Power transmission line local wind field deviation calibration method, device and system

Abstract

The invention provides a power transmission line local wind field deviation calibration method, device and system. The method comprises the following steps: acquiring multi-source wind field data; constructing an anisotropic covariance model reflecting wind field directivity correlation and terrain non-stationary characteristics; generating background wind fields on the buffer zone grids on the two sides of the power transmission line by using wide-area observation data; correcting a background wind field by point location observation data, performing data fusion and deviation calibration based on an anisotropic covariance model, and outputting a calibrated wind field value of each grid point in the buffer zone grid and the uncertainty of the calibrated wind field value; according to the calibrated wind field value and the uncertainty thereof, calculating an overproof probability that the wind field along the power transmission line exceeds a preset equipment toughness threshold; the risk level map is generated based on the over-standard probability, the problem that the precision of power transmission line local wind field calibration is insufficient under the complex terrain is effectively solved, and the accuracy of buffer area wind field space distribution is remarkably improved through multi-source data fusion and anisotropic modeling.

Description

Technical Field

[0001] This invention relates to the field of power meteorological forecasting and early warning technology, and in particular to a method, device and system for calibrating local wind field deviations of transmission lines. Background Technology

[0002] As the backbone network of the power system, the safe and stable operation of transmission and distribution lines is crucial to the economy and people's livelihood. However, the corridors of these lines often traverse complex geographical environments such as valleys, slopes, coastlines, and suburbs. The local dynamic flow and thermal effects result in an exceptionally complex wind field structure near the ground (10-30 meters high), exhibiting strong spatial heterogeneity and anisotropy.

[0003] Currently, the perception and early warning of wind environment along transmission lines mainly rely on two types of data sources: First, macroscopic meteorological products, including numerical weather prediction, reanalysis data, and wind field information obtained from meteorological radar and satellite inversion. While these data can effectively characterize the wind field structure at medium and large scales, their spatial resolution is typically only a few kilometers to tens of kilometers, making it difficult to accurately capture near-surface wind shear, flow around, and acceleration effects caused by micro-topography and surface roughness variations within the transmission line scale. This results in significant deviations in the local wind field characterization at tower locations. Second, on-site wind measurement equipment, such as three-dimensional ultrasonic anemometers, can accurately acquire three-dimensional wind speed, wind direction, and turbulence parameters at a given location using high-frequency (1-10Hz) methods. Although their observation accuracy is high, limitations in cost and maintenance conditions mean they are usually only sparsely deployed along the transmission line, with limited representative radii for their observation points, making it impossible to directly form a continuous wind field covering the entire buffer zone of the transmission line. Summary of the Invention

[0004] This invention provides a method, apparatus, and system for calibrating local wind field deviations of transmission lines, in order to solve the technical problem of insufficient calibration accuracy of local wind fields of transmission lines under complex terrain in the prior art.

[0005] On one hand, the present invention provides a method for calibrating local wind field deviations in transmission lines, comprising: Acquire multi-source wind field data, including point observation data collected by ground-based fixed-point high-frequency anemometers and wide-area observation data collected by remote sensing wind field observation equipment. An anisotropic covariance model reflecting the directional correlation of wind field and the non-stationary characteristics of topography is constructed. Background wind fields are generated on the buffer zone grid on both sides of the transmission line using wide-area observation data; The background wind field is corrected using point observation data, and data fusion and bias calibration are performed based on an anisotropic covariance model. The calibrated wind field value and its uncertainty are output for each grid point in the buffer zone grid. Based on the calibrated wind field value and its uncertainty, the probability of the wind field along the transmission line exceeding the preset equipment toughness threshold is calculated. A risk level map is generated based on the probability of exceeding the standard.

[0006] According to the method for calibrating local wind field deviations of transmission lines provided by the present invention, an anisotropic covariance model reflecting the directional correlation of wind field and the non-stationary characteristics of terrain is constructed, as shown in the following formula: k(x,x′)=σ²·Matérn_ν(||Λ(θ)(x x′)|| / l); Where k(x,x′) is the anisotropic covariance function, θ is the prevailing wind direction, Λ(θ) is the scale matrix rotating with the prevailing wind direction θ, l is the spatial correlation scale parameter, σ² is the variance, ν is the smoothness parameter, x and x′ represent two different geographical locations, and ||Λ(θ)(x′)|| = 0. x′)|| represents the distance between two points after adjustment by the scaling matrix.

[0007] According to the method for calibrating local wind field deviation of transmission lines provided by the present invention, when the terrain area within the buffer zone meets the preset rules, the spatially relevant scale parameters are replaced by spatially variable scale functions, as shown in the following formula: l(x) = l0(z0(x) / z)^b; Where z0(x) is the surface roughness length at geographic location x, z is the reference roughness, b is the sensitivity parameter, and l0 is the reference correlation scale; The preset rules include at least one of the following: The terrain slope is greater than the set slope threshold; The land use type has changed. The spatial rate of change of the surface roughness length exceeds the set range.

[0008] According to the present invention, a method for calibrating local wind field deviations of transmission lines is provided, which generates a background wind field on a buffer zone grid on both sides of the transmission line using wide-area observation data, including: Wide-area observation data is uniformly interpolated onto a regular grid of buffer zone with a preset width, based on the centerline of the transmission line, using the Gaussian interpolation method, to form a spatially continuous background wind field.

