Power distribution network toughness improving method and system considering unmanned aerial vehicle deicing in ice disaster scene

Abstract

The invention discloses a power distribution network toughness improving method and system considering unmanned aerial vehicle deicing in an ice disaster scene, and belongs to the technical field of power systems, and the method comprises the steps: calculating the icing thickness and fault probability of each line through a pre-constructed icing thickness prediction model and a fault probability calculation model, and generating a seriously icing line set; calculating a multi-dimensional toughness evaluation index of the power distribution network in combination with a pre-constructed ice disaster scene power distribution network load change curve, and calculating a comprehensive toughness evaluation index; an optimal flight path and an operation time sequence of the unmanned aerial vehicle are generated, and the unmanned aerial vehicle is scheduled to execute deicing operation; generating a recovery strategy based on the real-time power distribution network state data and a pre-constructed dynamic recovery optimization model, updating the comprehensive toughness evaluation index, dynamically adjusting the weight coefficient distribution of the comprehensive toughness evaluation index based on an updating result, synchronously updating the target function weight coefficient of the dynamic recovery model, and iteratively executing the above steps to obtain the comprehensive toughness evaluation index. The comprehensive toughness evaluation index is continuously higher than the preset threshold value.

Description

Technical Field

[0001] The present invention belongs to the technical field of power systems, and more specifically, relates to a disaster recovery technology for using unmanned aerial vehicles (UAVs) for de-icing when lines are severely ice-covered in an extreme ice disaster scenario to eliminate fault risks and thus enhance the resilience of the distribution network. Background Art

[0002] In recent years, extreme weather has occurred frequently, and power grid disasters caused by severe climates have been intensifying. Among them, ice disasters, as representatives of typical extreme disasters, long-term icing can trigger accidents such as line breaks, insulator flashovers, and pole tilting and collapse, which have a greater impact on the power system. When repairing lines under extreme ice disaster weather, power-off operations need to be carried out on severely damaged lines first to ensure that repair personnel are not at risk of personal injury, which seriously affects the de-icing efficiency of the repair team, increases the power outage time, and causes serious economic losses.

[0003] The prior art document (CN118693820A) discloses a method for enhancing the resilience of a distribution network with multi-resource coordination of the source, network, and load in an ice disaster scenario. By constructing a multi-resource coordinated scheduling model and a solution strategy based on the penalty-based Gauss-Seidel method, there is a problem of insufficient prediction accuracy of the ice-covered thickness in its line icing model as it does not consider the dynamic influence of line loss heating on the line temperature rise on the ice-covered thickness; its resilience assessment only uses the reduction of the load loss amount as a single index, and there is a problem that it cannot comprehensively represent the synergistic effect of the distribution network's defense ability and recovery efficiency; although it adopts a traditional repair team and multi-department coordinated scheduling strategy, there are problems of low de-icing efficiency during the disaster, complex post-disaster recovery operations, and high economic costs. Summary of the Invention

[0004] To solve the deficiencies in the prior art, the present invention provides a method for enhancing the resilience of a distribution network considering UAV de-icing in an extreme ice scenario. First, use UAVs for fault location, establish a calculation model for the ice-covered thickness under ice disasters, and find the lines with severe ice-covered thickness during the disaster; then, according to the characteristics of distribution network faults in the ice disaster scenario, formulate reasonable resilience assessment indicators. Finally, establish a UAV line de-icing model, including locating the severely ice-covered lines of the distribution network and dispatching UAVs for de-icing, to minimize load power-off to the greatest extent, thereby enhancing the resilience of the distribution network in the ice disaster scenario.

[0005] The present invention adopts the following technical solutions.

[0006] The first aspect of the present invention provides a method for enhancing the resilience of a distribution network considering UAV de-icing in an ice disaster scenario, including the following steps: Real-time collect the meteorological data, line operation parameters, and equipment attribute data of the distribution network in the ice disaster environment, calculate the ice-covered thickness and fault probability of each line through a pre-constructed ice-covered thickness prediction model and fault probability calculation model, and generate a set of severely ice-covered lines; Calculate the multi-dimensional resilience evaluation index of the distribution network in combination with the load change curve of the pre-built ice disaster scenario of the distribution network, and calculate the comprehensive resilience evaluation index. When the comprehensive resilience evaluation value is lower than the preset threshold, trigger the UAV collaborative de-icing instruction; Generate the optimal flight path and operation time sequence of the UAV through the pre-built UAV de-icing and restoration model, and dispatch the UAV to perform de-icing operations; During the UAV de-icing operation, generate a restoration strategy based on the real-time distribution network status data and the pre-built dynamic restoration optimization model, update the comprehensive resilience evaluation index according to the distribution network operation data after implementing the restoration strategy, and dynamically adjust the weight coefficient allocation of the comprehensive resilience evaluation index based on the update result. Synchronously update the objective function weight coefficient of the dynamic restoration model, and iteratively execute the above steps until the comprehensive resilience evaluation index is continuously higher than the preset threshold.

[0007] Optionally, the calculation of the ice coating thickness of each line by the pre-built ice coating thickness prediction model includes: Calculate the change in the surface temperature of the line based on the Joule heat effect generated by the line energization current and the ambient temperature; Determine the critical non-icing time before ice coating formation through the transient heat balance equation of the conductor surface during the ice disaster process. The transient heat balance equation of the line surface characterizes the dynamic change process of the line surface temperature during the ice disaster through the product of the specific heat capacity of the line material and the change rate of the line surface temperature, and includes the heat balance relationship of Joule heat generated by the line operation, convective heat transfer loss between the line and the environment, heat loss of supercooled water droplets heated to the condensation point, latent heat absorption of raindrop evaporation on the line surface, and heat radiation; Calculate the ice coating thickness growth in segments after the ice coating start time according to the real-time freezing rain intensity, wind speed, and air water content, where the ice coating thickness is zero before reaching the critical non-icing time, and the ice coating accumulation is dynamically calculated after exceeding the critical non-icing time.

[0008] Optionally, the generation of the set of severely ice-coated lines includes: Preset the ice coating thickness safety threshold and the failure probability safety threshold; Real-time monitor the ice coating thickness and failure probability of each line. When both the ice coating thickness and the failure probability exceed their safety thresholds, dynamically mark the corresponding line as a severely ice-coated line and add it to the set of severely ice-coated lines; Periodically remove the lines that meet the natural recovery conditions, where the natural recovery conditions include that the ice coating thickness is lower than the first recovery threshold and the failure probability is lower than the second recovery threshold.

[0009] Optionally, the calculation of the multi-dimensional resilience evaluation index of the distribution network includes: The total load recovery index is calculated based on the time integral ratio of the actual recovered load during the ice disaster period to the theoretical maximum load. The fault recovery time index is calculated based on the ratio of the time taken for fault recovery to the total duration of the ice disaster. The network loss index is calculated based on the time integral ratio of the actual line network loss to the reference network loss.

