A tower gap insulation optimization design method and system considering bird-crossing fault characteristics

Abstract

The present invention discloses a tower gap insulation optimization design method and system that takes into account the characteristics of bird-related faults, and relates to the technical field of power system insulation protection. The method comprises: obtaining the structural parameters and operating environment information of the target tower, constructing a three-dimensional gap simulation model including local insulation armor, and superimposing a bird foreign body disturbance model; performing electric field simulation calculations on the three-dimensional gap simulation model after superimposing the bird foreign body disturbance model, fusing to form a discharge risk map, and identifying key risk areas where the local discharge risk value exceeds a set threshold; using the thickness function and coverage boundary function of the insulation armor as design variables, constructing an optimization function with the minimum discharge risk integral value and the minimum armor material volume; calling the optimization algorithm to iteratively solve the optimal parameter set, and outputting a non-uniform thickness distribution structure and spatial coordinates. The present invention realizes the intelligent optimization of local insulation structure for bird-related discharge risks, significantly improving the safety and pertinence of tower gap insulation.

Description

Technical Field

[0001] The present invention relates to the technical field of power system insulation protection, and in particular to a tower gap insulation optimization design method and system taking bird-related fault characteristics into consideration. Background Art

[0002] During the operation of high-voltage transmission lines, birds frequently move between towers and conductors, particularly across ecologically sensitive areas such as mountains, forests, water sources, and farmland. Bird disturbance and defecation are significant factors in flashovers. While common bird prevention measures, such as installing bird scarers and adding insulator shields, can reduce the probability of failures to a certain extent, they fail to systematically identify and mitigate the risk of bird-related discharges from a structural design perspective.

[0003] Currently, traditional tower insulation design primarily sets static electrical clearances based on voltage levels, supplemented by empirical insulation redundancy. High-risk areas are often addressed by manually thickening insulation armor or adding armor belts at equal intervals. This approach lacks the support of sophisticated simulation analysis and is unable to incorporate differentiated optimization based on local electric field distribution characteristics. Summary of the Invention

[0004] To address the above issues, the present invention proposes a tower gap insulation optimization design method and system that considers the characteristics of bird-related faults. The bird disturbance behavior is converted into electric field interference source modeling, forming a local insulation structure intelligent design method guided by risk maps and driven by multi-objective functions to improve the effectiveness and pertinence of local insulation configuration.

[0005] The present invention achieves the above-mentioned purpose through the following technical solutions:

[0006] A tower gap insulation optimization design method considering bird-related fault characteristics, the method comprising:

[0007] Obtain the structural parameters and operating environment information of the target tower and build a three-dimensional gap simulation model with local insulation armor;

[0008] A bird foreign body disturbance model is superimposed on the three-dimensional gap simulation model to describe the disturbance effect of bird activity on the gap electric field distribution. The bird foreign body disturbance model includes a conductor that simulates bird interference and a conductive path that simulates the path of bird excrement. The conductive path generates a random conical trajectory based on bird excretion behavior statistics and wind disturbance.

[0009] The electric field simulation calculation is performed on the three-dimensional gap simulation model after superimposing the bird foreign body disturbance model. The peak value of the electric field intensity, the rate of change of the electric field gradient, and the density of the equipotential lines are extracted as discharge risk indicators. These are integrated to form a discharge risk map, and key risk areas where the local discharge risk value exceeds the set threshold are identified.

[0010] Taking the thickness and coverage of the local insulating armor as design variables, an optimization function is constructed with the objectives of minimizing the weighted integral value of the discharge risk map and minimizing the volume of the local insulating armor, with complete coverage of the critical risk area as a spatial constraint.

[0011] A multi-objective optimization algorithm is used to iteratively solve the optimization function to obtain the optimal parameter set that meets the insulation requirements. According to the discharge risk map, the thickness of the local insulating armor is made to form a non-uniform distribution structure in the key risk area as the local discharge risk value increases. The structural design scheme including the armor thickness distribution and spatial position coordinates is output.

[0012] As a preferred embodiment of the present invention, the structural parameters include: the three-dimensional shape and inclination angle of the tower extracted by a directional tower detection model for synthetic aperture radar images, as well as the conductor hanging point height, crossarm length, phase spacing, and hardware geometric parameters of the target tower, wherein the hardware geometric parameters include the grading ring diameter, insulator string length, and hanging point coordinates;

[0013] The operating environment information includes: insulator defect types and contamination levels identified based on the improved YOLOv7 model, as well as historical bird damage failure heat maps of the tower area, typical bird activity height statistics, regional meteorological parameters, and terrain background data. The regional meteorological parameters include annual average humidity, maximum wind speed, salt density distribution, and annual average number of thunderstorm days.

[0014] The local insulating armor is a hydrophobic composite material covering the cross arm end and the hardware area, and the relative dielectric constant Satisfy the dynamic equation ,in It is a local field strength, and the surface hydrophobic contact angle is not less than 110 degrees.

[0015] As a preferred solution of the present invention, the directional tower detection model adopts sparse hint coding and shape adaptive feature fusion technology, specifically including:

[0016] The model input is a synthetic aperture radar image of the target area and a set of initial prompt point coordinates. The prompt points are the possible locations of towers.

[0017] The model frontend includes a sparse cue encoder that maps cue point coordinates into low-dimensional sparse vector embeddings that are fused with image features extracted by multiple Transformer encoding layers, using a cross-attention mechanism between point embeddings and image position embeddings.