[0009] According to the present invention, a method for calibrating local wind field deviations in transmission lines corrects the background wind field using point observation data, and performs data fusion and deviation calibration based on an anisotropic covariance model, outputting the calibrated wind field value and its uncertainty for each grid point within the buffer zone grid, including: Based on background wind field and point observation data, a residual field characterizing local bias is constructed; Based on Gaussian process regression, the residual field and anisotropic covariance model are integrated to calculate the posterior mean and posterior variance of the residual field at each grid point on the regular grid of the buffer zone. The posterior mean is superimposed with the background wind field to obtain the calibrated wind field value; The posterior variance of the residual field and the uncertainty of the background wind field are combined to output the total uncertainty of each grid point.

[0010] According to the present invention, a method for calibrating local wind field deviations in transmission lines is provided, which constructs a residual field characterizing local deviations based on background wind field and location observation data, including: At the location of the ground-based fixed-point high-frequency wind measurement equipment, the residuals between the observed wind field and the background wind field are calculated, and the residual field r(x) is assumed to follow a Gaussian process with zero mean and covariance defined by an anisotropic covariance model; the residuals are shown in the following formula: r_i = u_sonic(x_i) - m(x_i); Where r_i is the residual between the observed wind field and the background wind field, x_i is the location of the ground-based fixed-point high-frequency wind measurement equipment, u_sonic(x_i) is the observed wind field at the location, and m(x_i) is the background wind field. The posterior mean includes: ; in, For in position The posterior mean of the residuals; σ_s² is the covariance matrix between the test points and the observation points; K_ss is the covariance matrix between the observation points; σ_s² is the variance of the observation error; I is the identity matrix; r is a vector composed of the residuals at all observation points. Posterior variance includes: ; in, For position Prior variance at the location; for The transpose of the matrix; For the residual field at position The posterior variance; The calibrated wind field values ​​include: ; in, The calibrated wind field value; For in position Background wind field; The total uncertainty for each grid point includes: ; in, Background wind field at location Uncertainty at the location; For in position The variance of the background wind field.

[0011] According to the present invention, a method for calibrating local wind field deviations in transmission lines is provided, and the probability of exceeding the standard is shown in the following formula: p(x)=p(U(x)>T) =1 Φ((T μ(x)) / σ(x)); Where p(x) is the probability of exceeding the standard at geographical location x, U(x) is the actual wind field at location x; T is the preset equipment resilience threshold, μ(x) is the calibrated wind field value at geographical location x; σ(x) is the standard deviation of the uncertainty at geographical location x; and Φ is the cumulative distribution function of the standard normal distribution.

[0012] According to the method for calibrating local wind field deviations of transmission lines provided by the present invention, after acquiring multi-source wind field data, it further includes: Spatiotemporal registration and quality control of multi-source wind field data are performed, specifically including: Set a synchronization time window to perform time registration between point observation data and wide-area observation data; A unified map projection is used to map the node positions of point observation data to the grid points or wind vectors of wide-area observation data to the same coordinate system; For the point observation data, hard threshold filtering is performed based on the preset wind speed range and tilt angle limit, and outlier data points are removed using the sliding median absolute deviation method. For missing data in the point observation data, the neighborhood spatiotemporal interpolation method or the background wind field generated by wide-area observation data is used to fill in the missing data.

[0013] On the other hand, the present invention also provides a local wind field deviation calibration device for transmission lines, comprising: The data acquisition module acquires multi-source wind field data, including point observation data collected by ground-based fixed-point high-frequency anemometers and wide-area observation data collected by remote sensing wind field observation equipment. The model building module constructs an anisotropic covariance model that reflects the directional correlation of wind field and the non-stationary characteristics of terrain. The background wind field module generates a background wind field on the buffer zone grid on both sides of the transmission line using wide-area observation data. The data calibration module uses point observation data as high-weight observations and background wind field as prior. Based on the anisotropic covariance model, it fuses and calibrates the background wind field and outputs the calibrated wind field value and its uncertainty for each grid point in the buffer zone grid. The exceedance calculation module calculates the probability of the wind field along the transmission line exceeding the preset equipment toughness threshold based on the calibrated wind field value and its uncertainty. The risk level module generates a risk level map based on the probability of exceeding the standard.

[0014] In another aspect, the present invention also provides a local wind field deviation calibration system for transmission lines, wherein the system uses the aforementioned local wind field deviation calibration device for transmission lines.

[0015] The method, apparatus, and system for calibrating local wind field deviations in transmission lines provided by this invention effectively solve the problem of insufficient calibration accuracy of local wind fields in complex terrain. Through multi-source data fusion and anisotropic modeling, the accuracy of wind field spatial distribution in buffer zones is significantly improved. Based on the calculation of the probability of exceeding the limit due to uncertainty, high-risk hidden areas ignored by traditional methods can be identified, providing refined decision support for line wind protection reinforcement and operation and maintenance scheduling. The risk level map intuitively displays vulnerable sections of the line, helping to prioritize the deployment of inspection resources and reduce the incidence of wind-induced tripping accidents. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating the method for calibrating local wind field deviations in transmission lines provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of the transmission line local wind field deviation calibration device provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of the local wind field deviation calibration system for transmission lines provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0019] Figure 1This is a schematic flowchart of the local wind field deviation calibration method for transmission lines provided in an embodiment of the present invention.