[0010] Optionally, generating the optimal flight path and operation timing sequence of the unmanned aerial vehicle (UAV) through a pre - constructed UAV ice removal and recovery model, and scheduling the UAV to perform ice removal operations includes: Obtaining the coordinate information of the set of severely ice - covered lines, the parameters of the UAV cluster, and the thermodynamic characteristic parameters. Calculating the ice removal energy consumption and operation time of each line through the pre - constructed UAV ice removal and recovery model, and generating an optimal collaborative operation plan. Scheduling the UAV fleet to perform ice removal operations according to the optimal collaborative operation plan, and real - time monitoring and updating the ice - covered state of the lines.

[0011] Optionally, the pre - constructed UAV ice removal and recovery model includes: The UAV ice removal and recovery model takes the maximization of the comprehensive resilience evaluation index during the ice disaster period as the objective function. The constraint conditions of the UAV ice removal and recovery model include: Global operation time constraint, requiring that the total ice removal operation time of all lines does not exceed the preset maximum allowable operation time. Battery safety constraint, ensuring that the remaining battery power of the UAV meets the minimum requirement for returning to the nearest charging station. Personnel quantity constraint, requiring that the number of operators configured for each line is not less than the number of UAVs allocated. Resource capacity constraint, verifying that the capacity of the ice removal equipment carried by the UAV meets the demand for clearing the ice - covered lines.

[0012] Optionally, generating a recovery strategy based on real - time distribution network state data and a pre - constructed dynamic recovery optimization model includes: Real - time collecting the distribution network node voltage, branch current, and UAV ice removal progress data, and updating the line on - off state parameters. Invoking the dynamic recovery optimization model and converting it into a second - order cone programming model, and setting the objective function as the weighted comprehensive index of maximizing the load recovery efficiency and operation economy. Solving the second - order cone programming model to generate a recovery strategy, including network reconfiguration instructions and distributed power generation output plans, where the instructions include the switch action sequence and new energy inverter control parameters. Verifying whether the recovery strategy meets the node voltage safety range and line current - carrying capacity constraints, and if so, issuing an execution instruction.

[0013] Optionally, the pre-built dynamic restoration optimization model includes: The objective function of the dynamic restoration optimization model is the weighted sum of the load restoration efficiency index and the operation economy index in the post-disaster restoration stage; The constraint conditions of the dynamic restoration optimization model include node power balance constraint, voltage safe operation range constraint, line current-carrying capacity upper limit constraint, new energy output limit constraint, state linkage constraint for forced disconnection of un-deiced lines, and progress feedback constraint to meet the preset minimum closing ratio.

[0014] Optionally, updating the comprehensive resilience evaluation index according to the operation data of the distribution network after implementing the restoration strategy, and dynamically adjusting the weight coefficient allocation of the comprehensive resilience evaluation index includes: When the updated comprehensive resilience evaluation index is lower than the preset threshold or the average improvement amount in consecutive multiple iterations is lower than the set threshold, start the weight adjustment mechanism; Based on the deviation degree between the actual performance of each sub-index and the benchmark value, inversely adjust the weight allocation ratio. When the sub-index value is lower than the preset benchmark value, increase the weight coefficient of this index; Maintain the sum of weight coefficients constant through normalization processing, and synchronously update the weight of the objective function of the dynamic restoration optimization model to form a collaborative optimization closed-loop of evaluation and restoration strategy.

[0015] The second aspect of the present invention provides a distribution network resilience improvement system considering UAV de-icing in an ice disaster scenario. Based on the method for improving the resilience of a distribution network considering UAV de-icing in the first aspect of the present invention, this system includes: A data acquisition module for real-time collecting meteorological data, line operation parameters, and equipment attribute data in the ice disaster environment; An icing monitoring module for generating a set of severely-iced lines through an icing thickness prediction model and a fault probability calculation model; A resilience evaluation module for calculating multi-dimensional resilience evaluation indexes and a comprehensive resilience evaluation index according to the load change curve of the distribution network in the ice disaster scenario, and triggering a UAV collaborative de-icing instruction when the index is lower than the threshold; A UAV scheduling module for generating the optimal flight path and operation time sequence of the UAV, and scheduling the UAV to perform de-icing operations; A dynamic restoration module for generating a restoration strategy based on the real-time distribution network state data, and dynamically adjusting the weight coefficient of the evaluation index and the parameters of the restoration model; A state update module for updating the set of severely-iced lines and the operation state of the distribution network according to the feedback of the UAV de-icing operation.

[0016] Compared with the prior art, the beneficial effects of the present invention at least include: 1. The present invention constructs a multi-factor icing prediction model considering the thermal effect of load changes, the coupling effect of vertical ice force and wind force in the distribution network, solves the technical defect that traditional methods ignore the dynamic change of line operation temperature and the influence of complex mechanical environment, significantly improves the prediction accuracy of icing thickness, and provides an accurate line state assessment basis for ice disaster prevention strategies.

[0017] 2. The present invention establishes a comprehensive resilience evaluation index system integrating load recovery volume, fault recovery time and network loss volume, overcomes the limitation of the existing technology that solely relies on power supply reliability indicators, comprehensively quantifies the defense and recovery capabilities of the distribution network in the ice disaster scenario, and provides a multi-dimensional decision-making basis for formulating differential recovery strategies.

[0018] 3. The present invention breaks through the technical bottleneck that traditional manual de-icing requires power outage operation through a dynamic recovery mechanism of collaborative de-icing by drones during the disaster and collaborative repair by humans and machines after the disaster, realizes rapid de-icing under the condition of non-stop operation of the line, significantly shortens the fault recovery cycle and reduces the risk of high-altitude operation for personnel.

[0019] 4. The present invention constructs a dynamic recovery model integrating the collaborative optimization of load recovery efficiency and operation economy, solves the technical defect that the existing technology separates the improvement of resilience and the control of economic cost, and significantly reduces the network loss and new energy power generation cost during the ice disaster recovery stage on the premise of ensuring the safety of node voltage and the constraint of line current-carrying capacity, realizing the unity of the technical feasibility and economic optimality of the resilience improvement strategy.

[0020] 5. The present invention solves the technical problem that traditional static weight allocation cannot adapt to the dynamic change of multi-objective priorities in the ice disaster recovery stage by establishing a dynamic feedback mechanism for the weight coefficients of the comprehensive resilience evaluation index, reversely corrects the weight ratios of load recovery, fault time and network loss based on the real-time operation data after the implementation of the recovery strategy, realizes the closed-loop collaborative optimization of resilience evaluation and recovery strategy, and effectively improves the adaptive ability and recovery efficiency of the distribution network recovery strategy in the middle and late stages of the ice disaster. Description of the Drawings

[0021] Figure 1 is the flowchart of the method for improving the resilience of the distribution network of the present invention; Figure 2 is the load change curve graph of the system affected by the ice disaster; Figure 3 is the modified IEEE33-node distribution network diagram; Figure 4 is the graph of the total load power, fan and photovoltaic output data of each node in 24 hours; Figure 5 is the comparison graph of the recovered load amounts of different schemes; Figure 6 is the comparison graph of the resilience index values of different schemes. Specific implementation mode

[0022] To make the objectives, technical solutions and advantages of the present invention clearer, the following will clearly and completely describe the technical solutions of the present invention with reference to the accompanying drawings in the embodiments of the present invention. The described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the spirit of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0023] In Embodiment 1 of the present invention, a method for improving the resilience of a distribution network considering ice removal by unmanned aerial vehicles (UAVs) in an ice disaster scenario is provided. As Figure 1 shown, it includes the following steps: Step 1: Collect the meteorological data, line operation parameters and equipment attribute data of the distribution network in the ice disaster environment in real time, calculate the ice thickness and failure probability of each line through a pre-constructed ice thickness prediction model and a failure probability calculation model, and generate a set of lines with severe ice accumulation.

