Unmanned aerial vehicle task scheduling and optimal path planning method for power distribution line fault scenarios

Abstract

This invention belongs to the field of power distribution network fault emergency repair and UAV intelligent dispatching technology, and relates to a UAV task orchestration and optimal path planning method for power distribution line fault scenarios. After standardizing multi-source data, it is mapped to a set of topological segments, calculating a set of candidate fault segments, generating a set of task units, and constructing a comprehensive objective function with the shortest power restoration time or minimum power outage impact as the core objective. A task sequence is assigned to each UAV, tasks are executed, and status and evidence information are transmitted back. It is determined whether the UAV needs to trigger rolling replanning, freezing completed or uninterrupted tasks. After feasibility verification, tasks are re-executed, and fault evidence and location are written back as feedback. The method proposed in this invention can significantly shorten the critical decision-making time in the emergency response link, locate high-impact fault segments faster under the same resource conditions, and promote faster power restoration.

Description

[0001] This invention belongs to the field of emergency repair and intelligent dispatching technology for power distribution network faults, and relates to a method for UAV mission orchestration and optimal path planning for power distribution line fault scenarios. Background Technology

[0002] Power distribution networks are widely distributed, with numerous devices and a complex operating environment. They are susceptible to faults such as line breaks and pole collapses due to natural factors like typhoons and rainfall. After a fault occurs, power supply reliability and repair timeliness directly impact critical user safety and social impact, requiring power distribution operation and maintenance and emergency repair to locate and repair faulty sections within a short period. In recent years, drones have been widely used due to their advantages in speed and flexibility. However, existing drone inspection or emergency support solutions are mostly based on routine inspection approaches, with single evaluation indicators. In emergency scenarios, the fault point is often unclear; planning drone flight routes solely based on geographical location or straight-line distance can easily lead to delays in identifying critical sections, thus prolonging the overall power outage time. The high degree of uncertainty at emergency sites may also result in secondary faults, false alarms, or multiple suspected faults coexisting. New on-site clues constantly generated during drone operations can also significantly alter the priority ranking of candidate faults. Furthermore, power distribution emergency response is often limited by weather conditions, actual flight conditions, and the power and payload of the drone. In summary, existing technical solutions for solving the emergency application of drones in power distribution networks have limitations such as misaligned business objectives, coarse model granularity, and insufficient dynamic adaptability. Currently, there is a lack of feasible solutions for system security. Summary of the Invention

[0003] This invention addresses the need for rapid response to power outages caused by faults such as tripping, grounding, line breaks, pole falls, tree obstructions, and foreign object entanglement in distribution lines. It unifies the modeling of multi-source information, including distribution network topology segmentation, fault alarms and repair reports, weights of important users and loads, arrival and operational capabilities of repair teams, drone endurance and payload capacity, airspace and no-fly zones, weather conditions, and communication coverage and link reliability. With the shortest power restoration time or minimal power outage impact as the core optimization objective, it enables drones to perform tasks, prioritize tasks, allocate tasks, plan flight routes, calculate ETA / ETC time, and perform rolling replanning during execution in emergency reconnaissance, close-range verification, and necessary support operations. It can output executable tasks, flight routes, and timelines for command and dispatch, and write drone detection results back to the topology segment to support isolation, dispatch, and power restoration confirmation.

[0004] To achieve the above objectives, the present invention is implemented using the following technical solution: (1) Data Access and Standardization: Multi-source data is obtained from the distribution network automation system, power outage management system, alarm parsing service, 95598 repair reporting system, hidden danger database, UAV ground station, weather station, and communication assessment module. After coordinate transformation, topology consistency verification, time alignment, and spatial index construction, all information is uniformly mapped to the topology segment set S. The segment set is represented as: .in This is a set of topology segments, derived from the distribution topology segmentation model. Indicates the first Each section is defined by switches, boundary points, and branch line divergences; The number of segments is obtained through topological analysis.

[0005] (2) Evaluation of candidate fault segments: For each segment Calculate the confidence level of the fault Weighting of power outage impact The confidence level is calculated using the following formula: .in It is a normalization function. The segment alarm intensity characteristics (e.g., the degree of alarm location) derived from fault alarm analysis. The segmental repair clustering intensity characteristics derived from user repair clustering. Prior characteristics of historical hazards / faults in the section from the hazard database. The fusion weights are derived from the system parameter configuration.

