Multi-unmanned aerial vehicle communication perception cooperation method in power grid fault scene

Abstract

The invention provides a multi-unmanned aerial vehicle communication sensing cooperation method in a power grid fault scene, and the method combines unmanned aerial vehicle dynamic topology reconstruction and a multi-stage optimization strategy, and achieves efficient information acquisition for emergency scenes of damaged ground base stations and large-area power failure in disaster areas after a power grid fault. Power flow entropy is introduced to analyze power grid node vulnerability, communication unmanned aerial vehicle deployment is optimized, and an optimal communication coverage strategy is solved based on a multi-target grey wolf optimization algorithm; meanwhile, a graph theory method and a shortest path search algorithm Dijkstra are used for optimizing the inspection path of the unmanned aerial vehicle, so that the fault diagnosis efficiency is improved; under a two-stage optimization framework, in the first stage, important load nodes are preferentially covered and routing inspection is executed, in the second stage, supplementary optimization is carried out on an uncovered area, and finally a global optimal solution considering unmanned aerial vehicle resource efficiency, communication quality and information acquisition is obtained. The method improves the recovery efficiency of a post-disaster power grid, and is suitable for the emergency response of a power system in an extreme disaster environment.

Description

Technical Field

[0001] The present invention relates to the technical field of UAV emergency communication, and particularly to a multi-UAV communication perception collaboration method in the scenario of power grid faults. Background Art

[0002] Modern distribution systems have evolved into a highly automated cyber-physical coupling system, where the communication system plays a key role in power system dispatching and distribution automation.

[0003] In extreme situations such as natural disasters, the communication systems of distribution automation power equipment are often interrupted due to the damage of ground infrastructure, resulting in the failure of status perception and information interaction, seriously affecting the repair efficiency and power grid restoration. Therefore, in addition to the physical network resilience, the resilience of the communication system should also be considered. The technology of using UAVs as temporary communication base stations for communication relay has been applied at present. Due to its characteristics of strong scheduling real-time performance, dynamic network formation, and flexible and unrestricted coverage area, it has special application value in power system emergency communication.

[0004] Traditional power system emergency repair methods have great limitations in extreme events. They mainly rely on ground vehicles for communication relay and repair. However, this method has high requirements for ground traffic conditions and is often difficult to implement in the harsh ground environment after extreme events. At the same time, before the repair is completed, the dispatching center is difficult to obtain the status information of power equipment in the affected area, and it is difficult to make timely dispatching and decisions on the power grid in the affected area, and the resilience and robustness of the power grid cannot be guaranteed. Therefore, it is necessary to study an emergency communication restoration mechanism for distribution automation equipment of power systems based on UAV communication-perception collaboration to ensure the rapid restoration of power grid communication. Summary of the Invention

[0005] For this reason, the present invention proposes a multi-UAV communication perception collaboration method in the scenario of power grid faults. Through dynamic topology reconstruction, the rapid reconstruction of the disaster area network is realized and timely decision-making responses are made to overcome the defect of insufficient topology reconstruction ability of the existing power emergency communication system in extreme disasters. The technical problems to be solved by the present invention are realized through the following technical solutions:

[0006] The present invention proposes a multi-UAV communication perception collaboration method in the scenario of power grid faults. The core idea is: considering the limited number of UAVs constraint, a multi-stage model is designed, and all UAVs start from the dispatching center. In the first stage, communication coverage and inspection are preferentially carried out for important load nodes. In the second stage, based on the deployment in the first stage, the remaining areas are investigated.

[0007] To achieve the above object, the present invention adopts the following technical solutions:

[0008] S1 Scenario setup. Considering the damage of the power grid and the interruption of base station communication in urban disasters, and due to the relatively high response delay of satellite communication, a multi-service collaboration scheme for UAV communication and sensing is designed. Normal communication devices directly transmit their status back to the ground station through UAVs, while devices with damaged communication functions need to be inspected. The required core parameters include the disaster area range, the coordinates of the dispatching center, the power grid topology and line status, the channel carrier frequency, bandwidth, environmental factor, as well as the initial position, quantity, and battery capacity of UAVs.

[0009] S2 Vulnerable node analysis. To accelerate the restoration of power supply, the vulnerability of power grid nodes is quantified by the power flow entropy method, and their importance is ranked to determine the maintenance priority. Nodes with high vulnerability have a greater impact on system stability and need to be diagnosed first. Based on this, the deployment of communication UAVs is optimized.

[0010] S3 Communication model and optimization. A three-dimensional airspace communication model is constructed to analyze the impact of factors such as path loss, signal-to-noise ratio, and channel capacity on communication quality, and to reveal the coupling relationship between the altitude and coverage radius of UAVs. Based on the spatio-temporal requirements of communication restoration, a multi-objective optimization model is established, comprehensively considering coverage weight, communication capacity, and energy consumption constraints, and the MOGWO algorithm is used to solve it.