[0024] Preferably, in Step 1, calculating the ice thickness of each line through a pre-constructed ice thickness prediction model includes: (1) Calculate the surface temperature of the line: (1) where: T s is the surface temperature of the line; T c is the ambient temperature; I is the current passing through the line; r is the resistance of the transmission line; v is the wind speed at the location of the transmission line before the ice disaster; d is the diameter of the transmission line.

[0025] (2) Calculate the non-icing time of the conductor considering line loss under ice disaster on the line surface, including: When icing starts, the heat balance equation on the surface of the transmission line can be expressed by the following formula: (2) where: q z is the Joule heat generated by the operation of the transmission line; q c is the convective heat transfer loss between the line and the environment; q w is the heat loss of supercooled water droplets heated to the condensation point; q e is the latent heat absorbed by the evaporation of raindrops on the line surface; q s is the radiation heat dissipation; The calculation formula of the transient heat balance equation on the conductor surface during the continuous process of ice disaster is: (3) In the formula: m is the mass per unit length of the transmission line; C p is the specific heat capacity of the transmission line material itself; q z0 is T s the line loss at the moment of T = 0°C, and T s is the surface temperature of the line.

[0026] The non-icing time of the conductor considering line loss under ice disaster is obtained from Equations (1) to (3) and is expressed by the following formula: (4) In the formula: is the non-icing time of the conductor considering line loss under ice disaster, and T0 represents the surface temperature of the line during stable operation before the disaster.

[0027] (3)Calculate the ice coating thickness of the line under ice disaster: (5) In the formula: D ice (t) is the ice coating thickness of the line at time t; ρ I and ρ W are the density of ice and the density of water, respectively; φ (t) is t the freezing rain amount at time v w (t) is the wind speed at the location of the transmission line at time t; A (t) is the water content of the ambient air.

[0028] It should be noted that in view of the technical defect that the existing ice coating prediction model ignores the dynamic coupling of the Joule heat effect of the line operating current and the environmental temperature, the present invention constructs a multi-physical field coupling model including the transient heat balance equation of the line surface, and calculates the critical non-icing time and ice coating growth rate before ice coating formation in real time, significantly improving the spatio-temporal resolution of ice coating thickness prediction and providing a reliable basis for accurate scheduling of disaster prevention resources.

[0029] Preferably, in step 1, calculating the failure probability of each line through a pre-constructed failure probability calculation model includes: (1)Calculate the ice wind load of each line under ice disaster, including: According to the ice coating thickness of the line at time t, the ice load per unit length of the line L I (t) can be obtained and expressed by the following formula: (6) In the formula: d is the line diameter; the wind load per unit length of the line L W (t) is: (7) In the formula: C is a constant, taking 6.964×10 -3 ; S is the span factor; From formula (6) and formula (7), the ice wind load borne by the transmission line per unit length can be obtained L IW (t), which can be expressed by the following formula: (8) (2) Calculate the line break failure rate, including: According to the design load and ultimate load of the line for ice load bearing, the failure rate of line break per unit length at time t can be obtained as: (9) In the formula: P f (t) is the failure rate of line break per unit length at time t; a 1W and b 1W are the design load and ultimate load of the line for ice load bearing capacity respectively; Then the failure rate model of the whole line can be expressed by the following formula: (10) In the formula: P l,i (t) is the failure rate of the whole line i, l is the corresponding line length, P f (t) is the failure rate of line break per unit length at time t.

[0030] Preferably, in step 1, generating the set of severely ice-covered lines includes: Preset the thresholds of ice thickness and failure probability. When it is monitored that the ice thickness and failure probability of a certain line exceed the safety threshold, then the line is automatically marked as a severely ice-covered line and added to the set ; Exemplarily, the thresholds can be set as: Ice thickness threshold: ; Failure probability threshold: ; Update the and of each line every 5 minutes. If a certain line meets both and , then add it to the set ; Clear the naturally recovered lines every 15 minutes, and set the natural recovery threshold to and ; The natural recovery refers to the situation where the environmental temperature rises and the ice coating slowly melts, and continuous monitoring is required.

[0031] It should be noted that in the subsequent steps, the ice thickness detection is triggered immediately after the drone de-ices. If it meets the standard (≤5mm), it is removed from the set .

[0032] The generation logic of the severely ice-coated lines is real-time monitoring - automatic marking - dynamic update.

[0033] Step 2: According to the operation data of the distribution network and the time series data of the line break probability output in Step 1, calculate the multi-dimensional resilience evaluation index of the distribution network in combination with the pre-constructed load change curve of the distribution network under the ice disaster scenario, and calculate the comprehensive resilience evaluation index. When the comprehensive resilience evaluation value is lower than the preset threshold, trigger the drone collaborative de-icing instruction.

[0034] Preferably, in Step 2, the pre-constructed load change curve of the distribution network under the ice disaster scenario includes: As Figure 2 shown, the figure includes the load curve of the distribution network under normal conditions f 0(t) and the load change curve of the distribution network under the influence of the ice disaster f (t), where the corresponding working conditions at each moment are; At time t0, the ice disaster begins to affect the distribution network; At time t1, icing appears on the line, and the dispatching center dispatches the drone to the severely ice-coated line for de-icing; At time t2, the drone arrives at the scene to perform de-icing operations. At this time, some lines fail due to excessive ice thickness, and the dispatching center actively disconnects the faulty lines through switch operations; At time t3, as the temperature rises, the ice disaster subsides, and the repair team and the drone carry out de-icing work on the actively disconnected ice-coated lines; At time t4, all line de-icing operations are completed, and the repair team restores the lines with line break faults; At time t5, the distribution network returns to the normal operation state.

[0035] In view of the characteristics of ice disasters and the dynamic changes of the distribution network load during the disaster process, the present invention first constructs a load change curve of the distribution network under the ice disaster scenario. Transmission lines are designed with a certain ice resistance. Therefore, in the initial stage of the ice and snow weather, the power grid usually does not immediately experience line breakage faults. However, as the ice and snow weather continues, the ice thickness on the lines gradually increases. When it exceeds the design load-bearing limit, the lines may fail. Further development may lead to more lines being out of service due to line breakage faults, thereby significantly reducing the load level of the distribution network. Therefore, after detecting serious ice coverage on the lines, the repair team will repair the damaged lines so as not to affect the normal power supply of the distribution network, and the load level gradually returns to the normal state. The above change process intuitively reflects the impact of ice disasters on the distribution network load and provides a theoretical basis for formulating resilience improvement strategies.