[0006] The formula for calculating the weight of power outage impact is: . For section The influence weights are used for subsequent optimization objectives. Section The set of affected users is obtained through topological tracing. user( The importance coefficient of a device (such as hospitals, base stations, and government agencies) is higher, and it comes from user ledgers or emergency configurations.

[0007] This generates a candidate segment set. , , The output is sent to the task unit generation module.

[0008] (3) Task unit generation: For each candidate segment , , Generate at least one type of task unit to form a task unit set: .in The task unit set is output to the optimization model. For the first Each task unit includes fields such as task type (reconnaissance / review / support), associated segment, entry waypoint, estimated execution time, payload requirements, and communication requirements. This refers to the number of task units. Tasks include the following types: Level 1 tasks involve rapid patrol and sweeping along the section corridor; Level 2 tasks involve hovering over suspicious points and conducting around-the-loop evidence collection; and Support tasks are generated under conditions such as abnormal weather, nighttime, and obstructed repairs.

[0009] (4) Optimization Modeling: With the shortest power restoration time or the minimum power outage impact as the core objective, a comprehensive objective function is constructed. First, the restoration time cost is defined: Among them are The lower the recovery time cost, the faster the critical section can be recovered. For the segment influence weight, For section fault confidence, For section ( The estimated recovery time or recovery moment is obtained from the task completion time output by the scheduling and the isolation or emergency repair process. This represents the number of segments.

[0010] The overall objective function is: .

[0011] in, For the overall optimization objective, The cost of flight and execution time is calculated by summing the flight time between waypoints and the mission execution duration. The risks involved, such as no-fly zone margins and weather conditions, are determined through risk assessment. These are communication penalty items, such as insufficient coverage / poor link quality leading to control risks, obtained from communication assessment; The weighting factor comes from the system policy configuration.

[0012] (5) Task scheduling and allocation and ETA / ETC recursion: Define the UAV set as For each drone Assign a sequence of tasks and calculate the estimated arrival and completion times. The recursive formula is: .in Let $\mathbf{k}$ be the estimated time for the UAV(k) to reach the $j$-th mission point. The estimated completion time for the drone (k) to complete the (j)th task; In order to complete the task ( The associated waypoints or corridor entrances; The flight time estimation function between two points is derived from factors such as distance, speed, and wind. This is a function for the task execution duration, configured based on the task type. The (j)th task unit is assigned to the unmanned aerial vehicle (k).

[0013] (6) Feasibility verification: The planning results for each UAV are constrained and verified, taking the range margin as an example: .in Accumulated flight time for the drone (k); Accumulate the mission execution time for the drone(k); The estimated time for the drone (k) to return from its last waypoint to its take-off and landing points; Budget for available time corresponding to the drone's (k) battery power; The safety margin ratio, such as reserved power, is configured by the strategy; This means that each drone must meet the requirements.

[0014] (7) Execution and Monitoring: The UAV executes the issued task and transmits status and evidence information back. The execution monitor continuously detects four types of events: new alarms, new clues discovered, communication degradation, and resource changes. Define the criteria for triggering rolling replanning: .in This is an indicator variable for whether a replanning process is triggered. The largest change in confidence level of a candidate segment comes from a new alert or a new clue; The confidence level change threshold; This represents the largest change in risk cost, such as communication degradation or weather changes. This is the threshold for risk changes; This refers to the change in battery power / battery remaining capacity, i.e., the change in resources. This is the threshold for resource changes.

[0015] (8) Rolling replanning: Freeze completed or uninterrupted tasks, and reorder and multi-aircraft reassignment of uncompleted tasks. Re-execute steps (3) to (6) to generate new task sequences, waypoint sequences and ETA / ETC timetables, and reissue them for execution after feasibility verification, forming a closed loop of planning-execution-replanning.

[0016] (9) Result Write-back and Closed Loop: After completing the key tasks, the fault evidence detected by the UAV, such as the location of the fallen pole, the location of the tree obstruction on the line, and the discharge traces, along with the positioning results, are written back to the topology section / tower identifier, and section-level conclusions are output for isolation decision-making, emergency repair dispatch, and power restoration confirmation. At the same time, emergency repair feedback, such as the completion of isolation operation and power restoration confirmation, is received as data input for the next round of evaluation and planning, realizing a closed loop for the entire emergency response process.

[0017] This invention transforms the traditional drone scheduling based on geographical distance into task orchestration and path optimization based on topological segments and recovery time benefits. During execution, it senses new clues, new alarms, communication degradation, and resource changes, triggering rolling replanning to continuously approach the emergency target of the shortest recovery time.