[0020] See Figure 1 The method for calibrating local wind field deviations in transmission lines includes the following steps.

[0021] Step 101: Obtain multi-source wind field data, including point observation data collected by ground-based fixed-point high-frequency wind measurement equipment and wide-area observation data collected by remote sensing wind field observation equipment.

[0022] In this step, the point-based observation data comes from three-dimensional ultrasonic anemometers deployed along the transmission line, with sampling frequencies ranging from 1 to 10 Hz or higher, enabling the acquisition of near-surface wind speed, direction, and turbulence parameters with high spatiotemporal resolution. Wide-area observation data includes meteorological radar reflectivity / radial wind, satellite cloud motion wind, or scatterometer wind fields, characterized by broad spatial coverage. Combining high-frequency point-based observations with wide-area surface wind fields can compensate for the limitations of single data sources in terms of spatial representativeness or resolution. The initial weighting ratio of the data acquired by ultrasound, radar, and satellite can be approximately 1:0.3:0.2, or other ratios can be set.

[0023] Step 102: Construct an anisotropic covariance model that reflects the directional correlation of wind field and the non-stationary characteristics of terrain.

[0024] Among them, an anisotropic covariance model reflecting the directional correlation of wind field and the non-stationary characteristics of terrain is constructed, as shown in the following formula (1): k(x,x′)=σ²·Matérn_ν(||Λ(θ)(x x′)|| / l)(1; Where k(x,x′) is the anisotropic covariance function, θ is the prevailing wind direction, Λ(θ) is the scale matrix rotating with the prevailing wind direction θ, l is the spatial correlation scale parameter, σ² is the variance, ν is the smoothness parameter, x and x′ represent two different geographical locations, and ||Λ(θ)(x′)|| = 0. x′)|| represents the distance between two points after adjustment by the scaling matrix.

[0025] In this embodiment, the anisotropic covariance function refers to a function that can adjust the spatial correlation structure according to the wind direction. Specifically, it can be implemented by transforming the geographical distance using a rotating scale matrix. By introducing the dominant wind direction θ, the covariance model can reflect the spatial correlation differences when the wind field propagates in different directions. The dominant wind direction θ refers to the wind direction angle that plays a dominant role in the wind field of the target area. It can be determined through historical observation data or numerical simulation results and is used to adjust the directional dependence of the covariance function. The scale matrix Λ(θ) is a linear transformation matrix that rotates according to the dominant wind direction θ. Specifically, it can be implemented through a combination of two-dimensional coordinate rotation transformation and anisotropic stretching parameters. It is used to convert geographical coordinates into stretched coordinate systems along and perpendicular to the wind direction, thereby reflecting the directional characteristics of wind field propagation. The spatial correlation scale parameter l is a parameter that controls the decay rate of the covariance function. It can be obtained through empirical estimation or optimization using the maximum likelihood method and is used to describe the effective range of the spatial correlation of the wind field.

[0026] Specifically, in the local wind field modeling of transmission lines, the spatial correlation of the wind field varies significantly in different directions due to the effects of terrain obstruction and wind deflection acceleration. For example, when the prevailing wind direction is southeast, the propagation distance of the wind field along this direction may be significantly greater than that in the vertical direction. By constructing an anisotropic covariance model, the geographical locations x and x′ are first substituted into the scaling matrix Λ(θ) for linear transformation to obtain the adjusted distance stretched along the prevailing wind direction. Subsequently, the Matérn covariance function is used to calculate the covariance value between the two points, where the smoothness parameter ν (e.g., preferably ν∈{3 / 2,5 / 2}) controls the differentiability of the covariance function, and the variance σ² characterizes the overall fluctuation intensity of the wind field. This model can effectively capture the extension correlation of the wind field along the prevailing wind direction and the rapid attenuation characteristics of the vertical wind direction, thus more accurately describing the non-stationary wind field structure under complex terrain.

[0027] In this embodiment, by introducing a direction-dependent scaling matrix and combining it with the Matérn function, quantitative modeling of the directional propagation characteristics of local wind fields along transmission lines is achieved. This ensures that the fusion process of background wind fields and observational data accurately reflects the spatial heterogeneity of the actual wind field. This embodiment solves the problem of local wind field estimation bias caused by neglecting the directional correlation of wind fields, significantly improving the accuracy of spatial interpolation of wind fields in complex terrain areas. For example, in scenarios where transmission lines cross mountain ridges, this model can accurately characterize the acceleration effect on the windward slope and the turbulent dissipation characteristics on the leeward slope, providing a reliable data foundation for subsequent wind-induced disaster risk assessment.

[0028] Step 103: Using wide-area observation data, generate background wind fields on the buffer zone grid on both sides of the transmission line.

[0029] In this step, background wind field generation refers to the process of interpolating wide-area data onto the line buffer grid, for example, converting satellite wind vector data into a continuous wind field on a regular grid using Gaussian interpolation. Specifically, using wide-area observation data, background wind fields are generated on the buffer zone grid on both sides of the transmission line, including: Wide-area observation data is uniformly interpolated onto a regular grid of buffer zone with a preset width, based on the centerline of the transmission line, using the Gaussian interpolation method, to form a spatially continuous background wind field.