[0011] S4 Graph theory modeling and path optimization of inspection routes. After ensuring communication coverage, UAVs need to inspect nodes with unknown status. Based on graph theory, an expression function for the UAV inspection route is given, and the weight matrix of the edges is established through communication positions and moving distances. To quickly solve the optimal inspection sequence, the Dijkstra algorithm is used to implement the shortest weighted path search, supporting dynamic adjustment of the weight ratio coefficient to meet the requirements of different task stages. Thus, one stage of the task is completed.

[0012] S5 Cooperative optimization of communication deployment and inspection routes in the second stage. After completing the tasks in the first stage, the inspection UAVs will return to the dispatching center to prepare for the second inspection, while the communication UAVs remain hovering. In the second stage, it is necessary to consider using as few communication UAVs as possible to complete the remaining tasks and consider moving the positions of UAVs based on the deployment in the first stage. The communication models in the two stages are the same, both considering using the MOGWO algorithm to solve, but the two stages have different objective functions at the level of communication UAV deployment. The first stage emphasizes priority, and the second stage emphasizes completion. The inspection models in the two stages are the same, both considering using the Dijkstra algorithm to solve, but the differences in communication layouts between the two stages will cause changes in the UAV inspection routes.

[0013] Compared with the prior art, the present invention has the following beneficial effects:

[0014] Aiming at emergency scenarios and damaged ground base stations, the present invention proposes a UAV task allocation mechanism based on communication-perception dual-mode collaboration, and establishes a step-by-step collaborative system that prioritizes fault location and then transmits the damaged status through state perception. It provides a strategy for the rapid acquisition of post-disaster damage conditions and provides support for subsequent post-disaster resource scheduling; the existing UAV path planning in power emergency scenarios mainly optimizes the physical movement trajectory and does not incorporate communication link requirements. The present invention proposes a coupling mechanism for grid node vulnerability assessment and communication quality parameters to construct a joint optimization framework for power equipment risk level and UAV communication resource configuration; in view of the constraint on the number of UAVs, a two-stage deployment strategy is constructed. UAVs are preferentially allocated to strengthen inspections in key areas, and then resources are reallocated to investigate non-critical areas, forming a global post-disaster information collection optimization plan under resource constraints. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for the embodiments are briefly introduced below.

[0016] Figure 1 A flow chart of a multi-UAV communication perception and coordination method in a power grid failure scenario provided by an embodiment of the present invention;

[0017] Figure 2 A schematic diagram of a structure for using a drone to assist in collecting power grid information in an urban disaster scenario provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to make the implementation objectives, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be described in detail and completely in conjunction with the accompanying drawings of implementation examples of the present invention. However, the technical solutions proposed by the present invention are not limited to the given examples.

[0019] like Figure 1 As shown, the present invention provides a multi-UAV communication perception coordination method in a power grid failure scenario. In the first phase, communication coverage and inspection are prioritized for important load nodes. In the second phase, based on the deployment in the first phase, the position of the UAV is adjusted to inspect the remaining areas. It is suitable for rapid inspection of post-disaster power restoration and mainly includes five steps:

[0020] Step S1, application scenario construction and basic parameter setting.

[0021] Considering the disaster situation in the urban scene, satellite links are not selected because satellite communications cannot meet the requirements of rapid response. In order to ensure the rapid acquisition of disaster area information, communication-perception multi-service drones are used for collaborative cooperation. The disaster area is set to S D. Given the power grid topology map as S = <N, E>, where N represents the number of nodes and E represents the set of node edges. When deploying the UAVs, considering the target users as the power grid nodes, the distance in the topology map is given according to the corresponding ratio of the actual map. Given the specific information of the lines before the disaster, such as factors like node voltage, phase angle, line current, etc., for power flow calculation. The dispatching center is located at Figure 2 the position shown, which is mainly responsible for collecting the data collected by the UAVs and can provide instructions for future emergency repairs, denoted as CC. The initial positions of the UAVs are all at the dispatching center. Considering the position deployment for signal coverage, the total power limit of each UAV is W E , mainly considering communication energy consumption and hovering energy consumption. For the inspection UAV, its inspection distance and the interval from the communication UAV are mainly considered. There are a total of M ALL UAVs and one inspection UAV. The parameters of the UAVs include the carrier frequency f, environmental coefficients a, b, and the bandwidth of each UAV is set to B. The signal transmission of each UAV needs to consider signal relaying, and another UAV is required within the communication range of the UAV to act as a communication bridge.

[0022] Step S2. To improve the efficiency of post-disaster power supply restoration, a priority ranking mechanism driven by node vulnerability is established. The failure of such nodes will significantly affect the balance of the power grid power flow distribution. Based on the power grid topology structure and historical line data, a power flow entropy model is constructed to quantify the impact of nodes on system stability and perform ranking. The priority ranking result provides a decision-making basis for the deployment location and resource allocation of communication UAVs, ensuring that vulnerable areas are given priority for communication guarantee and inspection perception.