[0018] A shape-adaptive feature fusion module is set in the middle of the model to dynamically adjust the convolution kernel size and feature map sampling area according to the aspect ratio of the target candidate box;

[0019] The model output is a detection result with rotation parameters including the direction angle and target frame coordinates;

[0020] The improved YOLOv7 model specifically includes:

[0021] The SPC module in the YOLOv7 backbone network is replaced by the RFB module, which contains multiple parallel channels with different convolution kernel sizes to enhance multi-scale feature perception.

[0022] A coordinate attention mechanism is introduced into the Head output module, which includes a position encoding module and a space-channel decoupling attention module to enhance feature direction selectivity;

[0023] In the bounding box regression part, the WIoU loss function is used to replace the original CIoU loss function, and the weight of the WIoU loss function is adaptively adjusted according to the quality of the predicted box.

[0024] As a preferred embodiment of the present invention, the method for constructing a three-dimensional gap simulation model containing local insulating armor includes:

[0025] An initial three-dimensional gap simulation model was established using the finite element method. The initial coverage of the local insulation armor was preset in areas with a high incidence of bird damage on crossarms. The initial coverage range was 60% to 75% of the length from the conductor suspension string point to the crossarm end.

[0026] Based on historical bird damage heat maps, terrain background data, and typical bird activity height statistics, adaptive grid densification technology was used in areas with high incidence of guano flashovers. The grid size of 10 cm in conventional areas was dynamically reduced to 2 cm in these areas.

[0027] The correlation distribution data of wind speed and equivalent salt density were imported, and the pollution accumulation rate was calculated through fluid dynamics simulation. Combined with the identified insulator defect type and pollution level, the accuracy of the electric field boundary conditions of the three-dimensional gap simulation model was verified and the surface potential gradient distribution was corrected.

[0028] As a preferred solution of the present invention, a bird foreign body disturbance model is superimposed on the three-dimensional gap simulation model, and the method includes:

[0029] A conductor is set up to simulate bird interference. The conductor is set to be an ellipsoid, with the size of the major axis in the range of 0.3 to 0.8 times the distance between the target towers, and the conductivity is set at 10 3 to 10 5 S / m, arranged in adjacent areas between cross-arm conductors;

[0030] A conductive path is provided for simulating the falling path of bird excrement, wherein the conductive path is anchored at the center of the lower surface of the conductor and is a cone with a shape perturbation function, wherein the cone has a length between 0.4 meters and 1.2 meters, a top diameter less than 20 mm, and a bottom diameter greater than 40 mm;

[0031] The centroid trajectory of the conductive path determines the starting height and dive angle according to the statistics of bird defecation behavior, and generates an offset trajectory in combination with the windage angle range set in the target area;

[0032] The conductors and conductive paths are both superimposed in the three-dimensional gap simulation model in the form of three-dimensional entities, and participate in the electric field boundary condition setting and simulation calculation.

[0033] As a preferred solution of the present invention, an electric field simulation calculation is performed on a three-dimensional gap simulation model superimposed with a bird foreign body disturbance model, and the method includes:

[0034] Perform 3D finite element steady-state electric field simulation and set conductor voltage and dielectric boundary conditions;

[0035] The electric field intensity, electric field gradient change rate, and equipotential line density in the area around the insulator are extracted. The equipotential line density is calculated as the number of equipotential surfaces per unit volume or the inverse of the average spacing. A directional factor consistent with the set wind direction is introduced into the calculation of the electric field gradient change rate to correct the sensitivity of the discharge risk indicator to the direction of the bird excretion path.

[0036] According to the scale characteristics of different spatial regions, graded weighted coefficients are set for the three discharge risk indicators, and the three indicators are integrated to form a discharge risk map.

[0037] An adaptive threshold is set according to the mean and standard deviation of the discharge risk indicators in each area, and the spatial collection above the set threshold is identified as the key risk area.

[0038] As a preferred solution of the present invention, the optimization function includes a first objective function and a second objective function;

[0039] The first objective function is the weighted integral value of the discharge risk map, expressed as: ;

[0040] The second objective function is the volume of the local insulating armor, expressed as:

[0041] ;

[0042] The electrical constraints are: ;

[0043] Where, The spatial domain occupied by the key risk area; is the discharge risk spectrum value; is the electric field gradient mode length, is the density of equipotential lines, 、 is the weighting factor; is the volume element in three-dimensional space; 、 are the thickness distribution function and coverage boundary function of the local insulating armor respectively; In the electric field simulation, the space point The electric field strength value at ; For line voltage level The corresponding critical value of the air gap breakdown electric field strength.

[0044] As a preferred solution of the present invention, the thickness distribution function of the local insulating armor is Defined as:

[0045] ;

[0046] Where, is the thickness scaling factor; In key risk areas The maximum value of is the nonlinear risk enhancement index; is the minimum thickness constraint.

[0047] As a preferred solution of the present invention, the iterative solution of the optimization function using a multi-objective optimization algorithm specifically includes:

[0048] The thickness distribution function and coverage boundary function of the local insulating armor are encoded as design variables to generate the initial population;

[0049] In each iteration, the three-dimensional gap simulation model is called to calculate the electric field distribution and discharge risk map corresponding to the design variables, and the objective function value is directly read through the result cache mapping table for repeated design variables;

[0050] For the spatial points where the gradient amplitude in the discharge risk map exceeds the set threshold, the domain of the coverage boundary function is expanded in the gradient direction, and the expansion distance is ,in is the empirical scaling factor;

[0051] For individuals with excessive electric field strength, the perturbation amplitude of the thickness distribution function is reduced;

[0052] Update the thickness distribution function based on the current discharge risk map, and introduce an inertia factor to suppress mutations;

[0053] Dynamically adjust the crossover rate and mutation rate according to the population objective function distribution entropy;

[0054] Extract the maximum gradient direction of the discharge risk map in the elite solution to guide local variation;

[0055] Output the minimum volume solution in the Pareto solution set that meets the spatial and electrical constraints, and generate spatial point set structural data of thickness distribution.