[0036] Preferably, in step 2, the multi-dimensional resilience evaluation indicators include the total load recovery amount indicator, the time taken for fault recovery indicator, and the network loss indicator.

[0037] Preferably, the calculation of the total load recovery amount indicator of the distribution network in the ice disaster environment includes: This indicator is constructed for the defensive ability of the resilient power grid. During the disaster, unmanned aerial vehicles are used to carry de-icing tools to de-ice severely ice-covered lines; this indicator is used to evaluate the load recovery of the distribution network under ice disasters. The larger the indicator value, the more the load of the distribution network is recovered. It can be expressed by the following formula: (11) In the formula: R ice is the total load recovery amount indicator of the distribution network under ice disasters, W is the load of the distribution network that has not experienced line breakage faults during the entire ice disaster process, W0 is the total load of the distribution network during this period without faults, t0 is the moment when the ice disaster begins to affect the distribution network, and t5 is the time point when the distribution network returns to normal operation.

[0038] Preferably, the calculation of the time taken for fault recovery indicator of the distribution network in the ice disaster environment includes: This indicator is constructed for the recovery ability of the resilient power grid. During the disaster, unmanned aerial vehicles are used to carry de-icing tools to de-ice severely ice-covered lines, and after the disaster, unmanned aerial vehicles are combined with traditional de-icing measures to de-ice faster; This indicator is used to evaluate the recovery speed of the distribution network under ice disasters. The larger the indicator value, the faster the distribution network recovers. It can be expressed by the following formula: (12) In the formula: R time is the time taken for fault recovery indicator, T’ is the time taken for fault recovery during the ice disaster, T 0 is the total duration of the ice disaster.

[0039] Preferably, the network loss index for calculating the distribution network in an ice disaster environment includes: The power loss dissipated in the form of heat during the power transmission process, which is mainly related to the current flowing through the line and the line resistance, can be expressed by the following formula: (13) In the formula: R loss is the network loss index; I ( t ) is the current flowing through the branch of the distribution network, is the total network loss of the branch when the distribution network operates normally, is the line network loss of the distribution network in the ice disaster scenario, is the resistance of branch ij.

[0040] It should be noted that in view of the technical limitation in the prior art that the resilience evaluation index is single, resulting in the restoration strategy tending to local optimization, the present invention realizes the collaborative quantification of the defensive and restorative forces during the entire ice disaster cycle through a three-dimensional dynamic evaluation system integrating the load restoration amount, fault time, and network loss, and dynamically allocates weight coefficients in combination with the analytic hierarchy process, providing a multi-objective optimization benchmark for formulating differential restoration strategies.

[0041] Preferably, the comprehensive resilience evaluation index integrated through the comprehensive resilience evaluation model includes: To improve the resilience of the distribution network under extreme ice disasters, the resilience evaluation model of the distribution network under ice disasters constructed by the present invention takes the comprehensive resilience evaluation index as the objective function. The comprehensive resilience evaluation index consists of three parts: the total load restoration amount, the time taken for fault restoration, and the network loss. Its mathematical model can be expressed by the following formula: (14) In the formula: R is the comprehensive resilience evaluation index; ω1, ω2, ω3 are the weights corresponding to each index, and their initial values are obtained by the analytic hierarchy process; exemplarily, the initial values can be set to 0.6, 0.3, and 0.1 respectively.

[0042] Preferably, the trigger for the collaborative ice removal instruction of the drone includes: When the comprehensive resilience evaluation index is lower than the preset threshold, a drone ice removal decision instruction is generated to trigger the ice removal task scheduling in step three; exemplarily, the preset threshold is set to 0.7 and can be dynamically adjusted according to the environmental temperature and the proportion of critical loads.

[0043] Step 3: According to the spatial distribution of the set of severely ice-covered lines, generate the optimal flight path and operation timing sequence of the drone through the pre-constructed drone ice removal and restoration model, and dispatch a group of drones equipped with laser ice removal devices to perform live-line ice removal operations on the target lines.

[0044] Preferably, step 3 includes: Step 3.1: Obtain the coordinate information, the parameters of the UAV cluster, and the thermodynamic characteristic parameters of the set of heavily iced lines; Step 3.2: Calculate the de-icing energy consumption and operation time of each line through a pre-constructed UAV de-icing recovery model, and generate an optimal cooperative operation plan; More preferably, the UAV de-icing recovery model includes: (1) Calculation of ice mass: (15) In the formula, m ij,i is the ice mass of the line to be de-iced ij ; d ij is the diameter of the line to be de-iced ij ; ρ i is the density of ice, taking 900 kg / m 3 ; l ij,i is the ice-covered length of the line to be de-iced ij ; D ij,ice is the ice-covered thickness of the line to be de-iced ij ; (2) Calculation of de-icing heat demand, including: Calculation of heating heat: (16) In the formula, Q ij,1 is the heat required to raise the temperature of ice to the critical melting temperature: C i is the specific heat capacity of ice, taking 2100 J / (kg·K); ∆ T i is the temperature difference for heating the ice-covered from the initial ambient temperature to the critical temperature; Calculation of phase change latent heat: (17) In the formula, Q ij,2 is the heat required for the phase change of critical ice melting into critical water: L i is the phase change latent heat of ice, taking 335000 J / kg; Calculation of heat loss: (18) In the formula, Q ij,3is the heat conduction loss during the whole process: the loss factor μ is set according to the actual laser ice melting situation and is set to 0.5 in the present invention; Total heat calculation: (19) In the formula, Q ij is the total heat required for the laser to melt a specific mass of ice; (3)Operation time constraint, including: (20) (21) In the formula, is the line to be de-iced ij the time required for de-icing, P is the actual working power of laser de-icing, is the preset maximum allowable operation time; is the global constraint, which is used to ensure that the de-icing operations of all lines are completed before the ice disaster deteriorates and to avoid the spread of faults; (4)Resource constraint verification, including: Number of operators constraint: (22) Resource capacity constraint: (23) In the formula, N person,j is the number of operators under the line j ; N drone,j is the number of unmanned aerial vehicles allocated for the line j ; D j is the resource requirement for the fault point j ; C drone,j is the resource capacity carried by the unmanned aerial vehicle allocated for the line j .

[0045] It should be noted that in the formula, the double subscript ij represents the physical line from node i to node j, associated with the line topology; the single subscript j represents the jth line in the set of ice-covered lines, used to identify the line number.

[0046] Further preferably, the generation of the optimal collaborative operation plan includes: (1)Task package division According to the maximum endurance time of the unmanned aerial vehicle and the battery capacity, the set of severely ice-covered lines is divided into several task packages , each task package satisfies: (24) wherein, is the de-icing time of the line j ; (2) Path optimization For each task package , the flight trajectory of the UAV is calculated using the shortest path algorithm, satisfying: (25) wherein, is the flight time of the m th section of the path, calculated by the flight distance and the cruising speed of the UAV; is the de-icing time of the mth line; (3) Battery replacement strategy When the remaining battery power of the UAV is lower than the safety threshold , plan the path to the nearest charging station , satisfying: (26) where v is the cruising speed of the UAV; (4) Instruction generation Output a scheduling instruction set including the following content: The takeoff and landing coordinate sequences of each UAV; The operation time window of each line; The location and time plan of the battery replacement point.