[0023] S2.1 First, it is necessary to input the power grid data for basic power flow calculation to obtain the active power flow P of the branch i0 (i = 1, 2,..., m), where m represents the total number of branches. Further, considering that a certain node a is disturbed, the power flow increment ΔE caused by each branch after the disturbance is calculated in turn ia = P ia - P i0 , where P ia is the power flow of the i-th branch after node a is disturbed. Finally, the disturbance of the node change to the system is calculated and the normalization processing η of the power flow increment is performed ia = ΔE ia / ΔE a . Thus, the power flow distribution entropy of node a is defined as:

[0024]

[0025] Among them, represents the power flow distribution entropy of node a, and N represents the target node. The magnitude reflects the impact on the system under overload disturbances of different nodes. When the η ia values of each branch are the same, at this time, all branches share the impact brought by the node disturbance, making the system more stable. If a single line bears this disturbance alone, then at this time, the node disturbance will have a huge impact on the system and easily lead to out-of-limit faults in the branches.

[0026] S2.3 Further, define the node vulnerability index to judge the importance of nodes:

[0027] In the formula, ΔE a represents the impact of the disturbance of node a on the entire system. The larger its value, the higher the corresponding node vulnerability index. represents the node power flow entropy. The smaller its value indicates that after node a is disturbed, the system is more unstable. Generally speaking, V a expresses the importance of node a. The higher the value, the more likely this node is to cause the system to be disordered and needs to be prioritized for inspection.

[0028] S2.4 Further, calculate V i for all nodes, i = 1, 2, … N, and assign the vulnerability index to the corresponding nodes to form a new network topology S' = <N, E, V>. Among them, V corresponds to the vulnerability index of each node. The larger its value, the greater the impact of the change of node i on the system, that is, the higher the importance.

[0029] Step S3, first build the communication model, and then consider the optimization objective modeling and solution. There are two communication models in total: for the communication link between the UAV and the ground control center, due to the high probability of NLOS propagation caused by urban building occlusion, the LOS / NLOS hybrid channel model is used to evaluate the link reliability; while the air-to-air communication between UAVs is applicable to the free space propagation model to simplify the link characteristic analysis due to the open environment.

[0030] S3.1 When the communication target is the control center:

[0031]

[0032]

[0033]

[0034]

[0035] Among them, represents the probability that the communication path between UAV M and the control center CC is LOS, Indicates the probability that the communication path between the UAV M and the dispatching center CC is NLOS. a and b are environmental coefficients. θ represents the angle between the UAV and the ground terminal. is the hovering height of the UAV, represents the two-dimensional position of the UAV. And X CC is used to express the two-dimensional position of the dispatching center. Equation (5) represents the distance between the UAV M and the dispatching center CC.

[0036] S3.2 Further, calculate the communication losses for LOS and NLOS paths and obtain the average path loss, signal-to-noise ratio, and channel capacity.

[0037]

[0038]

[0039]

[0040]

[0041]

[0042] R M,CC = Blog 2 (1 + SNR M,CC )(11);

[0043] Among them, represents the communication loss for the LOS path between the UAV M and the dispatching center CC. represents the communication loss for the NLOS path between the UAV M and the dispatching center CC. c represents the speed of light, and ψ NLOS represents the additional loss caused by the NLOS path. Equation (8) gives the average path loss value between the UAV M and the dispatching center CC. Equation (9) gives the signal-to-noise ratio between the UAV M and the dispatching center CC. Among them, W M represents the communication power of the UAV, and σ 2 represents the noise power. And Equation (10) gives the hovering mechanical power, where represents the hovering consumption when the UAV height is H 0 at that time, is the actual hovering position of the UAV M. Γ > 0 is the motor speed multiplier. Equation (11) calculates the channel capacity between the UAV M and the dispatching center CC.

[0044] S3.3 Further, build the communication model between communication UAVs. When the communication target of the UAV is M', there is:

[0045]

[0046]

[0047]

[0048] R M,M' = Blog 2 (1 + SNR M,M' )(15);

[0049] Equation (12) represents the distance between UAV M and UAV M'. Equation (13) gives the communication loss between UAV M and M'. Equation (14) gives the calculation formula for the signal-to-noise ratio between UAV M and UAV M'. Equation (15) represents the channel capacity between UAV M and UAV M'. B is the UAV broadband.

[0050] S3.4 Further, build a model for the ground coverage rate. Set the field of view angle of the UAV to α, and thus obtain the coverage radius as When the target node satisfies , then it is considered that the node is covered by the UAV, and it is denoted as x i = 1, otherwise 0.