[0056] A tower gap insulation optimization design system considering bird-borne fault characteristics includes:

[0057] Data input module, used to obtain the structural parameters and operating environment information of the target tower;

[0058] a three-dimensional modeling module for constructing a three-dimensional gap simulation model including a local insulating armor based on the structural parameters, and superimposing a bird foreign body disturbance model on the three-dimensional gap simulation model to describe the disturbance effect of bird activities on the gap electric field distribution;

[0059] The electric field analysis module is used to perform electric field simulation calculations on the three-dimensional gap simulation model after superimposing the bird foreign body disturbance model. It extracts the electric field intensity peak, electric field gradient change rate, and equipotential line density as discharge risk indicators, integrates them into a discharge risk map, and identifies key risk areas where the partial discharge risk value exceeds the set threshold;

[0060] An optimization function construction module is used to construct an optimization function based on the discharge risk map. The optimization objectives include minimizing the weighted integral value of the discharge risk map and minimizing the volume of the local insulating armor.

[0061] The parameter optimization module is used to iteratively solve the optimal parameter set using the thickness distribution function and coverage boundary function of the local insulating armor as design variables using a multi-objective optimization algorithm to meet spatial and electrical constraints.

[0062] The structural output module is used to generate a non-uniformly distributed structure in which the armor thickness increases with the discharge risk value in the key risk area based on the optimization results. The module outputs a data set containing the thickness distribution and spatial position coordinates to guide the layout of local insulating armor.

[0063] The beneficial effects of the present invention are: introducing a bird foreign body disturbance modeling mechanism, and realizing high-fidelity simulation modeling of the impact of bird behavior on the gap electric field disturbance by constructing a conductor and a conical conductive path to simulate the bird body and excretion path; establishing a discharge risk map with electric field strength, electric field gradient and equipotential line density as input, quantifying the local risk spatial distribution, and breaking through the limitations of previous static gap design; constructing a dual-objective optimization function to minimize the volume of local insulating armor while controlling the bird-related discharge risk, thereby improving design economy and engineering adaptability; proposing a thickness response function and a dynamic disturbance compression mechanism to form a continuous and physically reasonable non-uniform distribution structure of local armor thickness in high-risk areas; establishing a full-process integrated system from data input-three-dimensional modeling-electric field simulation-optimization calculation-structure output to realize the automated design output of local insulating structures, improve design efficiency and enhance customization capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be derived from these drawings without inventive effort. Among them:

[0065] Figure 1 is a flow chart of the method of the present invention;

[0066] Figure 2 This is a system structure diagram of the present invention. DETAILED DESCRIPTION

[0067] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the described embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field are within the scope of protection of the present invention.

[0068] like Figure 1 FIG. 1 is an embodiment of the present invention, which provides a tower gap insulation optimization design method considering bird-related fault characteristics, including the following steps:

[0069] S1: Obtain the structural parameters and operating environment information of the target tower and construct a three-dimensional gap simulation model with local insulation armor.

[0070] Structural parameters include: the three-dimensional shape and tilt angle of the tower extracted by the directional tower detection model for synthetic aperture radar images, as well as the conductor hanging point height, crossarm length, phase spacing, and hardware geometric parameters of the target tower, among which the hardware geometric parameters include the grading ring diameter, insulator string length, and hanging point coordinates;

[0071] Operating environment information includes: insulator defect types and contamination levels identified using an improved YOLOv7 model, historical bird damage heat maps of the tower area, typical bird activity height statistics, regional meteorological parameters, and terrain background data. Regional meteorological parameters include annual average humidity, maximum wind speed, salt density distribution, and the average number of thunderstorm days per year.

[0072] The local insulating armor is a hydrophobic composite material covering the cross arm end and the hardware area, with a relative dielectric constant of Satisfy the dynamic equation ,in is the local electric field strength (unit: kV / mm), and the surface hydrophobic contact angle is not less than 110 degrees.

[0073] In a bird-related fault protection design application for a 220kV straight tower-type transmission line, based on the method of the present invention, synthetic aperture radar images of the tower area were obtained using an unmanned aerial vehicle (UAV). The images were processed using an improved directional tower detection model to extract the three-dimensional structural parameters of the target tower, including: the height of the conductor hanging point from the ground was 32.6 meters, the horizontal length of the crossarm was 3.2 meters, and the phase spacing was 2.1 meters; the hardware used a grading ring structure with an average ring diameter of 450 mm.

[0074] Based on tower operation information, an insulator recognition model based on an improved YOLOv7 network structure (introducing a coordinate attention mechanism and RFB module) was used to detect and classify insulator images. The recognition results showed obvious cracks on the right-phase insulator of the tower's crossarm, and large areas of contamination deposits on the left phase. The insulator was identified as being in the early stages of a mild flashover, with a contamination level assessment of II.

[0075] Regional environmental information was obtained by sourcing historical wading bird failure databases and local meteorological monitoring data. The target area has an average annual humidity of 72%, a maximum wind speed of 18.2 meters per second, a salt density of 0.12 mg / cm², and an average annual number of thunderstorms of 35 days. Bird monitoring radar and historical records indicate that corvids predominate in the area, with activity typically concentrated between 25 and 35 meters. The peak wading bird season occurs from March to May.