[0030] Among them, the Gaussian interpolation method refers to a probabilistic interpolation technique based on Gaussian processes. Specifically, it can be implemented by combining the covariance matrix with the observation error model. By introducing a spatial correlation model, discrete wide-area observation data is smoothed to generate a continuously distributed background wind field. The buffer zone regular grid refers to a rectangular grid area formed by expanding outward from the centerline of the transmission line. Specifically, it can be divided using preset width parameters, such as 500 meters to 5000 meters (preferably 2000 meters), and combined with grid resolution parameters (such as 50-200 meters, preferably 100 meters) to ensure coverage of potential wind field anomaly areas on both sides of the line.

[0031] Specifically, when mapping discrete wide-area observation data onto a buffer zone regular grid using Gaussian interpolation, an interpolation weight matrix based on spatial covariance structure is first established. For example, an anisotropic covariance model is used to calculate the spatial correlation between grid points and observation points. Then, a background wind field estimate for each grid point is generated through linear combination. This process effectively integrates the spatial distribution characteristics of wide-area observation data, forming a high-resolution background wind field consistent with the transmission line alignment, providing a spatially continuous reference field for subsequent calibration of point observation data.

[0032] In this embodiment, by combining Gaussian interpolation with a covariance model, the interpolation weights can be adaptively adjusted. For example, the spatial correlation weights along the wind direction can be enhanced under the prevailing wind direction, thereby more accurately reflecting the spatial distribution pattern of the actual wind field. This application can transform low-resolution wide-area observation data into a high-resolution background wind field that matches the spatial distribution of transmission lines, effectively solving the problem of mismatch between the background field and the line alignment.

[0033] Step 104: Correct the background wind field using point observation data, and perform data fusion and bias calibration based on the anisotropic covariance model, outputting the calibrated wind field value and its uncertainty for each grid point in the buffer zone grid.

[0034] In this step, data fusion and bias calibration refers to the statistical optimization process of correcting the background field using point observations. For example, the posterior distribution of the residual field is calculated based on Gaussian process regression, and the background field is superimposed to obtain the calibration result. The calibration process may include decomposing the wind vector into a westerly component (abbreviated as u') and a southerly component (abbreviated as v'), and then performing subsequent bias calculations, data fusion, and calibration on the westerly component (abbreviated as u') and the southerly component (abbreviated as v') respectively, outputting the calibrated wind field value and its uncertainty for each grid point within the buffer zone grid.

[0035] Step 105: Based on the calibrated wind field value and its uncertainty, calculate the probability of the wind field along the transmission line exceeding the preset equipment toughness threshold.

[0036] In this step, the probability of exceeding the limit calculation refers to quantifying the likelihood that the wind field will exceed the equipment's withstand threshold, for example, by calculating the cumulative probability based on the normal distribution characteristics of the calibration values. Preset equipment resilience thresholds can generally include icing, galloping, or wind resistance limits.

[0037] Step 106: Generate a risk level map based on the probability of exceeding the standard.

[0038] In this step, risk level map generation refers to mapping the probability results to visual risk zones, such as using four colors—red, orange, yellow, and blue—to represent different risk levels. Generally, the higher the probability of exceeding the limit, the higher the corresponding risk level. For example, L1 (p<5%), L2 (5%–20%), L3 (20%–40%), L4 (40%–70%), and L5 (≥70%), where L1 to L5 represent risk levels, and the percentages following them indicate the probability of exceeding the limit.

[0039] This embodiment effectively solves the problem of insufficient accuracy in local wind field calibration for transmission lines in complex terrain. Through multi-source data fusion and anisotropic modeling, the accuracy of wind field spatial distribution in buffer zones is significantly improved. Based on the calculation of the probability of exceeding uncertainties, high-risk hidden areas ignored by traditional methods can be identified, providing refined decision support for line wind protection reinforcement and operation and maintenance scheduling. The risk level map intuitively displays vulnerable sections of the line, helping to prioritize the deployment of inspection resources and reduce the incidence of wind-induced tripping accidents.

[0040] In one embodiment of this specification, when the terrain area within the buffer zone meets the preset rules, the spatially related scale parameter is replaced by a spatially variable scale function, as shown in the following formula (2): l(x) = l0(z0(x) / z)^b(2; Where z0(x) is the surface roughness length at geographic location x, z is the reference roughness (the reference roughness can be understood as farmland roughness, which can be set to 0.1m), b is the sensitivity parameter (which can be set to -0.4); and l0 is the reference correlation scale. The preset rules include at least one of the following: The terrain slope is greater than the set slope threshold; The land use type has changed. The spatial rate of change of the surface roughness length exceeds the set range.

[0041] In this embodiment, the spatial scaling function refers to a mathematical expression that dynamically adjusts the spatial correlation scale based on the surface roughness length. Specifically, the ratio of surface roughness to reference roughness can be used as an adjustment factor, and nonlinear scaling can be achieved through an exponential function. The surface roughness length is a parameter reflecting the strength of the frictional effect of surface cover on airflow, and can be obtained through field measurements or remote sensing inversion methods. The reference correlation scale refers to the baseline spatial correlation length under standard surface roughness conditions, and can be determined statistically based on historical observation data. The sensitivity parameter is a coefficient that controls the degree of influence of surface roughness changes on the spatial correlation scale, and can be calibrated through parameter optimization or machine learning methods.