[0051] Further, establish an optimization model for the coverage of passing UAVs. This design considers designating the corresponding links of adjacent UAVs for communication relay to ensure stable data transmission. Denote the connection variable between UAVs as x M,M' ∈ {0, 1}. If a communication link connection is selected between UAV M and UAV M', then x M,M' = 1. Due to task constraints, at least two UAVs are required to connect to the dispatching center, and its variable is denoted as x M,CC ∈ {0, 1}. From this, the link matrix M C between UAVs and between UAVs and the dispatching center can be obtained, which can be expressed as:

[0052]

[0053] Among them, the diagonal elements represent self-loops and are all set to 0. x CC,1 represents the variable of the connection between the dispatching center and UAV 1, and the meanings of the remaining similar variables are the same. And x 1,M represents the connection between UAV 1 and UAV M. This matrix is a symmetric matrix. This matrix is mainly used to clarify the connection relationship and its variable values.

[0054] S3.5 Further, according to the obtained sorting of node vulnerability indicators, preferentially provide communication coverage for important nodes, and the coverage utility function can be expressed as:

[0055]

[0056] Furthermore, in addition to ensuring priority coverage, it is also necessary to guarantee the channel capacity of the signal. After traversing the entire communication process, the utility function of the average channel capacity can be obtained:

[0057]

[0058] Equation (18) represents the average channel capacity, including the channel capacities between UAVs and between UAVs and the dispatching center. Due to the symmetry of the variables, the corresponding capacities will be calculated twice during the calculation process, so a coefficient of 2 needs to be given in the denominator.

[0059] Furthermore, although increasing the transmission power of UAVs can enhance communication quality, it will lead to a rapid decay of the battery life. To consider the balance between power and efficiency, the power utility function of a single UAV and the overall average power utility function are defined:

[0060]

[0061]

[0062] Equation (19) represents the power utility of a single UAV. The denominator includes the power consumption of communication between UAVs, communication between UAVs and the dispatching center, and UAV hovering. Equation (20) represents the overall average power utility. Due to the symmetry of the variables, the corresponding capacities will be calculated twice during the calculation process, so a coefficient of 2 needs to be given in the denominator.

[0063] Furthermore, the optimization objective function for the static deployment in the first stage of communication is constructed as follows:

[0064] maxμ 1 ·U D +μ 2 ·U R +μ 3 ·U W (21);

[0065]

[0066] The μ 1 , μ 2 , μ 3 in the objective function are coefficients for balancing various indicators. In the constraints, represents that the safety distance between UAVs must be greater than the set value D A ; D A ≥D M means that the safety distance must be greater than the coverage radius; otherwise, there will be a situation of overlapping coverage; means that each UAV selects at most one UAV for communication relay; x M,M' ·d M,M'≤R means that the distance of the UAV cannot exceed the maximum direct communication distance R, otherwise, the signal will be interrupted; means that at least two UAVs need to be connected to the dispatching center CC, but the total number of connections will not exceed the total number of UAVs; x M,CC ·d M,CC ≤R means that the communication distance between the UAV and the dispatching center CC cannot exceed the maximum direct communication distance R; for this constraint, let Q be a positive real number. When the left side < 0, it means that the coverage target is not within the coverage radius. To ensure the inequality holds, the right side needs to satisfy x i = 0 to meet the constraint. On the contrary, if the left side ≥ 0, it proves that the target is within the coverage radius, and there is x i = 1; represents the upper and lower limits of the UAV's height; X M ∈X map represents the boundary constraint of the UAV, which must be within the disaster area; U M ≥1 means that the power of the UAV cannot exceed its set maximum value.

[0067] S3.6 Further, the present invention is solved by using the multi-objective grey wolf optimization algorithm. The MOGWO (Multi-Objective Grey Wolf Optimization) algorithm is a multi-objective optimization method based on the swarm intelligence behavior of grey wolves. By simulating the social hierarchy and cooperative hunting mechanism of the grey wolf population, it realizes the collaborative optimization of multiple conflicting objectives. After initializing the grey wolf population, this algorithm uses the three-level leaders (non-dominated solutions) of α, β, and δ to guide other individuals to approach the Pareto front, balances global exploration and local exploitation by dynamically adjusting the search weights, and maintains an elite archive during the iteration process to retain diverse non-dominated solutions. Finally, it outputs the Pareto solution set that optimally balances multiple objectives, and is applicable to complex scenarios such as UAV path planning that need to consider energy consumption, efficiency, and safety. Its specific steps are designed as follows:

[0068] First, initialize the grey wolf population Y i =(y i1 ,y i2 ,…y iD ), i = 1, 2, …, n, where n represents the population size and D represents the decision variable dimension. Parameter initialization: where linearly decreases from 2 to 0 during the iteration process. where is a random vector within [0, 1]. And set the maximum number of iterations L T , and save the initial non-dominated solution.

[0069] Furthermore, calculate the objective function value of each grey wolf individual. where f 1 , f 2 , f 3 correspond to the three optimization indicators in formula (21) respectively.