[0076] Based on the above comprehensive parameters, the insulation armor was installed in a localized area covering the ends of the crossarms and near the hardware. A composite insulating coating material with enhanced hydrophobicity was selected, whose relative dielectric constant satisfies the dynamic equation. This material maintains stable relative dielectric properties in strong electric fields and exhibits excellent resistance to electrical corrosion. The hydrophobic angle formed by the surface treatment was tested to 113°, meeting the design requirement of ≥110°, ensuring that a conductive film is not easily formed in dirty environments.

[0077] This embodiment adopts a directional tower detection model that combines sparse cue coding with shape adaptive feature fusion technology to extract structural parameters such as the three-dimensional geometry and tilt angle of the target tower from synthetic aperture radar images.

[0078] 1. Data Input

[0079] The input is a high-resolution SAR image of the target area and a set of cue points generated using a priori spatial rules (such as transmission line paths). The cue points are sparse coordinate sets that infer the locations of towers.

[0080] 2. Front-end sparse hint encoder

[0081] The cue point coordinates are embedded as low-dimensional sparse position vectors and fed into the front-end encoding module along with the image features. The encoder employs a multi-layer Transformer architecture to extract high-level relationships between the cue location and the local image structure. A cross-attention mechanism is used to enhance the responsiveness of image features to the spatial location of the cue point.

[0082] 3. Middle shape adaptive feature fusion module

[0083] Shape-Aware Convolution (SHAPE-AWARENESS CONV) with a dynamic kernel resizing mechanism is introduced. This module automatically adjusts the convolution receptive field based on the appearance and geometric features of candidate objects, making the model more geometrically invariant to tilted towers and partially deformed objects. The convolution kernel shape dynamically changes based on the object's aspect ratio.

[0084] 4. Output module

[0085] The model output is: the spatial coordinates of the tower center point , height estimate (Based on image scale calibration), tilt angle (based on candidate box rotation parameters), crossarm orientation estimation parameters (optional).

[0086] This embodiment adopts an improved YOLOv7 model with multi-module fusion to achieve accurate identification of defect types and contamination levels in transmission line insulator images.

[0087] 1. Improvement of backbone network structure

[0088] The Receptive Field Block (RFB) structure is introduced in the SPPCSPC module of the YOLOv7 backbone network. The RFB module contains multi-scale receptive field paths, including 3×3, 5×5, and dilated convolution paths, which are fused in parallel to enhance the contextual feature representation of the insulator region.

[0089] 2. Feature Enhancement Module

[0090] To improve the model's robustness in localizing insulator targets, a combined position and channel attention mechanism is introduced before the head output module. Position information is output by the position information encoding module, while channel weighting uses the Squeeze-and-Excitation module to optimize feature weights.

[0091] 3. Coordinate Prediction Mechanism

[0092] The head output introduces a position annotation attention module, combined with a spatial-channel fusion strategy to enhance small object selection. The coordinate output is a quadrilateral bounding box, and supports rotation angle parameter estimation (for tilted insulator recognition).

[0093] 4. Loss Function Adjustment

[0094] The bounding box regression component uses a weighted IoU loss function (WIoU) instead of the traditional CIoU. The WIoU weighting function incorporates prediction bias and boundary matching quality factors to automatically adjust the sensitivity of prediction errors. This is particularly suitable for complex scenes with arc interference or pollution flashover backgrounds.

[0095] 5. Output results

[0096] Model output: bounding box location of each target, classification label (such as crack, porcelain chipping, light pollution, heavy pollution), classification confidence, and tower number (associated through image registration).

[0097] In this embodiment, a three-dimensional gap simulation electric field model of the local insulating armor area was constructed for a typical high-incidence area of ​​bird-related discharge faults in a certain 500kV linear tower area for subsequent electric field simulation analysis and structural optimization design.

[0098] 1. Construction of the initial 3D model

[0099] Finite element software (such as ANSYS Maxwell) is used to establish a three-dimensional gap simulation structural model. The modeling objects include: the crossarm body (steel component, given conductivity), the conductor suspension string (modeled as a conductor and grading ring structure), the hardware area and its insulator string structure (defining the dielectric constant), and the ground plane area (set as a zero potential boundary).

[0100] In the model, the insulating armor is designed to cover the right end of the crossarm and the area around the conductor hanging point. Its initial range is set to the area extending from the hanging point toward the end of the crossarm, with a length of 70% of the total length of the crossarm.

[0101] 2. Data-driven high-risk area delineation

[0102] By combining the bird-damaging discharge defect records and map sheet numbering within the region over the past five years, a bird hotspot distribution map was generated. This map was overlaid with typical topographic information (such as rivers, woodlands, and substation corridors) to identify key spatial units with a heat level of 0.7 or above. Based on the flight altitude distribution of Columbidae and Falconidae birds within this region, the primary bird disturbance altitude in the target spatial area was calculated to be concentrated between 26 and 38 meters.

[0103] 3. Adaptive mesh refinement settings

[0104] In 3D modeling, the standard mesh size for the entire modeling area was set to 10 cm. Within the high-heat subregions of the wading birds, the model automatically reduced the mesh size to 2 cm to improve the accuracy of the electric field simulation and ensure accurate capture of the localized electric field concentration at the tip. An octree partitioning strategy was used to dynamically generate encrypted regions, and the corresponding region labels were recorded for post-simulation stratification.

[0105] 4. Boundary condition correction

[0106] Combined with historical wind speed and salt density data for the target area (maximum annual wind speed of 22.1 m / s, typical salt density of 0.18 mg / cm²), the data was input into the fluid-structure interaction simulation module to develop a model of the air flow field and the interface interaction between the structure and contamination. The simulation yielded a non-uniform distribution map of surface contamination coverage at different locations within the target area.