[0018] During the input data access and standardization phase, four types of data are acquired and preprocessed. In the grid-side information, distribution network topology and segment information are exported from the distribution automation system in CIM / XML format, and the outage impact range is extracted from the real-time event table of the outage management system. Both types of data undergo coordinate transformation to CGCS2000, topology consistency verification, and segment ID normalization, and a basic topology table is generated using the segment ID as the primary key. Regarding alarm and repair information, fault alarm parsing results are obtained from overcurrent and protection action events transmitted by FTUs, DTUs, and fault indicators via the 104 protocol. The alarm parsing service extracts the segment location and quantifies the alarm intensity. User repair clustering results are obtained from the 95598 customer service system or the State Grid App backend, using DBSCAN spatial clustering and temporal density analysis, followed by reverse geocoding to the segment level. After second-level time and spatial alignment, the two types of data output segment-level alarm intensity values ​​and repair density. The knowledge base information is derived from historical fault records of the power transmission line condition monitoring system. The number of faults in each section over the past three years is statistically analyzed and multiplied by a seasonal coefficient to obtain the prior confidence level. In the resource and constraint information, the UAV take-off and landing point coordinates are obtained from RTK-GPS positioning via the MAVLink protocol from the ground station system; remaining battery power is read from the BQ40Z80 battery management chip; and maximum speed is derived from flight control parameters. No-fly zones are represented by GeoJSON data downloaded from the Civil Aviation Administration's UOM system. Meteorological data, including wind speed, rainfall intensity, and visibility, is collected in real-time by a micro-weather station. Communication coverage prediction is based on RSRP values ​​calculated using operator road test grids and ray tracing models. The above data undergoes preprocessing, including coordinate unification, filtering and denoising, and spatial indexing. The hazard database information is derived from historical fault records of the power transmission line condition monitoring system, real-time monitoring data from the distribution automation system, and defect registrations from manual inspections. Through integration and statistical analysis, historical hazard or fault prior characteristics for each section are formed.

[0019] In the data alignment and standardization process, all information is aligned using the segment ID as the unique key through left joins. Alarm intensity, repair density, prior probability, power outage indicator, environmental risk index, and communication penalty coefficient are combined to form a feature vector, stored in an in-memory data structure. This structure is continuously maintained by the data access and standardization service: grid-side and knowledge base information is statically loaded hourly; alarm and repair information is updated in real-time (second-level) via Kafka stream processing; drone status is updated at a frequency of 10Hz; and meteorological data is updated at a frequency of 1Hz.

[0020] The standardized data serves as the input for generating candidate fault segments. It is used for each planning read, from which the confidence features and influence weights of each segment are extracted. This allows for the calculation of the comprehensive fault confidence and the generation of a set of candidate fault segments, providing a consistent data foundation for subsequent task unit generation and optimization modeling.

[0021] Compared with the prior art, the advantages and positive effects of the present invention are as follows: (1) The method proposed in this invention introduces the section fault confidence and power outage impact weight into the optimization objective function in a unified manner, so that the task priority and path planning directly serve the power restoration objective, which can significantly shorten the critical decision time of the emergency response link, and more quickly locate high-impact fault sections and promote faster power restoration under the same resource conditions.

[0022] (2) The output of UAVs can directly support isolation decision-making and emergency repair dispatch. This invention generates candidate fault segments by using topological segments as modeling units, and explicitly associates isolable boundaries, affected user sets, and segment business value, so that clues discovered by UAVs can be written back to the operable segment level, thereby quickly supporting the reduction of power outage range, selection of isolation schemes, and dispatch and positioning of emergency repair teams, significantly improving the executability of command and dispatch, with high positioning efficiency in complex scenarios and strong resistance to uncertainty. A task-unit-based emergency operation process modeling method is proposed to realize the integrated arrangement and output of reconnaissance, verification, and support operations, significantly improving the organizational efficiency of on-site handling.

[0023] (3) A monitoring and rolling replanning mechanism was established to avoid wasting critical emergency time on low-value tasks; rolling replanning was initiated based on dynamic disturbance trigger criteria, supporting freezing or queue reordering and recalculation, which improved the dynamic adaptability of UAV scheduling and the efficiency of resource utilization. A feasibility verification system covering multiple constraints such as endurance and return margin, no-fly zone and weather risks, and communication reliability was established to improve the feasibility and safety of the project. Detailed Implementation

[0024] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described below with reference to specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0025] Numerous specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways than those described herein, and therefore the invention is not limited to the specific embodiments disclosed in the following specification.