[0070] Furthermore, perform non-dominated solution archiving. Filter out the non-dominated solutions in the current population through non-dominated sorting, and update the non-dominated solutions stored in the elite archive (Archive). Its archive space will be divided into multiple grids, and leaders are preferentially selected from the least crowded grids. The α wolf (optimal solution), β wolf (sub-optimal solution), and δ wolf (third-best solution) are selected through the roulette wheel strategy. Each ω wolf (the bottom layer of the wolf pack, moving following the leader) changes its position according to the leader. First, calculate the distance:

[0071]

[0072]

[0073] Furthermore, update the specific position:

[0074]

[0075] Furthermore, linearly decrease, dynamically adjust. Repeat the above process until the maximum number of iterations is reached, and output the Pareto optimal solution.

[0076] Step S4, patrol path graph theory modeling and path optimization. In the case where the first-stage position deployment has been obtained and the key area has been covered by communication, if the electrical equipment also has communication functions, the status information of the target equipment can be directly obtained. If the electrical equipment cannot transmit information through the drone, it means that this area is in an unknown state, and it is necessary to send a patrol drone to check the specific situation. Consider using the method of graph theory to plan the patrol path and use the Dijkstra algorithm to solve for the optimal path.

[0077] S4.1 Take the fault point results after position deployment as the input of the patrol target, and propose the weight matrix G(i,j) (i = 1,..., K + 1, j = 1,..., K + 1) of the Dijkstra algorithm. These node positions correspond to the node coordinate positions that still need to be patrolled in the graph. Its expression is:

[0078]

[0079] where L com represents the distance between the patrol drone and the communication relay drone, and L movRepresents the inspection distance between nodes of the inspection UAV. The distance length can reflect the energy consumption of the UAV from the side, d 1 , d 2 is the corresponding weight ratio. Generally speaking, the propulsion energy consumption is much greater than the communication energy consumption, so it can be considered that d 2 = td 1 , where t is a positive real number. Since the main purpose of the inspection UAV is to check the specific situation, its flight altitude is generally low and the change range is not large. To prevent collisions between UAVs, consider its altitude to be constant at H r , and only consider its weight from the two-dimensional level. Set the starting point and the end point of the inspection UAV's mission to be the dispatching center, denote its node number as 1, and require it to pass through all the maintenance points. Each maintenance node can be regarded as a starting or leaving point during the inspection process.

[0080] S4.2 Define the specific expressions of L com and L mov for updating the weight. There are a total of M communication UAVs, and the faulty nodes will definitely be covered by a corresponding UAV. The faulty points within the range can be classified as N l , l = 1, …, M. The inspection UAV selects the UAV closest to it for communication. The relevant distance can be calculated as:

[0081]

[0082] where are the two-dimensional coordinates of the nodes of UAV l, i, and j respectively, and are all two-dimensional vectors. And X CC is used to express the two-dimensional position of the dispatching center, and is also a two-dimensional vector.

[0083] S4.3 Further, start the path search. First, traverse all nodes to construct the weight matrix. For each pair of nodes i and j, update the weight matrix G(i, j) using formula (25). If i and j are the same node, the weight is 0; otherwise, the matrix weight is the weighted sum of the communication distance and the moving distance.

[0084] Further, starting from node 1 of the dispatching center, traverse all j nodes, and the corresponding path cost is G(1, :). Then for each node i, its G(i, :) can be obtained. For each calculated node, consider using the Dijkstra algorithm to consider the obtained weight matrix. It is actually an algorithm for finding the shortest path from a single source point to all other nodes in a non-negative weight graph by gradually determining the current shortest path node, and each step selects the current local optimum. Set that the operation process must traverse all nodes to stop, and the ultimate goal is the dispatching center.

[0085] Further, output the shortest path of the inspection UAV.

[0086] Step S5, collaborative optimization of two-stage communication deployment and inspection path. So far, the first-stage UAV deployment has been completed. However, considering the UAV quantity limit, there are still some nodes not detected. Therefore, second-stage deployment optimization is required. When the inspection UAV returns to the dispatching center again, consider replenishing its power. For the remaining areas to be detected, it is still preferred to first make a preliminary judgment by relying on UAV communication and then consider dispatching UAVs for specific diagnosis. Since there are fewer points to be detected at this time, not all UAVs are needed to execute tasks, and only the position deployment of UAVs needs to be changed based on the original layout, while the remaining UAVs can choose to return to the dispatching center.

[0087] S5.1 First, determine the basic deployment through the objective function. At this time, the objective function will change. Instead of considering coverage priority, as few UAVs as possible are used to ensure the completion of all tasks. Record the values of x i that are 1 in the first stage, exclude them from the coverage target, and calculate the number of target in the remaining area as X R , and at this time the objective function can be rewritten as:

[0088] max - λ 1 N U +λ 2 ·U R +λ 3 ·U W (28);

[0089] Among them, λ 1 , λ 2 , λ 3 are the corresponding weight coefficients. On the basis of the original constraints, it is also necessary to add: to ensure that all targets are covered, 1 ≤ N U ≤ M to ensure that the UAV quantity does not exceed the limit. Still consider using the multi-objective grey wolf optimization algorithm in S3 for solution.