[0107] The map is mapped to the surface of the insulator in the electric field simulation model and used as an additional boundary condition to correct the dielectric constant disturbance. Finally, the complete three-dimensional electric field boundary setting conditions are derived for subsequent electric field simulation and risk map generation stages.

[0108] S2: A bird foreign body disturbance model is superimposed on the three-dimensional gap simulation model to describe the disturbance effect of bird activities on the gap electric field distribution. The bird foreign body disturbance model includes a conductor that simulates the interference of the bird body and a conductive path that simulates the falling path of bird excrement. The conductive path generates a random conical trajectory based on the statistics of bird excretion behavior and wind disturbance.

[0109] In this embodiment, in order to evaluate the disturbance effect of bird-borne foreign matter on the electric field distribution in the gap between transmission line towers, a bird-borne foreign matter disturbance model was constructed by superimposing it on the three-dimensional gap simulation model.

[0110] 1. Modeling the Conductor of a Bird

[0111] According to the regional bird damage database, the target area's most frequent wading bird species are medium-sized birds from the Corvidae and Falconidae families, with an average body length of 28–41 cm. Considering the tower structure design, with a phase spacing of approximately 2.2 meters, the model sets the bird conductor as an ellipsoid with a major axis length of 0.6 times the phase spacing (approximately 1.32 meters). The ratio of the major and minor axes of the transverse cross-section is set to 1.5:1. The conductivity is set to 5.0×10 4 S / m, corresponding to the average water conductivity characteristics of poultry tissues.

[0112] The ellipsoid is positioned between the two-phase conductors on the crossarm, with the spatial height set at 0.5 meters below the conductors, the direction perpendicular to the conductors, and the axis passing through the midpoint of the gap to simulate the interference behavior of the bird on the gap electric field in a static perching or gliding state.

[0113] 2. Modeling the Conductive Pathway in Guano

[0114] The excretion cone is anchored at the center point of the lower surface of the bird's conductor, which is set as a solid conical channel with a length of 0.9 meters, a top diameter of 15 mm, and a bottom diameter of 45 mm. This structure introduces a shape perturbation function : ;in is the initial radius, is the disturbance amplitude (set to 3 mm), is the coordinate of the cone axis direction, is the total length of the cone. This structure is used to simulate the diffusion path of the excretion column after being disturbed by wind in the air.

[0115] 3. Path deviation trajectory calculation

[0116] Based on bird defecation behavior statistics, the dive angle was set within a range of 35°–60°, with the starting altitude set at the lower boundary of the conductor. Based on the target terrain and typical wind speed statistics, the horizontal offset angle was set between –10° and +15°. The dive vector and windage angle were superimposed to generate a three-dimensional offset trajectory of the centroid of the conductive pathway cone. This trajectory was then used to rotate the cone and transform its position in the 3D simulation model.

[0117] 4. Simulation model integration and boundary condition assignment

[0118] Conductors and conductive pathways are embedded in the gap simulation model using 3D solid modeling, and the material conductivity boundary properties are assigned. During the electric field simulation, these conductors act as highly conductive interfering objects and participate in the boundary condition setting, ensuring that the overall model accurately reflects the concentrated electric field distortion caused by bird intrusion.

[0119] After the electric field solution is completed, the equipotential line density and the rate of change of the local electric field intensity near the foreign object are quantitatively extracted to construct the disturbance input layer of the subsequent discharge risk map.

[0120] S3: Perform electric field simulation calculations on the three-dimensional gap simulation model after superimposing the bird foreign object disturbance model, extract the electric field intensity peak, electric field gradient change rate and equipotential line density as discharge risk indicators, integrate them to form a discharge risk map, and identify key risk areas where the local discharge risk value exceeds the set threshold.

[0121] This embodiment uses a three-dimensional finite element steady-state electric field simulation method to calculate the tower clearance model with a superimposed bird and foreign object disturbance model, and constructs a discharge risk map based on multi-dimensional indicators to identify key risk areas.

[0122] Symmetrical voltage boundary conditions were applied to each conductor endpoint in the model, with a typical voltage level of ±250 kV. Crossarms, hardware, and the conductive portion of the conductors were set as constant-potential conductor boundaries, while the ground was assigned zero potential. The insulator body was assigned a relative permittivity of 6.5 based on the material model. High-conductivity solid boundaries were set for the bird and drain cone to allow for localized charge accumulation and dissipation.

[0123] After the simulation is solved, the following electric field indicators are extracted in each grid cell:

[0124] Electric field strength : Export as scalar field;

[0125] Electric field gradient rate of change : Calculate the normalized modulus of the vector difference of the electric field change of adjacent units;

[0126] Density of equipotential lines : Divide the unit cube into several slices and count the number of equipotential surfaces And reflected as density value , or the inverse of the average spacing ,in is the unit volume, is the average equipotential surface distance.

[0127] In order to enhance the model's sensitivity to risk responses in the direction of the excretion path, a directional adjustment factor is introduced when calculating the rate of change of the electric field gradient. The specific steps are as follows:

[0128] Assume that the main direction vector of wind deviation is , the electric field gradient vector Project to Direction, define the directional gradient as: ;Will Replace the original scalar gradient term to participate in risk indicator fusion

[0129] Based on the scale differences of spatial regions (e.g., near the wire hanging point, electrode tip, and proximity to foreign objects), the three discharge risk indicators are fused using a hierarchical weighting strategy. The weighting formula is as follows:

[0130] ;

[0131] in 、 、 The risk weight factor corresponding to the regional type (such as the cutting-edge area ).