[0026] Example 1 The hardware deployment of this invention adopts a collaborative architecture of command / dispatch terminal—communication link—UAV terminal—on-site repair terminal to meet the engineering requirements of rapid response, dynamic scheduling, and closed-loop data transmission for emergency repairs of power distribution line faults. The command / dispatch terminal can be deployed in the power distribution emergency command center or the municipal / county-level repair command station, consisting of a computing server, a data access gateway, and a visualization workstation. It is used to access the power distribution network topology segmentation model, fault alarms and repair reports, important user and load weights, hidden danger databases / historical databases, meteorological and no-fly zone data, and communication coverage assessment information, and to run core software modules such as candidate segment evaluation, task unit generation, optimization solution, and rolling replanning. The communication link adopts public cellular networks, private network data links, ground station relays, or multi-link redundancy to realize UAV status transmission, image / video evidence transmission, and task command issuance. To improve reliability under emergency conditions, link monitoring and bandwidth adaptive strategies can be configured on the ground station side. The UAV terminal includes a flight platform, flight control and navigation module, airborne communication module, mission execution and data acquisition module, and necessary payloads, such as visible light cameras, infrared cameras, supplementary lighting, loudspeakers, and drop mechanisms, selected according to the mission. It uses an airborne computing unit to parse mission instructions, execute waypoints, report status, and annotate key clues. The on-site repair terminal can consist of a mobile terminal or a vehicle-mounted terminal, used to receive mission-route-timetable and segment-level write-back results, enabling coordination with the repair team's arrival, isolation operations, and power restoration confirmation; it also supports repair feedback transmission as a data source for rolling replanning and closed-loop updates. Candidate fault segments are the basis for subsequent mission unit generation and optimization modeling. This embodiment uses topology segmentation as the main line, decomposing the line into a set of segments: .

[0027] in This is a set of topology segments, derived from the distribution topology segmentation model. Indicates the first Each section is defined by a switch / boundary point / branch branch; The number of segments is obtained through topological analysis.

[0028] Then, the fault candidate score and confidence level are calculated for each segment to form a candidate segment set and sort it, using a fusion method: .

[0029] in It is a section The confidence level of the faulty segment is output to the candidate segment set. It is the normalization function, given by the algorithm configuration. Section alarm intensity characteristics derived from fault alarm analysis The segmental repair clustering intensity characteristics derived from user repair clustering. Prior characteristics of historical hazards or faults in the section from the hazard database. The fusion weights are derived from the system parameter configuration.

[0030] Meanwhile, to reflect the business objectives of emergency scenarios, shorten the duration of power outages, and prioritize the restoration of critical users, the impact weight of power outages is calculated for each section: .

[0031] For section The influence weights are used for subsequent optimization objectives. Section The set of affected users is obtained through topological tracing. user( The importance coefficient is derived from user ledgers or emergency configurations.

[0032] Therefore, the candidate segment set is structured as follows: , , ), which serves as the input data generated by the task unit.

[0033] The candidate segment set enters the task unit generator, generating three types of tasks: Level 1 tasks involve rapid patrol along the segment corridor; Level 2 tasks involve hovering / flying around suspicious points for evidence collection; and support tasks are generated under conditions such as weather / nighttime / repair obstruction. Specifically, Level 1 tasks are wide-area reconnaissance, Level 2 tasks are close-range verification, and optional tasks include lighting, verbal communication, data drop, and guidance. This hierarchical structure (Level 1 → Level 2) allows the emergency response process to employ a speed-first, accuracy-later strategy: quickly narrowing down suspected fault locations before conducting detailed verification and evidence collection of high-value suspicious points, thereby reducing unnecessary flights and waiting.

[0034] To ensure that subsequent optimizations can uniformly handle different task types, task units are abstracted into sets: .

[0035] in The task unit set is output to the optimization model. For the first Each task unit includes fields such as task type (reconnaissance, verification, support), associated segment, entry waypoint, estimated execution time, payload requirements, and communication requirements. This represents the number of task units.

[0036] The output of the task unit generator must satisfy the following: each task unit can be mapped to a specific waypoint or corridor segment, thereby supporting the generation of paths, waypoints, and timetables.

[0037] The next step is to evaluate candidate segments. Input metrics include: fault alarm analysis, user repair report clustering, and a potential hazard database. The output is a set of candidate segments. Subsequently, different levels of tasks are generated based on this set. The key constraint is that the candidate segment output must not only indicate where the fault might be, but also the business value of prioritizing that segment. and This should also be output for use in the subsequent construction of the objective function.