[0090] S5.2 Further, after obtaining the optimized positions in the second stage, according to the coverage deployment map in the first stage, traverse the relative distances between all UAVs in the first stage and those in the second stage, find the minimum value among them to form the moving pair (O, D), and adjust the deployment in the first stage. If there are still redundant UAVs, choose to return to the dispatching center.

[0091] S5.3 Further, after determining the position deployment in the second stage, consider using the algorithm in step S4 again to find the shortest inspection path, and finally realize the post-disaster inspection work for the entire area.

[0092] In view of the urban disaster scenario where ground base stations are damaged and there is a large-scale power outage in the disaster area, the present invention proposes a multi-UAV communication and sensing cooperation method under the power grid fault scenario, which relates to the field of UAV emergency communication technology. In view of the emergency scenario and the damaged situation of ground base stations, the present invention proposes a UAV task allocation mechanism based on communication-sensing dual-mode cooperation, and establishes a step-by-step cooperation system that first locates faults and then transmits the damaged state back through state sensing. It provides a strategy for quickly obtaining the damaged situation after the disaster and supports subsequent post-disaster resource scheduling; the existing UAV path planning in the power emergency scenario mainly optimizes the physical movement trajectory and does not incorporate the communication link requirements. The present invention proposes a coupling mechanism between the vulnerability assessment of power grid nodes and communication quality parameters, and constructs a joint optimization framework for the risk level of power equipment and the allocation of UAV communication resources; in view of the UAV quantity constraint, a two-stage deployment strategy is constructed. First, UAVs are preferentially allocated to strengthen the inspection of key areas, and then resources are reallocated to investigate non-key areas, forming an optimized global post-disaster information collection plan under resource constraints.

[0093] The above content is a further detailed description of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is only limited to these descriptions. For those of ordinary skill in the technical field to which the present invention belongs, without departing from the concept of the present invention, several simple deductions or substitutions can be made, and all should be regarded as belonging to the protection scope of the present invention.

Claims

1. A multi-UAV communication perception coordination method in a power grid fault scenario, characterized in that: include: S1 Scenario Construction; The present invention takes into account the damage to the power grid and the interruption of base station communication in urban disasters. Due to the high response delay of satellite communication, a UAV communication-perception multi-service collaborative solution is designed; the status of the equipment with normal communication is directly transmitted back by the UAV, while the equipment with damaged communication function needs to be inspected; the required core parameters include the scope of the disaster area, the coordinates of the dispatching center, the topology and line status of the power grid, the channel carrier frequency, bandwidth, environmental coefficient, and the initial position, number and battery capacity of the UAV; S2 Vulnerable node analysis: To speed up power supply restoration, the vulnerability of power grid nodes is quantified through the power flow entropy method, and their importance is ranked to determine the maintenance priority; nodes with high vulnerability have a greater impact on system stability and need to be diagnosed first, and the deployment of communication drones is optimized accordingly; S3 communication model and optimization; construct a three-dimensional airspace communication model, analyze the impact of factors such as path loss, signal-to-noise ratio, channel capacity, etc. on communication quality, and reveal the coupling relationship between drone height and coverage radius; based on the spatiotemporal requirements of communication recovery, establish a multi-objective optimization model, comprehensively consider coverage weight, communication capacity and energy consumption constraints, and use the MOGWO algorithm to solve it; S4 Inspection path graph theory modeling and path optimization; After ensuring communication coverage, it is necessary to conduct drone inspections on nodes in unknown states; Based on the graph theory method, an expression function for the drone inspection path is given, and the edge weight matrix is ​​established through the communication position and the moving distance; In order to quickly solve the optimal inspection sequence, the Dijkstra algorithm is used to realize the shortest weighted path search, and the weight ratio coefficient can be dynamically adjusted to meet the needs of different task stages; This completes the first stage of the task; S5 The communication deployment and inspection path of the second stage are coordinated and optimized; after completing the task in the first stage, the inspection drone will return to the dispatch center and prepare for the second inspection, while the communication drone will remain in a hovering state; in the second stage, it is necessary to consider using as few communication drones as possible to complete the remaining tasks, and consider moving the drone position based on the deployment in the first stage; the communication models of the two stages are consistent, and both consider using the MOGWO algorithm for solution, but the two stages have different objective functions at the communication drone deployment level, with the first stage emphasizing priority and the second stage emphasizing completion; the inspection models of the two stages are consistent, and both consider using the Dijkstra algorithm for solution, but the difference in communication layout between the two stages will cause the drone inspection path to change.