[0042] Specifically, when the terrain slope exceeds a set threshold, surface airflow is significantly lifted or subsided by the terrain, causing abrupt changes in the spatial correlation of the wind field. At this point, fixed spatial correlation scale parameters cannot accurately characterize the terrain effect. By introducing a spatially variable scaling function, the correlation scale is dynamically adjusted based on the ratio of surface roughness length to a reference value, enabling the model for undulating terrain areas to adapt to changes in surface friction effects. For example, at the boundary between woodland with high roughness and smooth grassland, the difference in surface roughness length causes the spatial correlation scale to automatically shrink, thus capturing abrupt wind field changes more precisely. The land use type switching trigger mechanism in the preset rules can identify areas of abrupt changes in vegetation cover or building distribution, avoiding model errors caused by discontinuities in surface attributes.

[0043] In this embodiment, by introducing a spatially variable scaling function, relevant scale parameters can be dynamically adjusted according to real-time surface roughness, ensuring the model maintains physical consistency even under non-uniform terrain conditions and effectively reducing wind field estimation errors caused by fixed parameter assumptions. This application solves the problem of mismatch between spatially relevant scale parameters and surface properties in wind field modeling of complex terrain areas, improving the accuracy of wind field calibration in transmission line buffer zones. By dynamically adjusting the spatially relevant scale, the modulation effect of surface roughness changes on the near-surface wind field can be more accurately reflected, reducing the uncertainty in wind speed estimation in sensitive areas such as steep slopes and vegetation transition zones.

[0044] In one embodiment of this specification, the background wind field is corrected using point observation data, and data fusion and bias calibration are performed based on an anisotropic covariance model. The calibrated wind field value and its uncertainty for each grid point within the buffer zone grid are output, including: Based on background wind field and point observation data, a residual field characterizing local bias is constructed; Based on Gaussian process regression, the residual field and anisotropic covariance model are integrated to calculate the posterior mean and posterior variance of the residual field at each grid point on the regular grid of the buffer zone. The posterior mean is superimposed with the background wind field to obtain the calibrated wind field value; The posterior variance of the residual field and the uncertainty of the background wind field are combined to output the total uncertainty of each grid point.

[0045] The correction process may include decomposing the wind vector into a westerly component (u') and a southerly component (v'), and then performing subsequent deviation calculations, data fusion, and calibration on the westerly component (u') and the southerly component (v') respectively. The residual field refers to the field quantity constructed by the difference between the observed wind field and the background wind field at a fixed-point high-frequency anemometer location. Specifically, it can be calculated by point-by-point difference between the measured value and the background value at the observation point, and is used to characterize the systematic deviation of the background wind field under the influence of local micro-topography. Gaussian process regression is a data fusion method based on probability and statistics theory, specifically implemented using covariance matrix operations and Bayesian inference, used to optimally estimate the residual field while considering the spatial correlation of the wind field. The posterior mean and posterior variance refer to the conditional expectation and variance of the residual field obtained through Gaussian process regression at unobserved locations, specifically implemented by covariance matrix inversion and linear combination operations, used to quantify the spatial distribution and uncertainty of the residual field. Total uncertainty refers to the comprehensive error range of the calibrated wind field values. Specifically, it can be achieved by superimposing the posterior variance of the residual field with the variance of the background wind field, and is used to reflect the cumulative effect of multi-source data errors during the calibration process.

[0046] In this embodiment, by introducing an anisotropic covariance model, the spatial variation of wind field residuals with prevailing wind direction and surface roughness is explicitly characterized, making the data fusion process more consistent with actual physical mechanisms. By superimposing posterior variance and background variance, the entire chain of uncertainty quantification is achieved. This application effectively solves the problem of fusing sparse anemometer data with wide-area background wind fields in complex terrain areas, significantly improving the spatial consistency and accuracy of local wind field calibration for transmission lines. Through residual field construction and Gaussian process regression, the wind field distortion characteristics caused by micro-topography can be accurately captured, avoiding the systematic bias generated by traditional interpolation methods in unobserved areas. Simultaneously, the quantified output of total uncertainty provides a risk decision basis for line wind protection design, supporting maintenance personnel in identifying high-risk sections of wind-induced disasters and formulating differentiated prevention and control strategies.

[0047] In one embodiment of this specification, a residual field characterizing local bias is constructed based on background wind field and location observation data, including: At the location of the ground-based fixed-point high-frequency wind measurement equipment, the residuals between the observed wind field and the background wind field are calculated, and the residual field is assumed to follow a Gaussian process with zero mean and covariance defined by an anisotropic covariance model; the residuals between the observed wind field and the background wind field are shown in the following formula (3): r_i=u_sonic(x_i)-m(x_i)(3); Where r_i is the residual between the observed wind field and the background wind field, x_i is the location of the ground-based fixed-point high-frequency wind measurement equipment, u_sonic(x_i) is the observed wind field at the location, and m(x_i) is the background wind field. At the location of the ground-based fixed-point high-frequency wind measurement equipment, the wind vector can be decomposed into a westerly component (u') and a southerly component (v'), and the deviation of the westerly component (u') and the southerly component (v') are calculated, data is fused, and calibrated respectively.