[0047] It should be noted that in the complete instruction set, the time window constraint ensures the synchronization of the de-icing operation with the power grid restoration plan, and the battery replacement strategy avoids task interruption and improves the operation continuity.

[0048] It is worth noting that aiming at the technical bottleneck of the long restoration cycle caused by the power outage operation required for traditional manual de-icing, the present invention realizes the continuous de-icing of the icing line without power outage by the collaborative design of the UAV swarm path optimization model and the battery safety replacement strategy, and generates the optimal flight trajectory and operation timing under the global operation time constraint, greatly improving the de-icing efficiency of the key line.

[0049] Step 3.3, dispatch the UAV swarm to perform the de-icing operation according to the said plan, and monitor and update the icing state of the line in real time.

[0050] Further preferably, the real-time monitoring and updating of the icing state of the line includes: Detect the change of the surface temperature of the line through the infrared thermal imager carried by the UAV to verify the de-icing elimination state; When the detected icing thickness When the line is restored, it is marked as restored and removed from the set of severely iced lines.

[0051] Preferably, the objective function of the UAV deicing recovery model based on maximizing the comprehensive toughness evaluation index includes: The objective function is: (27) Where: The time span of the ice disaster duration is a collection, and R is the comprehensive resilience assessment indicator.

[0052] With the goal of maximizing the comprehensive resilience index throughout the ice disaster process, the use of drones equipped with laser de-icers to de-ice faulty lines without stopping operations can reduce the number of faulty lines that have been shut down, reduce the loss load, and thus improve the resilience of the distribution network.

[0053] The deicing operation of drones mainly relies on laser deicers. Both drones and laser deicers need power supply to work. Therefore, the deicing process of drones is limited by resources such as battery power, laser deicing tools, and the number of operators. The deicing time of laser deicers is related to the thickness and length of ice on the line. It is stipulated that drones and laser deicers share batteries. Operators are on standby directly below the drones. When the battery is low, the battery replacement time is ignored.

[0054] Step 4. During the UAV de-icing operation, a recovery strategy is generated based on the real-time distribution network status data and the pre-built dynamic recovery optimization model. The comprehensive resilience evaluation index is updated according to the distribution network operation data after the recovery strategy is executed. The weight coefficient allocation of the comprehensive resilience evaluation index is dynamically adjusted based on the update result, and the objective function weight coefficient of the dynamic recovery model is synchronously updated. Steps 2-4 are iteratively executed until the comprehensive resilience evaluation index is continuously higher than the preset threshold.

[0055] Preferably, step 4 comprises: Step 4.1, real-time collection of distribution network node voltage, branch current and drone deicing progress data; Step 4.2, calling the pre-built dynamic recovery optimization model and converting it into a second-order cone programming model; Further preferably, the dynamic recovery optimization model includes an objective function and a distribution network Distflow power flow constraint, wherein: The objective function is: During the ice disaster recovery phase ( t 3≤ t ≤ t 5) By dynamically adjusting the network topology and distributed generation (DG) output, the comprehensive index of post-disaster recovery efficiency and grid operation economy is maximized. Its mathematical expression is: (28) In the formula, , the set of time for the recovery stage after the ice disaster ends; is the load recovery efficiency index at time t; is the economic operation index at time t; and are weight coefficients, which are generated by proportionally mapping the weights ω1 and ω2 of the comprehensive resilience evaluation index. The specific mapping rules are as follows: , , which can be dynamically adjusted according to the recovery progress. Exemplarily, they can be respectively set to 0.7 and 0.3. Select = 0.7, indicating that load recovery is prioritized, and power supply to key loads such as hospitals and communication base stations is quickly restored. = 0.3, indicating that economic operation is optimized, and network losses and new energy generation costs are reduced; Specifically, the load recovery efficiency index includes: used to measure the proportion of key loads that have been restored in the recovery stage, and is calculated by the following formula: (29) Among them, is the set of key load nodes. Exemplarily, the division of key load nodes is shown in Table 2 of Embodiment 3; is the load power restored by node j at time t; is the original load power of node j; The economic operation index includes: used to comprehensively evaluate the network losses and DG generation costs in the recovery stage: (30) Among them, is the benchmark cost, that is, the total cost when not optimized, including network losses and DG generation costs; is the actual cost at time t.

[0056] It should be noted that the objective function of the dynamic recovery optimization model is to maximize the comprehensive index of load recovery efficiency and operation economy in the post-disaster recovery stage. By using double weight coefficients to balance key load supply guarantee and cost control, it ensures that the distribution network can quickly and economically recover to the normal state after the disaster.

[0057] The Distflow power flow constraint conditions of the distribution network include: (1) Node power balance constraints, including: Active power balance: (31) Reactive power balance: (32) Node power decomposition: (33) (34) Voltage-current relationship: (35) (36) (37) (38) Where: represents the set of the first nodes of the node branches ending at node ; represents the last node of the node branch starting at node ; P ki,t and Q ki,t respectively represent the active power and reactive power flowing from node k to node i at time t; r ki and x ki respectively represent the resistance and reactance of branch ki; I ki,t represents the current flowing from node k to node i at time t; represents the active power flowing from node to node at time ; represents the reactive power flowing from node to node at time ; represents the current value flowing from node to node at time ; represents the active power injected from the outside into node at time ; represents the reactive power injected from the outside into node at time ; V j,t represents the voltage amplitude of node j at time ; and respectively represent the resistance and reactance of branch ; and represent the current at node The active power output of wind power and photovoltaic power; and represent the reactive power output of wind power and photovoltaic power at time at node and respectively represent the active power and reactive power provided by SOP to node at time ; and respectively represent the active power and reactive power consumed by node at time ; M represents the relaxation constant, which is an arbitrarily large positive number; α ij,t represents the on-off state of branch ij.

[0058] (2) Operating safety constraints (39) (40) In the formula: represents the voltage amplitude at time ; represents the current flowing through branch at time ; and

[0059] respectively represent the upper and lower voltage limits; (41) In the formula: represents the active power flowing through line at time ; represents the reactive power flowing through line at time ;

[0060] (4) New energy output constraints (42) In the formula: and respectively represent the upper limits of the active power output of wind power and photovoltaic power at node at time ;

[0061] Further preferably, the dynamic recovery optimization model further includes line state linkage constraints, including: (43) Among them, is the on - off state of branch ij at time t, 0 = off, 1 = on; is the set of lines that have completed de - icing; The physical meaning of the line - state linkage constraint is that if line ij belongs to the severely ice - covered line and has not completed de - icing, then this line is forced to be disconnected to avoid the power supply operation of the restoration strategy on the un - de - iced line, resulting in secondary faults.