2. According to the method of multi-UAV communication perception coordination in power grid failure scenario in claim 1, it is characterized in that: The scenarios described in S1 are as follows: considering the disaster situation in the urban scenario, since satellite communication cannot meet the rapid response requirements, satellite links are not selected; to ensure the rapid acquisition of disaster area information, communication-perception multi-service drones are used for collaborative cooperation; the disaster area is set to S D ; The grid topology is given as S = <N, E>, where N represents the number of nodes, E represents the node edge set, and the dispatch center is recorded as CC; the initial position of the drone is located at the dispatch center, and the position deployment is considered for signal coverage. The total power limit of each drone is W E ; There are M ALL drones and an inspection drone; The parameters of the drone include carrier frequency f, environmental coefficients a and b, and the bandwidth of each drone is set to B; the signal transmission of each drone needs to consider signal relay, and another drone is needed within the communication range of the drone to serve as a communication bridge.

3. The method for multi-UAV communication perception coordination in a power grid fault scenario according to claim 2, characterized in that S2 include: S2.1 Input the grid data to perform basic power flow calculation and obtain the active power flow P of the branch. i0 (i=1,2,...,m), where m represents the total number of branches; consider a node a is disturbed, and calculate the flow increment ΔE caused by each branch after the disturbance in turn ia =P ia -P i0 , where P ia is the power flow of branch i after node a is disturbed, and finally the disturbance of the node change to the system is calculated And normalize the power flow increment η ia =ΔE ia / ΔE a ; The power flow distribution entropy of node a is defined accordingly; in, represents the power flow distribution entropy of node a, and N represents the target node; The size of reflects the impact of the system under the overload disturbance of different nodes; S2.2 Definition of node vulnerability indicators In the formula, ΔE a It indicates the impact of the disturbance of node a on the entire system. The larger its value is, the higher the vulnerability index of the corresponding node is. It represents the node power flow entropy. The smaller its value is, the more unstable the system is after node a is disturbed. In general, V a It expresses the importance of node a. The higher the value, the more likely the node is to cause system disorder and needs to be inspected first. S2.4 Calculate V of all nodes i ,i=1,2,...N, and assign the vulnerability index to the corresponding node to form a new network topology S'=<N,E,V>.

4. The method for multi-UAV communication perception coordination in a power grid fault scenario according to claim 3, characterized in that S3 include: S3.1 When the communication target is the dispatch center; represents the probability that the communication path between UAV M and the dispatch center CC is LOS, represents the probability that the communication path between the UAV M and the dispatch center CC is NLOS; a and b are environmental coefficients; θ represents the angle between the UAV and the ground terminal; is the drone's hovering height, represents the two-dimensional position of the drone; and X CC Used to express the two-dimensional position of the dispatch center; Formula (5) represents the distance between the UAV M and the dispatch center CC; S3.2 calculate the communication loss on the LOS and NLOS paths and derive the average path loss, signal-to-noise ratio, and channel capacity; R M,CC =Blog2(1+SNR M,CC )(11); It represents the LOS path communication loss between the UAV M and the dispatch center CC; represents the NLOS path communication loss between the UAV M and the dispatch center CC; c represents the speed of light, ψ NLOS represents that the NLOS path will cause additional loss; Formula (8) gives the average path loss value between the drone M and the dispatch center CC; Formula (9) gives the signal-to-noise ratio between the drone M and the dispatch center CC; where W M represents the UAV communication power, σ 2 represents the noise power; while formula (10) gives the hovering mechanical power, where Indicates the hovering consumption when the drone altitude is H0. is the actual hovering position of the UAV M; Γ>0 is the motor speed multiplier; Formula (11) calculates the channel capacity between the UAV M and the dispatch center CC; S3.3 builds the communication model between communication drones. When the communication target of the drone is M', there is; R M,M' =Blog2(1+SNR M,M' )(15); Formula (12) represents the distance between drone M and drone M'; Formula (13) gives the communication loss between drone M and drone M'; Formula (14) gives the signal-to-noise ratio calculation formula between drone M and drone M'; Formula (15) represents the channel capacity between drone M and drone M'; B is the drone bandwidth; S3.4 builds a model for ground coverage; the UAV’s field of view angle is set to α, and the coverage radius is obtained as When the target node satisfies When , the node is considered to be covered by the drone and is recorded as x i =1, otherwise 0; S3.5 The coverage utility function can be expressed as; Define the utility function of the average channel capacity; Formula (19) represents the average channel capacity, including the channel capacity between UAVs and between UAVs and the dispatch center; Define the power utility function of a single UAV and the overall average power utility function; Formula (19) represents the power utility of a single UAV. The denominator includes the power consumption of UAV-to-UAV communication, UAV-to-dispatch center communication, and UAV hovering. Formula (20) represents the overall average power utility. Due to the symmetry of the variables, the corresponding capacity is calculated twice during the calculation process, so a coefficient of 2 is required in the denominator. maxμ1·U D +μ2·U R +μ3·U W (20); x M,M' =x M',M ,x M,M =0,W M,M =0 x M,CC ·d M,CC ≤R μ1, μ2, and μ3 in the objective function are the coefficients of balancing various indicators; in the constraints, The safe distance between drones must be greater than the set value D A ;D A ≥D M This means that the safety distance must be greater than the coverage radius, otherwise repeated coverage will occur; Indicates that each drone selects at most one drone for communication relay; x M,M' ·d M,M' ≤R means the distance between drones cannot exceed the maximum direct communication distance R, otherwise the signal will be interrupted; Indicates that at least two drones need to connect to the dispatch center CC, but the total number of connections will not exceed the total number of drones; x M,CC ·d M,CC ≤R means that the distance between the UAV and the dispatch center CC cannot exceed the maximum direct communication distance R; For this constraint, Q is set to a positive real number. When the left-hand side is less than 0, it means that the coverage target is not within the coverage radius. To ensure that the inequality holds, the right-hand side needs to satisfy x i = 0 to satisfy the constraint; otherwise, if the left formula ≥ 0, it proves that the target is within the coverage radius, and there is x i =1; Indicates the upper and lower limits of the drone's altitude; X M ∈X map Indicates the boundary constraint of the drone, which must be within the disaster area; U M ≥1 means the drone’s power cannot exceed its set maximum value; S3.6 The present invention uses a multi-objective grey wolf optimization algorithm to solve the problem, and the specific steps are designed as follows: Initialize the gray wolf population Y i =(y i1 ,y i2 ,...y iD ), i = 1, 2, ..., n, where n represents the population size and D represents the dimension of the decision variable; the parameter initialization is obtained in It decreases linearly from 2 to 0 during the iteration process; in is a random vector in [0,1]; and sets the maximum number of iterations L T , save the initial non-dominated solution; Calculate the objective function value for each gray wolf individual Where f1, f2, and f3 correspond to the three optimization indicators of formula (21); Archive the non-dominated solutions; filter out the non-dominated solutions in the current population through non-dominated sorting, and update the non-dominated solutions stored in the elite archive (Archive); the archive space will be divided into multiple grids, and the leader will be selected from the loosest grid first, and the α wolf (optimal solution), β wolf (second-best solution), and δ wolf (third-best solution) will be selected through the roulette strategy; each ω wolf (the bottom wolf pack, moving with the leader) changes its position according to the leader; calculate the distance; Update the specific location; a W Linearly decrease, A and C are adjusted dynamically; repeat the above process until the iteration reaches the maximum number of iterations and outputs the Pareto optimal solution.