[0132] For the global three-dimensional risk function Calculate its mean and standard deviation , and set dynamic thresholds : ,in is the adjustment factor, which is empirically set to 1.2–1.5;

[0133] Identify all The spatial point set is aggregated as the key risk area to provide input basis for subsequent thickness optimization and local structure adjustment.

[0134] S4: Taking the thickness and coverage of the local insulating armor as design variables, an optimization function is constructed with the goals of minimizing the weighted integral value of the discharge risk map and minimizing the volume of the local insulating armor, including the first objective function and the second objective function, and fully covering the key risk area as the spatial constraint.

[0135] Based on the generated discharge risk map , introducing the electric field gradient mode length and equipotential line density There are two types of weighted indicators. The weighted factors are defined as 、 , construct the first objective function:

[0136] ;

[0137] Where, The spatial domain occupied by the key risk area; is a volume element in three-dimensional space; the risk map value is preprocessed by normalizing the simulation results.

[0138] The local insulating armor thickness function is defined as , volume as the second objective function:

[0139] ;

[0140] Where, is the coverage boundary function, which is automatically expanded from the boundary of the key risk area.

[0141] Thickness distribution function of local insulating armor Defined as:

[0142] ;

[0143] Where, is the thickness scaling factor, ; In key risk areas The maximum value of is the nonlinear risk enhancement index, ; is the minimum thickness constraint, .

[0144] Constraints include:

[0145] Electrical constraints: For each , the electric field strength of the simulation output Must meet: ,in The breakdown threshold corresponding to the 220kV voltage level is approximately 3.0 kV / mm.

[0146] Spatial constraint: The domain of all thickness functions is limited to the armor boundary function, and the area outside is set to zero.

[0147] S5: A multi-objective optimization algorithm is used to iteratively solve the optimization function to obtain the optimal parameter set that meets the insulation requirements. Based on the discharge risk map, the thickness of the local insulation armor is made to form a non-uniform distribution structure within the key risk area as the local discharge risk value increases. A structural design scheme including the armor thickness distribution and spatial position coordinates is output.

[0148] This embodiment adopts a multi-objective optimization strategy based on the non-dominated sorting genetic algorithm (NSGA-II) to complete the dual-objective optimization process of minimizing the discharge risk integral value and minimizing the local insulation armor volume, thereby generating an optimal insulation thickness distribution dataset while satisfying spatial and electrical constraints.

[0149] The optimization design variables consist of two parts:

[0150] Thickness scaling factor for local insulating armor , nonlinear risk enhancement index ;

[0151] Coverage range boundary function parameters, such as boundary function stretching coefficient and regional coordinate normalization control amount.

[0152] The above variables are combined through real number encoding to form an initial population, and each group of individuals is a set of candidate structural configuration parameters.

[0153] For population individuals where the design variables have already appeared, the corresponding objective function values ​​are directly read from the electric field simulation cache mapping table to avoid repeated simulation solutions. The gap simulation module is only called for the first occurrence of a new solution, outputting the electric field distribution and discharge risk map.

[0154] The local risk value in the current risk map exceeds the set threshold The coverage area is automatically adjusted upwards, using the empirical form: ,in is an empirical scaling factor (recommended range 0.1–0.3) that enhances the spatial responsiveness of local regions according to the gradient modulus.

[0155] In the area where the electric field intensity is abnormally concentrated, the thickness perturbation amplitude is adaptively scaled by the degree of local electric field anomaly, and the following suppression function is set: , in order to avoid excessive structural changes in the strong field peak area; is the thickness disturbance amplitude (disturbance amount), is the initial disturbance reference value, is the electric field gradient response adjustment factor.

[0156] According to the first objective function value in the current population The distribution variance is used to adjust the crossover probability and mutation probability in real time. For example, when the variance convergence slows down, the crossover probability is reduced and the mutation rate is increased to break through the local optimum.

[0157] In the elite individual set after non-dominated sorting, the area with the maximum gradient of the discharge risk map is extracted, and a local perturbation vector is constructed according to its direction for fine-tuning of the elite individuals to achieve the expansion of the solution set boundary in the gradient-sensitive direction.

[0158] Finally, the minimum material volume solution in the Pareto solution set that satisfies the space boundary constraint and the electric field intensity constraint is selected to generate the corresponding The discrete thickness distribution point set is exported as the structural deployment design input file.

[0159] like Figure 2 FIG. 1 is another embodiment of the present invention, which provides a tower gap insulation optimization design system considering bird-related fault characteristics, including:

[0160] Data input module, used to obtain the structural parameters and operating environment information of the target tower;

[0161] A 3D modeling module is used to construct a 3D gap simulation model including local insulating armor, and to superimpose a bird foreign body disturbance model on the 3D gap simulation model to describe the disturbance effect of bird activities on the gap electric field distribution;

[0162] The electric field analysis module is used to perform electric field simulation calculations on the three-dimensional gap simulation model after superimposing the bird foreign body disturbance model. It extracts the electric field intensity peak, electric field gradient change rate, and equipotential line density as discharge risk indicators, integrates them into a discharge risk map, and identifies key risk areas where the partial discharge risk value exceeds the set threshold;

[0163] An optimization function construction module is used to construct an optimization function based on the discharge risk map. The optimization objectives include minimizing the weighted integral value of the discharge risk map and minimizing the volume of the local insulating armor.

[0164] The parameter optimization module is used to iteratively solve the optimal parameter set using the thickness distribution function and coverage boundary function of the local insulating armor as design variables using a multi-objective optimization algorithm to meet spatial and electrical constraints.