[0048] The posterior mean is shown in the following formula (4): (4); in, For in position The posterior mean of the residuals; σ_s² is the covariance matrix between the test points and the observation points; K_ss is the covariance matrix between the observation points; σ_s² is the variance of the observation error; I is the identity matrix; r is a vector composed of the residuals at all observation points. The posterior variance is shown in formula (5) below: (5); in, For position Prior variance at the location; for The transpose of the matrix; For the residual field at position The posterior variance; The calibrated wind field value is shown in the following formula (6): (6); in, The calibrated wind field value; For in position Background wind field; The total uncertainty for each grid point is shown in the following formula (7): (7); in, Background wind field at location Uncertainty at the location; For in position The variance of the background wind field.

[0049] In one embodiment of this specification, the probability of exceeding the standard is shown in the following formula (8): p(x)=p(U(x)>T) =1 Φ((T μ(x)) / σ(x))(8; Where p(x) is the probability of exceeding the standard at geographical location x, U(x) is the actual wind field at location x; T is the preset equipment resilience threshold, μ(x) is the calibrated wind field value at geographical location x; σ(x) is the standard deviation of the uncertainty at geographical location x; and Φ is the cumulative distribution function of the standard normal distribution. p(x) and 1 Φ((T μ(x) / σ(x) do not need to be exactly the same; they only need to be approximately the same. For example, the error between them should be less than 5%, meaning that p(x) = p(U(x) > T) ≈ 1. Φ((T μ(x)) / σ(x)) can be obtained.

[0050] In this embodiment, the calibrated wind field value μ(x) refers to the corrected wind field estimate obtained by fusing background wind field and point observation data. Specifically, it can be achieved by calculating the posterior mean of the residual field using Gaussian process regression and superimposing it with the background wind field, reflecting the spatial distribution of the wind field after correction for topography and local effects. The standard deviation σ(x) of the uncertainty is a quantitative index of the confidence of the wind field estimate, which can be calculated by fusing the posterior variance of the residual field and the uncertainty of the background wind field. It characterizes the combined impact of data error and model error during the calibration process. The preset equipment resilience threshold T can refer to the maximum wind speed limit that the transmission line equipment can withstand under specific design standards. Specifically, it can be set according to the line design parameters or historical fault data, for example, using 1.2 times the line design wind speed as the threshold. The cumulative distribution function Φ of the standard normal distribution is a mathematical tool used to calculate the probability after standardizing the wind field deviation, which can be achieved through table lookup or numerical integration methods.

[0051] Specifically, after completing the wind field data calibration, for each location x within the transmission line buffer zone grid, the calibrated wind field value μ(x) and its uncertainty standard deviation σ(x), combined with a preset equipment resilience threshold T, are used to calculate the probability that the actual wind field U(x) exceeds T through a probabilistic model. This model assumes that the actual wind field follows a normal distribution with mean μ(x) and standard deviation σ(x). By calculating 1... Φ((T The probability of exceeding the standard, p(x), is obtained by dividing μ(x) by σ(x). For example, when μ(x) is close to T and σ(x) is large, p(x) increases significantly, indicating a high risk of wind disaster at that location. The resulting risk level map can intuitively display the risk distribution along the line, providing a quantitative basis for operation and maintenance decisions.

[0052] In this embodiment, by introducing a probabilistic model, the calibrated wind field value and its confidence level are jointly used to calculate the probability of exceeding the standard, which can more objectively quantify wind disaster risk and avoid misjudgment or omission due to ignoring uncertainty. This application solves the problem of insufficient early warning accuracy caused by wind field estimation bias and lack of uncertainty. By quantifying the probability of exceeding the standard, high-risk areas can be identified and reinforcement measures can be prioritized.

[0053] In one embodiment of this specification, after acquiring multi-source wind field data, the method further includes: Spatiotemporal registration and quality control of multi-source wind field data.

[0054] Specifically, spatiotemporal registration and quality control of multi-source wind field data include: Set a synchronization time window to perform time registration between point observation data and wide-area observation data; A unified map projection is used to map the node positions of point observation data to the grid points or wind vectors of wide-area observation data to the same coordinate system; For the point observation data, hard threshold filtering is performed based on the preset wind speed range and tilt angle limit, and outlier data points are removed using the sliding median absolute deviation method. For missing data in the point observation data, the neighborhood spatiotemporal interpolation method or the background wind field generated by wide-area observation data is used to fill in the missing data.

[0055] Synchronization time window refers to aligning data collected by different observation devices in the time dimension. This can be achieved using a sliding window mechanism with fixed time intervals, such as matching the timestamps of observation data in minutes, to resolve time discrepancies caused by differences in device sampling frequencies. Unified map projection refers to converting spatial data from different sources to the same geographic coordinate system. This can be achieved using a universal transverse Mercator projection or a Lambert conformal conic projection, eliminating spatial misalignment caused by coordinate system differences. Hard threshold filtering involves setting an effective range for wind speed and wind direction angle based on physical laws. Specifically, the upper limit for wind speed can be set to 95% of the meteorological sensor's range, and the effective range for wind direction angle can be 0-360 degrees, directly eliminating abnormal data exceeding reasonable ranges. The sliding median absolute deviation method calculates the median and absolute deviation of a data sequence using a sliding time window. Specifically, a sliding interval with a window length of 5 minutes can be used to identify and eliminate outliers deviating from the median by more than three times the absolute deviation, suppressing the impact of short-term interference noise. Neighborhood spatiotemporal interpolation refers to filling missing data with observations from adjacent time periods or nearby spatial points. Specifically, Kriging interpolation or inverse distance weighted interpolation algorithms can be used to construct an interpolation model within the spatiotemporal neighborhood. Background wind field filling refers to using continuous wind fields generated from wide-area observation data as substitute values ​​for missing areas. Specifically, Gaussian process regression can be used to generate background fields to fill the gaps.