[0038] During the task unit generation phase, for each candidate segment ( Generate: Level 1 wide-area reconnaissance mission, quickly patrolling along the section corridor, with the goal of discovering obvious fault patterns or locating suspicious towers; Level 2 close-range verification mission, hovering or flying around suspicious locations to collect evidence; if equipped with a payload, infrared verification or ranging actions can be added; optional support mission, when conditions such as weather, night, or obstruction of the repair team's arrival occur, generate lighting, announcement, throwing, and guidance mission units as a supplement to improve emergency response efficiency.

[0039] The process involves obtaining a set of task units that determine whether a specific payload is required, whether execution is necessary in a well-connected area, and whether a time window exists. Task orchestration and allocation are then performed, along with feasibility verification. Finally, a comprehensive objective function is constructed, with the shortest power restoration time or minimal power outage impact as the core objective. The confidence level of candidate segments (…) is used as the basis for this objective. ), weighting of power outage impact ( ) and expected recovery time ( ) constitutes the recovery time cost term, of which ( This is calculated based on the task completion time and the emergency repair process. Definition: .

[0040] Among them The lower the recovery time cost, the faster the critical section can be recovered. For the segment influence weight, For section fault confidence, For section ( The estimated recovery time / recovery moment is obtained from the task completion time output by the scheduler and the isolation / repair process. This represents the number of segments.

[0041] Taking into account flight efficiency, risk, and communication, a comprehensive objective function is formed: .

[0042] Among them Overall optimization objective The cost of flight and execution time is calculated by summing the flight time between waypoints and the mission execution duration. As a risk consideration, this embodiment uses the no-fly zone margin and weather level (obtained from risk assessment); As a communication penalty item, this embodiment selects insufficient coverage or poor link quality leading to control risks as the judgment factor, which is obtained from communication evaluation; To balance the factors, in this embodiment... , , In actual deployment, adjustments can be made appropriately based on weather conditions and communication environment, while maintaining constraints.

[0043] Define the conditions for generating the UAV's path, waypoints, and timetable; output the mission sequence and waypoint sequence; and provide information that can be used for calculation. The arrival / completion time of ). Define the set of drones as ( ), and perform ETA / ETC recursion on the mission sequence of the (k)th UAV: .

[0044] in Let be the estimated arrival time of the UAV (k) at its (j)th mission point. The estimated completion time for the drone (k) to complete the (j)th task. With the task ( The associated waypoints or corridor entrances, The flight time estimation function between two points is derived from distance, speed, and wind. This is a function for the task execution duration (configured by the task type). The (j)th task unit is assigned to the unmanned aerial vehicle (k).

[0045] Feasibility verification is performed to ensure that the output flight routes and schedules can be actually executed. This embodiment uses constraints including range, battery capacity, return time, no-fly zone avoidance, and communication coverage thresholds. Taking range capacity as an example, it is expressed by the following constraints: .

[0046] in The cumulative flight time of the drone (k) Accumulate the mission execution time for the drone(k). The estimated time for the drone (k) to return from its last waypoint to its takeoff and landing points. The available time budget corresponding to the drone's (k) battery power. The safety margin ratio (e.g., reserved power) is configured by the policy. This means that each drone must meet the requirements.

[0047] If the feasibility verification fails, the task scheduling and allocation will be readjusted. In this embodiment, this includes reducing low-yield tasks, adjusting allocations or changing waypoints, until the mission is initiated and executed by the UAV after the verification is passed.

[0048] After a drone completes a mission, its execution is monitored, generating four types of events: new alarms, new clues discovered, communication degradation, and resource changes. Then, it is determined whether the drone needs to trigger replanning. If so, the drone is frozen or interrupted, then task queueing and reallocation are performed, the flight path and schedule are recalculated, and the mission is restarted. This ensures continuous optimization of recovery time even amidst uncertain power grid information and dynamic environmental changes.

[0049] When changes in candidate segment confidence, risk, or resource redundancy exceed a threshold, the following formula is used to determine whether replanning is necessary: .

[0050] in This is an indicator variable for whether a replanning process is triggered. The largest change in confidence level of the candidate segment (from a new alarm or a new clue). The confidence level change threshold. The maximum change in risk cost (communication degradation / weather change). For risk change threshold, This represents the change in battery level / battery remaining capacity (resource changes). This is the threshold for resource changes.