5. The method for multi-UAV communication perception coordination in a power grid fault scenario according to claim 4, characterized in that S4 include: S4.1 takes the fault point results after the position deployment as the input of the inspection target, and proposes the weight matrix G(i,j) (i=1,...,K+1,j=1,...,K+1) of the Dijkstra algorithm. These node positions correspond to the coordinate positions of the nodes that still need to be inspected in the figure; Its expression is; Where L com represents the distance between the inspection drone and the communication relay drone, and L mov Indicates the inspection distance between nodes of the inspection drone; the distance length can indirectly reflect the energy consumption of the drone, d1, d2 are the corresponding weight ratios; where d2 = td1, t is a positive real number; the initial point and the end point of the inspection drone are both set to the dispatch center, and its node code is recorded as 1; S4.2 Definition of L com and L mov The specific expression of is used to update the weight; there are M communication drones in total, and the faulty node will definitely be covered by a corresponding drone, and the faulty points within the range can be classified into N l ,l=1,...,M; the inspection UAV selects the UAV closest to it to communicate; the relevant distance can be calculated as; in are the two-dimensional coordinates of the drone nodes l, i, and j, respectively, and are all two-dimensional vectors; and X CC It is used to express the two-dimensional position of the dispatch center, which is also a two-dimensional vector; S4.3 starts the path search; first, traverse all nodes to construct the weight matrix, and for each pair of nodes i and j, use formula (25) to update the weight matrix G(i, j); If i and j are the same node, the weight is 0; otherwise, the matrix weight is the weighted sum of the communication distance and the moving distance; the Dijkstra algorithm is used to calculate the shortest path and output it.

6. The method for multi-UAV communication perception coordination in a power grid fault scenario according to claim 5, characterized in that S5 include: S5.1 Recording one stage x i If the value is 1, it is excluded from the coverage target, and the number of targets in the remaining area is calculated as X R , then the objective function can be rewritten as: max-λ1N U +λ2·U R +λ3·U W (27); Among them, λ1, λ2, λ3 are the corresponding weight coefficients; they need to be added to the original constraints; To ensure that all targets are covered, 1≤N U ≤M to ensure that the number of drones does not exceed the limit; the multi-objective gray wolf optimization algorithm in S3 is still considered for solution; S5.2 After obtaining the optimized position of the second stage, according to the coverage deployment diagram of the first stage, traverse the relative distances between all drones in the first stage and the drones in the second stage, find the minimum value to form a moving pair (O, D), and adjust the deployment of the first stage; if there are still extra drones left, choose to return to the dispatch center; S5.3 After determining the location deployment of the second phase, consider using the algorithm in step S4 again to find the shortest inspection path, and finally realize the post-disaster inspection work in the entire area.

Images