[0056] Specifically, the spatiotemporal registration and quality control process first eliminates time discrepancies in multi-source data through a time synchronization window, for example, aligning satellite remote sensing data and ground-based anemometer data with minute-level timestamps. Then, the geographic coordinates of the point observation data are transformed to a projected coordinate system consistent with the wide-area observation data, for example, converting the latitude and longitude coordinates of the 3D ultrasonic anemometer to UTM coordinates. Hard threshold filtering is applied to the point data, for example, marking records with wind speeds exceeding 40 m / s or wind direction angles exceeding 0-360 degrees as invalid data. When using the sliding median absolute deviation method, for example, calculating the median of the wind speed sequence within a 5-minute window, if a data point deviates from the median by more than 3 times the absolute deviation, it is identified as an outlier and removed. For missing data, if the missing duration is less than 10 minutes, linear interpolation is performed using observations from three adjacent time steps; if the missing area is covered by wide-area observation data, values ​​from the corresponding location in the background wind field are used to fill in the gaps.

[0057] In this embodiment, multi-step collaborative processing achieves consistent calibration of multi-source data in both time and space dimensions, and a composite quality control mechanism enhances data reliability. This application addresses the fusion errors caused by differences in spatiotemporal references and quality issues in multi-source wind field data, improving the spatiotemporal consistency of data sources and providing high-precision input for subsequent wind field deviation calibration. For example, time registration eliminates wind speed phase deviations caused by asynchronous device sampling frequencies, unified spatial projection avoids the impact of coordinate transformation errors on wind field interpolation, and the composite quality control mechanism reduces the interference of abnormal data on background field generation and residual calculation, thereby ensuring the reliability of the final calibrated wind field.

[0058] The following example illustrates the specific steps for correcting the background wind field using point observation data, performing data fusion and bias calibration based on an anisotropic covariance model, and outputting the calibrated wind field value and its uncertainty for each grid point within the buffer zone grid: Step 1: Wind field vector decomposition: The observation data at the aforementioned locations and the wind vector in the background wind field are decomposed into a westerly component (u') and a southerly component (v'), respectively.

[0059] Step 2: Construct the residual field characterizing local bias: For the u' component and v' component respectively, based on the background wind field and point observation data, residual fields characterizing local bias are constructed; that is, the residual field r_u(x) of the u' component and the residual field r_v(x) of the v' component are calculated.

[0060] Step 3: Calculate the posterior statistic of the residual field: Based on Gaussian process regression, the residual fields of the u' component and the v' component are respectively fused with the anisotropic covariance model. The posterior mean and posterior variance of the residual field of the u' component and the posterior mean and posterior variance of the residual field of the v' component are calculated in parallel at each grid point on the regular grid of the buffer zone.

[0061] Step 4: Obtain the calibrated wind field components: The posterior mean of the u' component is superimposed on the u' component of the background wind field to obtain the calibrated westerly component value. The posterior mean of the v' component is then superimposed with the v' component of the background wind field to obtain the calibrated southerly wind component value. .

[0062] Step 5: Wind field vector synthesis: The calibrated westerly component value for each grid point South wind component value Vector synthesis is performed to calculate the final wind speed and wind direction values ​​for the grid point, which are then used as the calibrated wind field values.

[0063] Step 6: Output the total uncertainty: The residual field posterior variances of the u' and v' components are fused (e.g., by taking the root of the square or by calculating according to the law of error propagation), and then superimposed with the uncertainty of the background wind field to output the total uncertainty of each grid point.

[0064] Based on the same general inventive concept, this invention also protects a local wind field deviation calibration device for transmission lines, such as... Figure 2 As shown, Figure 2 This is a schematic diagram of the structure of the local wind field deviation calibration device for transmission lines provided in an embodiment of the present invention. The local wind field deviation calibration device for transmission lines provided by the present invention will be described below. The local wind field deviation calibration device for transmission lines described below can be referred to in correspondence with the local wind field deviation calibration method for transmission lines described above.

[0065] The local wind field deviation calibration device for transmission lines includes: The data acquisition module 201 acquires multi-source wind field data, including point observation data collected by ground-based fixed-point high-frequency wind measurement equipment and wide-area observation data collected by remote sensing wind field observation equipment. Model building module 202 constructs an anisotropic covariance model that reflects the directional correlation of wind field and the non-stationary characteristics of terrain; Background wind field module 203 uses wide-area observation data to generate background wind field on the buffer zone grid on both sides of the transmission line; The data calibration module 204 uses point observation data as high-weight observations, background wind field as prior, and anisotropic covariance model to fuse and calibrate the background wind field, and outputs the calibrated wind field value and its uncertainty for each grid point in the buffer zone grid. The exceedance calculation module 205 calculates the probability of the wind field along the transmission line exceeding the preset equipment toughness threshold based on the calibrated wind field value and its uncertainty. The risk level module 206 generates a risk level map based on the probability of exceeding the standard.

[0066] Based on the same general inventive concept, this invention also protects a local wind field deviation calibration system for transmission lines, such as... Figure 3 As shown, Figure 3 This is a schematic diagram of the structure of the local wind field deviation calibration system for power transmission lines provided in an embodiment of the present invention.

