Re-acquire the current environmental data each time there is a change, obtain the real-time location information of the network-connected drone, and update and optimize the inspection route based on the real-time location information and the re-acquired current environmental data; The updated optimized inspection route is sent to the network distribution drone, and the network distribution drone performs inspection according to the updated optimized inspection route until the network distribution drone completes the inspection.

3. The method according to claim 2, characterized in that in, After determining in real time whether the current environmental data has changed, the method further includes: If the current environmental data does not change during the inspection process of the network distribution drone, determine the hovering power consumption, dwell time, and actual flight speed of the network distribution drone at each task point in the optimized inspection route, and obtain the average flight speed, total flight power consumption, frontal area, preset network distribution drone drag coefficient, and flight distance between each task point of the network distribution drone, and calculate the unit energy consumption data of the network distribution drone based on the total flight power consumption and average flight speed; Obtain air density and calculate the air drag coefficient experienced by the network distribution drone during the optimized inspection route based on the average flight speed, total flight power consumption, frontal area, preset network distribution drone drag coefficient, air density, and preset air drag coefficient calculation formula. The preset air drag coefficient calculation formula is: Wherein, k is the air resistance coefficient; P drag is the total power consumption during flight; C d is the preset network distribution drone drag coefficient; ρ is the air density; A is the windward area; v avg is the average flight speed; Obtain the total number of mission points and calculate the total energy consumption data of the networked drone based on the total number of mission points, hovering power consumption, dwell time, actual flight speed, average flight speed, unit energy consumption data, flight distance between each mission point, air resistance coefficient, and the preset total energy consumption calculation formula of the networked drone. If the total energy consumption data is lower than the preset energy consumption threshold, the optimized inspection route will be sent to the distribution network drone after the preset inspection time interval, so that the distribution network drone can complete the next inspection according to the optimized inspection route; Accordingly, after sending the optimized inspection route to the network distribution drone after the preset inspection time interval, the method further includes: If the network distribution drone is identified and begins to conduct inspections according to the optimized inspection route, it will determine in real time whether there are any changes in the current environmental data; Re-acquire the current environmental data each time there is a change, obtain the real-time location information of the network-connected drone, and update and optimize the inspection route based on the real-time location information and the re-acquired current environmental data; The updated optimized inspection route is sent to the network distribution drone, and the network distribution drone performs inspection according to the updated optimized inspection route until the network distribution drone completes the inspection.

4. The method according to claim 3, characterized in that in, After calculating the total energy consumption data of the network-distributed UAV, the method further includes: If the total energy consumption data is higher than the preset energy consumption threshold, the optimized inspection route, hovering power consumption, stay time, actual flight speed, average flight speed, unit energy consumption data, flight distance between each mission point, and air resistance coefficient will be input into the preset route planning model to update the optimized inspection route.

5. The method according to claim 3, characterized in that in, The preset total energy consumption calculation formula for network distribution drones is: Among them, E total is the total energy consumption data; N is the total number of task points; i is the index of the current task point; P h is the hovering power consumption; t i is the residence time; P m is the unit energy consumption data; D i is the flight distance between each mission point of the network distribution UAV; V is the average flight speed; k is the air resistance coefficient; v i is the actual flight speed; g is the acceleration due to gravity.

6. The method according to claim 1, characterized in that in, After determining the optimized inspection route based on the route search map and the real-time status information of the network-connected UAV, the method further includes: Obtain the real-time temperature data, real-time amplitude data, and fault alarm count of the distribution network UAV, and calculate the fault probability of the distribution network UAV based on the real-time temperature data, real-time amplitude data, fault alarm count, and a preset fault probability calculation formula; If the fault probability exceeds a preset fault probability threshold, the location data of each abnormal repair point and the current location data of the network distribution drone are obtained, and the target abnormal repair point is determined based on the location data of each abnormal repair point and the current location data of the network distribution drone; The location data of the target abnormality repair point is sent to the distribution network drone, so that the distribution network drone can go to the target abnormality repair point for abnormality inspection based on the location data of the target abnormality repair point.

7. The method according to claim 6, characterized in that in, The preset failure probability calculation formula is: Among them, P f (t) is the failure probability; w1 is the preset temperature weight coefficient; T t is the real-time temperature data; w2 is the preset amplitude weight coefficient; V t is the real-time amplitude data; w3 is the preset fault alarm weight coefficient; A t is the number of fault alarms; b is the preset bias item.

8. The method according to claim 6, characterized in that in, After calculating the failure probability of the network-distributing drone, the method further includes: If the failure probability exceeds the preset critical warning threshold, and the failure probability does not exceed the preset critical warning threshold, then every time the preset monitoring interval is reached, the real-time temperature data, real-time amplitude data and number of fault alarms of the distribution network drone are updated, and the failure probability of the distribution network drone is updated based on the updated real-time temperature data, real-time amplitude data, number of fault alarms and the preset failure probability calculation formula; When the fault probability exceeds the preset fault probability threshold, the location data of each abnormal repair point and the current location data of the network distribution drone are re-acquired, and the target abnormal repair point is re-determined based on the location data of each abnormal repair point and the current location data of the network distribution drone; The location data of the re-determined target abnormality repair point is sent to the distribution network drone, so that the distribution network drone can go to the target abnormality repair point for abnormality inspection based on the location data of the target abnormality repair point.

9. A distribution network drone inspection route planning system based on a greedy algorithm, used to execute the method according to any one of claims 1 to 8, characterized in that: The system comprises: The model building module is used to obtain the inspection network diagram, inspection sub-area information and distribution network drone range data, and establish the distribution network drone inspection nest optimization model based on the inspection network diagram, inspection sub-area information and distribution network drone range data; The machine nest optimization module is used to obtain the machine nest optimization target, input the machine nest optimization target into the distribution network drone inspection machine nest optimization model, and obtain the machine nest optimization result; The inspection route construction module is used to obtain current environmental data and use a greedy search algorithm to construct an initial inspection route based on inspection sub-area information, machine nest optimization results, network distribution drone range data, and current environmental data; The inspection route optimization module is used to generate a route search map based on the initial inspection route, obtain the real-time status information of the network-connected drones, and determine the optimized inspection route based on the route search map and the real-time status information of the network-connected drones.

10. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 8.

Images