[0165] The structure output module is used to generate a non-uniform distribution structure of armor thickness in the key risk area according to the optimization results, and output a data set containing thickness distribution and spatial position coordinates to guide the layout of local insulation armor,

[0166] In summary, the present invention has clear technical innovations and engineering practical value, significantly improving the protection capability of the transmission tower gap insulation structure against bird-related discharge risks, and is particularly suitable for local insulation optimization design scenarios in bird-sensitive areas of medium and high voltage lines.

[0167] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any person skilled in the art can easily conceive of various modifications or substitutions within the technical scope disclosed in this application, and such modifications or substitutions should be included within the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A tower gap insulation optimization design method considering bird-related fault characteristics, characterized in that: The method comprises: Obtain the structural parameters and operating environment information of the target tower and build a three-dimensional gap simulation model with local insulation armor; A bird foreign body disturbance model is superimposed on the three-dimensional gap simulation model to describe the disturbance effect of bird activity on the gap electric field distribution. The bird foreign body disturbance model includes a conductor that simulates bird interference and a conductive path that simulates the path of bird excrement. The conductive path generates a random conical trajectory based on bird excretion behavior statistics and wind disturbance. The electric field simulation calculation is performed on the three-dimensional gap simulation model after superimposing the bird foreign body disturbance model. The peak value of the electric field intensity, the rate of change of the electric field gradient, and the density of the equipotential lines are extracted as discharge risk indicators. These are integrated to form a discharge risk map, and key risk areas where the local discharge risk value exceeds the set threshold are identified. Taking the thickness and coverage of the local insulating armor as design variables, an optimization function is constructed with the objectives of minimizing the weighted integral value of the discharge risk map and minimizing the volume of the local insulating armor, with complete coverage of the critical risk area as a spatial constraint. A multi-objective optimization algorithm is used to iteratively solve the optimization function to obtain the optimal parameter set that meets the insulation requirements. According to the discharge risk map, the thickness of the local insulating armor is made to form a non-uniform distribution structure in the key risk area as the local discharge risk value increases. The structural design scheme including the armor thickness distribution and spatial position coordinates is output.

2. The tower gap insulation optimization design method considering bird-related fault characteristics according to claim 1 is characterized in that: The structural parameters include: the three-dimensional shape and tilt angle of the tower extracted by the directional tower detection model for synthetic aperture radar images, as well as the conductor hanging point height, crossarm length, phase spacing, and hardware geometric parameters of the target tower, where the hardware geometric parameters include the grading ring diameter, insulator string length, and hanging point coordinates; The operating environment information includes: insulator defect types and contamination levels identified based on the improved YOLOv7 model, as well as historical bird damage failure heat maps of the tower area, typical bird activity height statistics, regional meteorological parameters, and terrain background data. The regional meteorological parameters include annual average humidity, maximum wind speed, salt density distribution, and annual average number of thunderstorm days. The local insulating armor is a hydrophobic composite material covering the cross arm end and the hardware area, and the relative dielectric constant Satisfy the dynamic equation ,in It is a local field strength, and the surface hydrophobic contact angle is not less than 110 degrees.

3. The tower gap insulation optimization design method considering bird-related fault characteristics according to claim 2 is characterized in that: The directional tower detection model adopts sparse hint coding and shape adaptive feature fusion technology, specifically including: The model input is a synthetic aperture radar image of the target area and a set of initial prompt point coordinates. The prompt points are the possible locations of towers. The model frontend includes a sparse cue encoder that maps cue point coordinates into low-dimensional sparse vector embeddings that are fused with image features extracted by multiple Transformer encoding layers, using a cross-attention mechanism between point embeddings and image position embeddings. A shape-adaptive feature fusion module is set in the middle of the model to dynamically adjust the convolution kernel size and feature map sampling area according to the aspect ratio of the target candidate box; The model output is a detection result with rotation parameters including the direction angle and target frame coordinates; The improved YOLOv7 model specifically includes: The SPC module in the YOLOv7 backbone network is replaced by the RFB module, which contains multiple parallel channels with different convolution kernel sizes to enhance multi-scale feature perception. A coordinate attention mechanism is introduced into the Head output module, which includes a position encoding module and a space-channel decoupling attention module to enhance feature direction selectivity; In the bounding box regression part, the WIoU loss function is used to replace the original CIoU loss function, and the weight of the WIoU loss function is adaptively adjusted according to the quality of the predicted box.

4. The tower gap insulation optimization design method considering bird-related fault characteristics according to claim 2 is characterized in that: The method for constructing a three-dimensional gap simulation model containing local insulating armor includes: An initial three-dimensional gap simulation model was established using the finite element method. The initial coverage of the local insulation armor was preset in areas with a high incidence of bird damage on crossarms. The initial coverage range was 60% to 75% of the length from the conductor suspension string point to the crossarm end. Based on historical bird damage heat maps, terrain background data, and typical bird activity height statistics, adaptive grid densification technology was used in areas with high incidence of guano flashovers. The grid size of 10 cm in conventional areas was dynamically reduced to 2 cm in these areas. The correlation distribution data of wind speed and equivalent salt density were imported, and the pollution accumulation rate was calculated through fluid dynamics simulation. Combined with the identified insulator defect type and pollution level, the accuracy of the electric field boundary conditions of the three-dimensional gap simulation model was verified and the surface potential gradient distribution was corrected.