[0062] Further preferably, the dynamic restoration optimization model further includes a de - icing progress feedback constraint, including: (44) Among them, K is the de - icing completion threshold, and exemplarily, it can be set to 0.8; is the total number of lines that have completed de - icing; The physical meaning of the de - icing progress feedback constraint is that at least 80% of the de - iced lines are closed to ensure that the restoration strategy preferentially uses the safe lines to reconstruct the network, avoiding power supply islanding caused by excessive line disconnection.

[0063] Further preferably, the transformation of the dynamic restoration optimization model into a second - order cone programming model (Second Order Cone Programming, SOCP) includes: (1) Decomposition of line current - carrying capacity constraint (45) (2) Variable substitution (46) (3) Reconstruction of power flow equation (47) (48) (49) (50) (4) Introduction of second - order cone constraint (51) (5) Update of safety constraint (52) (53) It should be noted that in view of the technical deficiency that the existing restoration strategy separates the operation safety and economic objectives, the present invention constructs a dynamic restoration model that co - optimizes the load restoration efficiency and network loss cost, and combines the second - order cone programming relaxation method to solve the global optimal solutions of network reconstruction and distributed power generation output. On the premise of satisfying the node voltage safety constraint, the comprehensive economic cost in the post - disaster restoration stage is reduced.

[0064] Step 4.3, solve the second-order cone programming model and generate a restoration strategy, including network reconfiguration instructions and distributed generation output plans; Further preferably, the step 4.3 includes: Call a solver, input the second-order cone programming (SOCP) model into CPLEX, and output the optimal branch status and optimal DG output; Generate network reconfiguration instructions and DG output plans. Among them, the network reconfiguration instructions include generating a switch action sequence according to the optimal branch status, including closing and opening, and the DG output plan includes sending the optimal DG output to the inverter controller.

[0065] It should be noted that by controlling the switch equipment to adjust the network topology structure, combined with the distributed generation output plan, the distribution network is actively divided into multiple power supply islands. Through active island division, the areas not directly affected by the ice disaster are independently powered by distributed generation, reducing load losses.

[0066] Step 4.4, verify and execute the restoration strategy.

[0067] Further preferably, the step 4.4 includes: Verify whether the restoration strategy given in step 4.3 meets the operation safety constraints and new energy processing constraints in the Distflow power flow constraint conditions of the distribution network; If it is satisfied, control the relevant equipment to execute the restoration strategy, including controlling the switch equipment to adjust the topology according to the network reconfiguration instructions to form new power supply islands; writing the DG output plan into the inverter control protocol; Step 4.5, update the comprehensive resilience evaluation index according to the operation data of the distribution network after executing the restoration strategy, adjust the weight coefficient allocation of the comprehensive resilience evaluation index according to the update result, and synchronously update the weight coefficient of the objective function of the dynamic restoration model.

[0068] Further preferably, updating the comprehensive resilience evaluation index according to the operation data of the distribution network after executing the restoration strategy includes: Calculate the new comprehensive resilience evaluation index: (54) Wherein, is the value of the comprehensive resilience evaluation index in the post-disaster restoration stage; is the load restoration volume index in the post-disaster restoration stage, and the calculation formula is: (55) In the formula, is the actual restored load volume, is the theoretical maximum load volume, is the starting moment of the ice disaster; is the recovery time index, and the calculation formula is: (56) In the formula, is the reference recovery time, that is, the predicted recovery time without adopting the optimization strategy; is the actual recovery time, that is, the actual recovery time after adopting the optimization strategy; is the network loss index in the post-disaster recovery stage, and the calculation formula is: (57) In the formula, is the current value of branch ij during normal operation before the disaster, is the resistance value of branch ij.

[0069] Further preferably, the adjustment of the weight coefficient distribution of the comprehensive resilience evaluation index according to the update result includes: When the updated comprehensive resilience evaluation index or the average value of the index improvement in the last 3 iterations , start the weight adjustment mechanism; specifically, , is the comprehensive resilience evaluation index in the previous iteration cycle, and the dynamic weight adjustment is realized through the gradient feedback mechanism. The weight update formula is: (58) Among them, is the learning rate. Exemplarily, set ; It should be noted that the weight adjustment direction is inversely proportional to the actual performance of the sub-index. If a certain index value is low, its gradient is small, and the weight increase is limited; otherwise, the gradient is large and the weight increase is significant; After executing the gradient feedback mechanism to adjust the weights ω1, ω2, and ω3 of the comprehensive resilience evaluation index, synchronously update the weight coefficients of the objective function of the dynamic recovery optimization model and , to achieve the collaborative optimization of evaluation and recovery strategies. Step 4.6, iteratively execute the above steps until the comprehensive resilience evaluation index is continuously higher than the preset threshold.

[0070] Further preferably, step 4.6 further includes that if for three consecutive cycles , lock the weight coefficient; Terminate when the weight coefficient is locked or the maximum number of iterations is reached. Exemplarily, the maximum number of iterations is set to 10 times.

[0071] It should be noted that the weights are updated proportionally according to the gradients of the sub-indicators, with priority given to improving inefficient indicators; this can ensure that the model always focuses on the resilience dimensions that most need to be optimized, and ultimately maximize the comprehensive resilience of the distribution network.

[0072] It is worth noting that in order to address the technical problem of inaccurate multi-objective optimization caused by the rigidification of traditional weight allocation strategies, the present invention introduces a dynamic adjustment mechanism of weight coefficients based on gradient feedback, and reversely corrects the indicator weights according to the real-time load recovery rate and network loss deviation, so that the recovery strategy can adapt to the priority changes in different stages of the ice disaster, thereby improving the convergence speed and optimization accuracy of strategy iteration.

[0073] The present invention aims to improve the resilience of the distribution network. In combination with the development of drone technology, it proposes an extreme scenario in which ice is applied to the line during an ice disaster. It reasonably analyzes the characteristics of line failures under ice-covered conditions, proposes a resilience assessment index for the distribution network under ice disaster scenarios, and establishes a fault recovery model that considers drone deicing under extreme ice disasters. The simulation results show that the introduction of drones as a tool for deicing during ice disasters can effectively remove ice disasters, restore loads to the greatest extent, reduce power outage time and network losses, and make the entire recovery process have a lower economic cost, which corresponds to the resilience assessment index proposed in the invention and effectively improves the resilience of the distribution network under extreme ice disaster scenarios. In addition, the present invention uses drones for deicing, which can accurately, quickly and efficiently remove ice from the surface of the line on the one hand, and greatly save manpower and material resources on the other hand, and avoid unnecessary risks.

[0074] In embodiment 2, the present invention provides a distribution network resilience improvement system taking into account drone deicing in ice disaster scenarios. Based on the distribution network resilience improvement method taking into account drone deicing in ice disaster scenarios described in embodiment 1, the system includes: Data acquisition module, used to collect meteorological data, line operation parameters and equipment attribute data in real time under ice disaster environment; An ice-covering monitoring module is used to generate a set of severely ice-covered lines through an ice-covering thickness prediction model and a fault probability calculation model; The resilience assessment module is used to calculate the multi-dimensional resilience assessment index and the comprehensive resilience assessment index according to the load change curve of the distribution network in the ice disaster scenario, and trigger the UAV collaborative de-icing command when the index is lower than the threshold; The UAV scheduling module is used to generate the optimal flight path and operation sequence of the UAV and schedule the UAV to perform de-icing operations; Dynamic recovery module, used to generate recovery strategies based on real-time distribution network status data and dynamically adjust evaluation index weight coefficients and recovery model parameters; The status update module is used to update the set of severely iced lines and the operating status of the distribution network based on the feedback from the drone deicing operation.