[0051] If the calculations indicate that the UAV needs to be replanned, the process involves freezing, reordering, reallocating, and recalculating flight routes and schedules for the UAV. This ensures that the new task and waypoint sequences re-enter the isomorphic solution output chain and are re-issued for execution. After completing the critical tasks, the fault evidence and location results detected by the UAV are written back to the topology segment / tower identifier, outputting segment-level conclusions for isolation decisions, emergency repair dispatch, and power restoration confirmation. Simultaneously, emergency repair feedback is received as data input for the next round of evaluation and planning, achieving a closed-loop process for emergency response.

[0052] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments for application in other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for UAV mission orchestration and optimal path planning for power distribution line fault scenarios, characterized in that, The steps are as follows: (1) Obtain multi-source data, and after standardization, map the data uniformly onto the topological segment set S, where set S is denoted as... ,in, Indicates the first Each section Number of segments; (2) For each segment Calculate the confidence level of the fault Weighting of power outage impact The failure probability of each segment is evaluated to obtain a set of candidate failure segments; (3) For each candidate fault segment, generate at least one type of task unit to form a task unit set; (4) Construct a comprehensive objective function with the shortest power restoration time or the minimum power outage impact as the core objective; (5) Assign a task sequence to each UAV, calculate the estimated arrival time and estimated completion time, and perform constraint verification on the planning results of each UAV; (6) The UAV executes the launched task and transmits the status and evidence information back. The execution monitor continuously detects four types of events: new alarms, new clues, communication deterioration, and resource changes, and determines whether the UAV needs to trigger rolling replanning. (7) Freeze completed or uninterruptible tasks, and reorder and multi-aircraft redistribute unfinished tasks; re-execute steps (3) to (5) to generate new task sequences, waypoint sequences and ETA / ETC timetables, and reissue them for execution after feasibility verification; (8) For UAVs that have completed the mission, the fault evidence and location results detected by the UAVs are written back to the topology section or tower identifier, and the section-level conclusions for isolation decision-making, emergency repair dispatch and power restoration confirmation are output; at the same time, emergency repair feedback is received as data input for the next round of evaluation and planning, so as to realize the closed loop of the entire emergency response process.

2. The UAV mission orchestration and optimal path planning method for power distribution line fault scenarios according to claim 1, characterized in that, The multi-source data acquisition platform in step (1) includes an automation system, a power outage management system, an alarm parsing service, a 95598 repair reporting system, a hidden danger database, a UAV ground station, a weather station, and a communication assessment module; The standardization process includes coordinate transformation, topology consistency verification, time alignment, and spatial index construction.

3. The UAV mission orchestration and optimal path planning method for power distribution line fault scenarios according to claim 1, characterized in that, Fault confidence in step (2) The calculation formula is as follows: ， in, For normalization function, For alarm intensity characteristics, The characteristics of the intensity of the reported repair clusters. These are historical hidden dangers or a priori characteristics of malfunctions. The fusion weights configured for system parameters; Power outage impact weight The calculation formula is as follows: ， in, For use in subsequent optimization objectives. Influence weight, For the affected section The set of affected users is obtained through topological tracing. For users Importance coefficient.

4. The UAV mission orchestration and optimal path planning method for power distribution line fault scenarios according to claim 1, characterized in that, The set of task units in step (3) is denoted as: ,in Let t1, t2, ..., t be the set of task units. m It represents different task units, including fields such as task type, associated segment, entry waypoint, estimated execution time, payload requirements, and communication requirements; The number of task units; the task includes primary tasks, secondary tasks and support tasks. The primary tasks are rapid patrols along the section corridor; the secondary tasks hover or fly around suspicious points to collect evidence; the support tasks are generated under conditions of abnormal weather, night, or obstructed repairs.

5. The UAV mission orchestration and optimal path planning method for power distribution line fault scenarios according to claim 1, characterized in that, In step (4), the process of constructing a comprehensive objective function with the shortest power restoration time or the minimum power outage impact as the core objective is as follows: First, define the recovery time cost: ， in In terms of recovery time cost, a smaller value indicates faster recovery of critical sections. For section Expected recovery time or recovery time Number of segments; The formula for calculating the comprehensive objective function is: ， in, For the overall optimization objective, This refers to the flight and execution time cost obtained by summing the flight time between waypoints and the mission execution time. In order to obtain the risk cost through risk assessment, The communication penalty term is obtained from the communication evaluation. The trade-off coefficients configured for system strategies.

Images