5. The tower gap insulation optimization design method considering bird-related fault characteristics according to claim 2 is characterized in that: Superimposing a bird foreign body disturbance model on the three-dimensional gap simulation model, the method comprising: A conductor is set up to simulate bird interference. The conductor is set to be an ellipsoid, with the size of the major axis in the range of 0.3 to 0.8 times the distance between the target towers, and the conductivity is set at 10 3 to 10 5 S / m, arranged in adjacent areas between cross-arm conductors; A conductive path is provided for simulating the falling path of bird excrement, wherein the conductive path is anchored at the center of the lower surface of the conductor and is a cone with a shape perturbation function, wherein the cone has a length between 0.4 meters and 1.2 meters, a top diameter less than 20 mm, and a bottom diameter greater than 40 mm; The centroid trajectory of the conductive path determines the starting height and dive angle according to the statistics of bird defecation behavior, and generates an offset trajectory in combination with the windage angle range set in the target area; The conductors and conductive paths are both superimposed in the three-dimensional gap simulation model in the form of three-dimensional entities, and participate in the electric field boundary condition setting and simulation calculation.

6. The tower gap insulation optimization design method considering bird-related fault characteristics according to claim 5 is characterized in that: The electric field simulation calculation is performed on the three-dimensional gap simulation model after superimposing the bird foreign body disturbance model, and the method includes: Perform 3D finite element steady-state electric field simulation and set conductor voltage and dielectric boundary conditions; The electric field intensity, electric field gradient change rate, and equipotential line density in the area around the insulator are extracted. The equipotential line density is calculated as the number of equipotential surfaces per unit volume or the inverse of the average spacing. A directional factor consistent with the set wind direction is introduced into the calculation of the electric field gradient change rate to correct the sensitivity of the discharge risk indicator to the direction of the bird excretion path. According to the scale characteristics of different spatial regions, graded weighted coefficients are set for the three discharge risk indicators, and the three indicators are integrated to form a discharge risk map. An adaptive threshold is set according to the mean and standard deviation of the discharge risk indicators in each area, and the spatial collection above the set threshold is identified as the key risk area.

7. The tower gap insulation optimization design method considering bird-related fault characteristics according to claim 1 is characterized in that: The optimization function includes a first objective function and a second objective function; The first objective function is the weighted integral value of the discharge risk map, expressed as: ; The second objective function is the volume of the local insulating armor, expressed as: ； The electrical constraints are: ; Where, The spatial domain occupied by the key risk area; is the discharge risk spectrum value; is the electric field gradient mode length, is the density of equipotential lines, 、 is the weighting factor; is the volume element in three-dimensional space; 、 are the thickness distribution function and coverage boundary function of the local insulating armor respectively; In the electric field simulation, the space point The electric field strength value at ; For line voltage level The corresponding critical value of the air gap breakdown electric field strength.

8. The tower gap insulation optimization design method considering bird-related fault characteristics according to claim 7 is characterized in that: Thickness distribution function of the local insulating armor Defined as: ； Where, is the thickness scaling factor; In key risk areas The maximum value of is the nonlinear risk enhancement index; is the minimum thickness constraint.

9. The tower gap insulation optimization design method considering bird-related fault characteristics according to claim 7, characterized in that: The iterative solution of the optimization function using a multi-objective optimization algorithm specifically includes: The thickness distribution function and coverage boundary function of the local insulating armor are encoded as design variables to generate the initial population; In each iteration, the three-dimensional gap simulation model is called to calculate the electric field distribution and discharge risk map corresponding to the design variables, and the objective function value is directly read through the result cache mapping table for repeated design variables; For the spatial points where the gradient amplitude in the discharge risk map exceeds the set threshold, the domain of the coverage boundary function is expanded in the gradient direction, and the expansion distance is ,in is the empirical scaling factor; For individuals with excessive electric field strength, the perturbation amplitude of the thickness distribution function is reduced; Update the thickness distribution function based on the current discharge risk map, and introduce an inertia factor to suppress mutations; Dynamically adjust the crossover rate and mutation rate according to the population objective function distribution entropy; Extract the maximum gradient direction of the discharge risk map in the elite solution to guide local variation; Output the minimum volume solution in the Pareto solution set that meets the spatial and electrical constraints, and generate spatial point set structural data of thickness distribution.

10. A tower gap insulation optimization design system considering bird-related fault characteristics, based on a tower gap insulation optimization design method considering bird-related fault characteristics according to any one of claims 1 to 9, characterized in that: include: Data input module, used to obtain the structural parameters and operating environment information of the target tower; a three-dimensional modeling module for constructing a three-dimensional gap simulation model including a local insulating armor based on the structural parameters, and superimposing a bird foreign body disturbance model on the three-dimensional gap simulation model to describe the disturbance effect of bird activities on the gap electric field distribution; The electric field analysis module is used to perform electric field simulation calculations on the three-dimensional gap simulation model after superimposing the bird foreign body disturbance model. It extracts the electric field intensity peak, electric field gradient change rate, and equipotential line density as discharge risk indicators, integrates them into a discharge risk map, and identifies key risk areas where the partial discharge risk value exceeds the set threshold; An optimization function construction module is used to construct an optimization function based on the discharge risk map. The optimization objectives include minimizing the weighted integral value of the discharge risk map and minimizing the volume of the local insulating armor. The parameter optimization module is used to iteratively solve the optimal parameter set using the thickness distribution function and coverage boundary function of the local insulating armor as design variables using a multi-objective optimization algorithm to meet spatial and electrical constraints. The structural output module is used to generate a non-uniformly distributed structure in which the armor thickness increases with the discharge risk value in the key risk area based on the optimization results. The module outputs a data set containing the thickness distribution and spatial position coordinates to guide the layout of local insulating armor.

Images