[0075] Preferably, the system further includes: A human-computer interaction module, which is used to visually display the icing distribution heat map, the UAV flight trajectory, and the change curve of the resilience index, and provide the monitoring and manual intervention interface for the whole process of the method; A data storage module, which is used to store the historical ice disaster case library, the UAV performance parameter library, and the distribution network topology database, and support the parameter calling requirements of all models in the method embodiments.

[0076] In order to more clearly introduce the prominent substantive features of the present invention and the significant progress brought to the prior art, an application example of implementing the present invention is introduced in Embodiment 3.

[0077] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings. The application example specifically includes: The third embodiment of the present invention uses the modified IEEE33 node system as shown in Figure 3 for example verification. It contains a total of 33 nodes, 32 branches, and 5 tie switches. The initial reference voltage is 12.66 kV, and the three-phase power reference value is 10 MV·A. The red nodes are critical loads, and the black nodes are ordinary loads. Distributed photovoltaics and wind turbines are located at nodes 7, 13, 27, and nodes 10 and 24 respectively. The safe range of the system node voltage is [0.95, 1.05]. The load power, wind turbine, and photovoltaic output data are as shown in Figure 4 shown. The example assumes that icing occurs at 4:00 under a blizzard environment, and the lines that may be affected by icing are 6, 12, 18, 21, 24, 32. The cost of load shedding compensation is 1.10 yuan / kW·h.

[0078] Table 1 DG Access Nodes and Capacities

[0079] Table 2 Node Load Levels and Weights

[0080] To verify the effectiveness of the resilience improvement method proposed by the present invention, 3 different schemes are set: Scheme 1: Consider network reconfiguration without considering the repair and restoration process of the lines; Scheme 2: Consider network reconfiguration and consider traditional repair teams for line de-icing during the disaster; Scheme 3: Consider network reconfiguration and consider the role of UAVs in line de-icing during the disaster, with 6 UAVs participating; After simulation calculations, the fault recovery effects corresponding to the 3 schemes are shown in Table 3: Table 3 Fault Recovery Effects Corresponding to 3 Schemes

[0081] As can be seen from Table 3, the load shedding amounts of Plan 2 and Plan 3 are 3144.1 kWh and 1792.2 kWh respectively, both much smaller than 6790.2 kWh of Plan 1. This shows that considering line de-icing measures under extreme ice disasters can effectively reduce the load shedding amount; the load shedding amount of Plan 3 is smaller than that of Plan 2, indicating that the unmanned aerial vehicle (UAV) can speed up the de-icing efficiency through dispatching for de-icing within a limited time. At the same time, the total line power loss of Plan 1 reaches 2619.7 kWh, significantly higher than that of Plan 2 and Plan 3. Although the power loss of Plan 2 decreases, it is still significantly higher than the scenario of using UAV for de-icing, which once again reflects the high efficiency of UAV de-icing. Considering the process of distribution network fault recovery with the participation of UAVs can greatly reduce the input of manpower and materials. Considering that the labor cost of the post-disaster repair team is high and personal safety problems are prone to occur, the UAV also has significant advantages in the economic cost during the fault recovery process.

[0082] From Figure 5 it can be seen that line icing occurs at 4 o'clock. To avoid causing disconnection faults, the dispatching center of Plan 2 disconnects the line with severe icing through switch operation, and sends out a repair team after the disaster to carry out de-icing operations and line repairs, and then the power supply can be restored through switch operation. Plan 3 can dispatch UAVs to carry out de-icing without line outage, which can significantly reduce the number of actively disconnected lines during the disaster. Due to the influence of extreme ice and snow weather, the de-icing efficiency of the manual repair team is very limited, and considering the different icing thicknesses of different lines, the materials and time required for de-icing each line are different. Therefore, considering the method of using UAVs to participate in de-icing during ice disasters can make the line return to the normal state more quickly, thus effectively supplying power to the distribution network and minimizing the load shedding amount during the fault recovery process.

[0083] At the same time, according to the proposed resilience evaluation index, the resilience index values of Plan 2 and Plan 3 are shown in Table 4: Table 4 Resilience indicators corresponding to two de-icing plans

[0084] From Table 4 and Figure 6 as shown, it is known that R ice , R time、 R loss are the three sub-resilience evaluation indicators corresponding to the article. The load level of the line de-icing measure using UAVs has increased significantly. This is because the UAV can achieve de-icing of the line without power interruption during the disaster and assist the repair team to remove the line icing faster after the disaster, enabling the distribution network to operate in a state closest to normal operation and having a higher system resilience level. Therefore, using UAVs for line de-icing can significantly improve the resilience of the distribution network under extreme ice disasters, which is consistent with the invention purpose.

[0085] The present disclosure may be a system, a method, and / or a computer program product. The computer program product may include a computer-readable storage medium having thereon computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0086] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: it is still possible to modify the specific implementation manners of the present invention or make equivalent replacements, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention shall be covered by the protection scope of the claims of the present invention.

Claims

1. A method for improving the resilience of a distribution network considering ice removal by unmanned aerial vehicles in an ice disaster scenario, characterized in that, The method includes the following steps: Collect in real time the meteorological data of the distribution network, the line operation parameters and the equipment attribute data under the ice disaster environment, calculate the ice thickness and the failure probability of each line through the pre-constructed ice thickness prediction model and the failure probability calculation model, and generate a set of severely ice-covered lines; calculate the multi-dimensional resilience evaluation index of the distribution network in combination with the pre-constructed load change curve of the distribution network in the ice disaster scenario, and calculate the comprehensive resilience evaluation index. When the comprehensive resilience evaluation value is lower than the preset threshold, trigger the unmanned aerial vehicle (UAV) collaborative de-icing instruction; Generate the optimal flight path and operation time sequence of the UAV through the pre-constructed UAV de-icing and restoration model, and dispatch the UAV to perform de-icing operations; during the UAV de-icing operation, generate a restoration strategy based on the real-time distribution network state data and the pre-constructed dynamic restoration optimization model, update the comprehensive resilience evaluation index according to the distribution network operation data after implementing the restoration strategy, and dynamically adjust the weight coefficient allocation of the comprehensive resilience evaluation index based on the update result, and synchronously update the weight coefficient of the objective function of the dynamic restoration model. Iteratively execute the above steps until the comprehensive resilience evaluation index is continuously higher than the preset threshold.

2. The method for improving the resilience of a distribution network considering ice removal by unmanned aerial vehicles in an ice disaster scenario according to claim 1, characterized in that: The calculation of the ice thickness of each line through the pre-constructed ice thickness prediction model includes: Calculate the change in the surface temperature of the line based on the Joule heat effect generated by the line energization current and the ambient temperature; Determine the critical non-icing time before ice formation through the transient heat balance equation of the wire surface during the ice disaster. The transient heat balance equation of the wire surface characterizes the dynamic change process of the wire surface temperature during the ice disaster through the product of the specific heat capacity of the wire material and the change rate of the wire surface temperature, and includes the heat balance relationship of Joule heat generated by the line operation, convective heat transfer loss between the line and the environment, heat loss of supercooled water droplets heated to the condensation point, latent heat absorption of raindrop evaporation on the wire surface, and heat radiation. According to the real-time freezing rain intensity, wind speed and air moisture content, calculate the ice thickness growth increment in segments after the ice formation start time, where the ice thickness is zero before the critical non-icing time, and dynamically calculate the ice accumulation after the critical non-icing time.

3. The method for improving the resilience of a distribution network considering ice removal by unmanned aerial vehicles in an ice disaster scenario according to claim 1, characterized in that: The generation of the set of severely ice-covered lines includes: Preset the ice thickness safety threshold and the failure probability safety threshold; Monitor in real time the ice thickness and the failure probability of each line. When both the ice thickness and the failure probability exceed their safety thresholds, dynamically mark the corresponding line as a severely ice-covered line and add it to the set of severely ice-covered lines; Periodically remove the lines that meet the natural recovery conditions, where the natural recovery conditions include that the ice thickness is lower than the first recovery threshold and the failure probability is lower than the second recovery threshold.

4. The method for improving the resilience of a distribution network considering ice removal by unmanned aerial vehicles in an ice disaster scenario according to claim 1, characterized in that: The calculation of the multi-dimensional resilience evaluation index of the distribution network includes: The total load recovery volume index, which is calculated based on the time integral ratio of the actual recovered load volume during the ice disaster to the theoretical maximum load volume; The fault recovery time index, which is calculated based on the ratio of the time used for fault recovery to the total duration of the ice disaster; The network loss index, which is calculated based on the time integral ratio of the actual line network loss to the reference network loss.

5. The method for improving the resilience of a distribution network considering ice removal by unmanned aerial vehicles in an ice disaster scenario according to claim 1, characterized in that: The generation of the optimal flight path and operation time sequence of the UAV through the pre-constructed UAV de-icing and restoration model, and the dispatch of the UAV to perform de-icing operations include: Obtain the coordinate information, UAV cluster parameters, and thermodynamic characteristic parameters of the severely ice-covered line set; Calculate the de-icing energy consumption and operation time of each line through a pre-constructed UAV de-icing recovery model, and generate an optimal collaborative operation plan; Dispatch the UAV group to perform de-icing operations according to the optimal collaborative operation plan, and monitor and update the line ice-covered state in real time.

6. The method for improving the resilience of a distribution network considering ice removal by unmanned aerial vehicles in an ice disaster scenario according to claim 5, characterized in that: The pre-constructed UAV de-icing recovery model includes: The UAV de-icing recovery model takes the maximization of the comprehensive resilience evaluation index during the ice disaster period as the objective function; The constraint conditions of the UAV de-icing recovery model include: Global operation time constraint, requiring that the total de-icing operation time of all lines does not exceed the preset maximum allowable operation time; Battery safety constraint, ensuring that the remaining battery power of the UAV meets the minimum requirements for returning to the nearest charging station; Personnel quantity constraint, requiring that the number of operators configured for each line is not less than the number of allocated UAVs; Resource capacity constraint, verifying that the capacity of the de-icing equipment carried by the UAV meets the requirements for clearing line ice coverage.

7. A method for improving the resilience of a distribution network considering ice removal by unmanned aerial vehicles in an ice disaster scenario, characterized in that: The generation of the recovery strategy based on the real-time distribution network state data and the pre-constructed dynamic recovery optimization model includes: Real-time collect the distribution network node voltage, branch current, and UAV de-icing progress data, and update the line on-off state parameters; Call the dynamic recovery optimization model and transform it into a second-order cone programming model, and set the objective function as the weighted comprehensive index of maximizing the load recovery efficiency and operation economy; Solve the second-order cone programming model to generate a recovery strategy, including network reconfiguration instructions and distributed power generation output plans, and the instructions include switch action sequences and new energy inverter control parameters; Verify whether the recovery strategy meets the node voltage safety range and line current-carrying capacity constraints. If it meets, issue an execution instruction.

8. A method for improving the resilience of a distribution network considering ice removal by unmanned aerial vehicles in an ice disaster scenario, characterized in that: The pre-constructed dynamic recovery optimization model includes: The objective function of the dynamic recovery optimization model is the weighted sum of the load recovery efficiency index and the operation economy index in the post-disaster recovery stage; The constraint conditions of the dynamic recovery optimization model include node power balance constraint, voltage safe operation range constraint, line current-carrying capacity upper limit constraint, new energy output limit constraint, state linkage constraint for forced disconnection of un-de-iced lines, and progress feedback constraint to meet the preset minimum closing ratio.

9. A method for improving the resilience of a distribution network considering ice removal by unmanned aerial vehicles in an ice disaster scenario, characterized in that: The update of the comprehensive resilience evaluation index based on the distribution network operation data after executing the recovery strategy, and the dynamic adjustment of the weight coefficient distribution of the comprehensive resilience evaluation index based on the update result include: When the updated comprehensive resilience evaluation index is lower than the preset threshold or the average value of the improvement amount for consecutive multiple iterations is lower than the set threshold, start the weight adjustment mechanism; Based on the deviation degree of the actual performance of each sub-index from the benchmark value, inversely adjust the weight distribution ratio. When the sub-index value is lower than the preset benchmark value, increase the weight coefficient of this index; Maintain the constancy of the total weight coefficient through normalization processing, and synchronously update the objective function weight of the dynamic recovery optimization model to form a collaborative optimization closed-loop of evaluation and recovery strategy.

10. A distribution network resilience improvement system considering ice removal by unmanned aerial vehicles in an ice disaster scenario, based on the method for improving the resilience of a distribution network considering ice removal by unmanned aerial vehicles according to any one of claims 1-9, characterized in that, This system includes: A data acquisition module for real-time collecting meteorological data, line operation parameters, and equipment attribute data in the ice disaster environment; An ice-coverage monitoring module, which is used to generate a set of lines with severe ice-coverage through an ice-coverage thickness prediction model and a fault probability calculation model; A resilience evaluation module, which is used to calculate multi-dimensional resilience evaluation indicators and comprehensive resilience evaluation indicators according to the load change curve of the distribution network in an ice disaster scenario, and trigger a drone collaborative de-icing instruction when the indicators are lower than the threshold; A drone scheduling module, which is used to generate the optimal flight path and operation time sequence of the drone, and schedule the drone to perform de-icing operations; A dynamic restoration module, which is used to generate a restoration strategy based on real-time distribution network state data, and dynamically adjust the evaluation index weight coefficient and restoration model parameters; A status update module, which is used to update the set of lines with severe ice-coverage and the operation status of the distribution network according to the feedback of the drone de-icing operation.

Images