[0067] The transmission line local wind field deviation calibration system 300 uses the transmission line local wind field deviation calibration device 200 described in the above embodiments. The hardware foundation of this system consists of a distributed sensing and computing network, whose core task is to provide a data input and computing platform for the wind field deviation calibration device (i.e., software functional modules). Specifically, ground-based high-frequency anemometers deployed along the transmission line serve as precise location data sources, while connected remote sensing equipment such as meteorological radar and satellites serve as wide-area wind field information sources. The raw data collected by these hardware devices is transmitted to a central server or cloud computing platform through a communication network. It is on this hardware platform that the core device for data fusion, model building, and deviation calibration runs. Finally, the risk level map generated by this device is distributed to the display terminal or mobile inspection equipment in the monitoring center, completing the entire process from data to decision.

[0068] Figure 4 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention.

[0069] like Figure 4 As shown, the electronic device may include a processor 410, a communication interface 420, a memory 430, and a communication bus 440. The processor 410, communication interface 420, and memory 430 communicate with each other via the communication bus 440. The processor 410 can call logic instructions from the memory 430 to execute a local wind field deviation calibration method for transmission lines.

[0070] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0071] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the transmission line local wind field deviation calibration method provided by the above methods.

[0072] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the transmission line local wind field deviation calibration method provided by the above methods.

[0073] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0074] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0075] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for calibrating local wind field deviation in transmission lines, characterized in that, include: Acquire multi-source wind field data, including point observation data collected by ground-based fixed-point high-frequency anemometers and wide-area observation data collected by remote sensing wind field observation equipment. An anisotropic covariance model reflecting the directional correlation of wind field and the non-stationary characteristics of topography is constructed. Using the wide-area observation data, a background wind field is generated on the buffer zone grid on both sides of the transmission line; The background wind field is corrected using the observation data at the aforementioned points, and data fusion and bias calibration are performed based on the anisotropic covariance model. The calibrated wind field value and its uncertainty for each grid point within the buffer zone grid are then output. Based on the calibrated wind field value and its uncertainty, the probability of the wind field along the transmission line exceeding the preset equipment toughness threshold is calculated. A risk level map is generated based on the aforementioned probability of exceeding the standard.

2. The method for calibrating local wind field deviation of transmission lines according to claim 1, characterized in that, The anisotropic covariance model that reflects the directional correlation of wind field and the non-stationary characteristics of terrain is constructed as shown in the following formula: k(x,x′)=σ²·Matérn_ν(||Λ(θ)(x x′)|| / l); Where k(x,x′) is the anisotropic covariance function, θ is the prevailing wind direction, Λ(θ) is the scale matrix rotating with the prevailing wind direction θ, l is the spatial correlation scale parameter, σ² is the variance, ν is the smoothness parameter, x and x′ represent two different geographical locations, and ||Λ(θ)(x′)|| = 0. x′)|| represents the distance between two points after adjustment by the scaling matrix.

3. The method for calibrating local wind field deviation of transmission lines according to claim 2, characterized in that, When the terrain region within the buffer zone meets the preset rules, the spatially relevant scale parameters are replaced by spatially variable scale functions, as shown in the following formula: l(x) = l0(z0(x) / z)^b; Where z0(x) is the surface roughness length at geographic location x, z is the reference roughness, b is the sensitivity parameter, and l0 is the reference correlation scale; The preset rules include at least one of the following: The terrain slope is greater than the set slope threshold; The land use type has changed. The spatial rate of change of the surface roughness length exceeds the set range.

4. The method for calibrating local wind field deviation of transmission lines according to claim 1, characterized in that, The process of generating a background wind field on the buffer zone grid on both sides of the transmission line using the wide-area observation data includes: The wide-area observation data is uniformly interpolated onto a regular grid of buffer zone with a preset width, based on the centerline of the transmission line, using the Gaussian interpolation method, to form a spatially continuous background wind field.

5. The method for calibrating local wind field deviation of transmission lines according to claim 1, characterized in that, The background wind field is corrected using the observed data at the stated locations, and data fusion and bias calibration are performed based on the anisotropic covariance model. The calibrated wind field value and its uncertainty for each grid point within the buffer zone grid are output, including: Based on the background wind field and the location observation data, a residual field characterizing local deviation is constructed. Based on Gaussian process regression, the residual field and the anisotropic covariance model are fused to calculate the posterior mean and posterior variance of the residual field at each grid point on the buffer zone regular grid. The posterior mean is superimposed with the background wind field to obtain the calibrated wind field value; The posterior variance of the residual field is combined with the uncertainty of the background wind field to output the total uncertainty of each grid point.

6. The method for calibrating local wind field deviation of transmission lines according to claim 5, characterized in that, Based on the background wind field and the location observation data, a residual field characterizing local bias is constructed, including: At the location of the ground-based fixed-point high-frequency wind measurement equipment, the residuals between the observed wind field and the background wind field are calculated, and the residual field r(x) is assumed to follow a Gaussian process with zero mean and covariance defined by an anisotropic covariance model; the residuals are shown in the following formula: r_i = u_sonic(x_i) - m(x_i); Where r_i is the residual between the observed wind field and the background wind field, x_i is the location of the ground-based fixed-point high-frequency wind measurement equipment, u_sonic(x_i) is the observed wind field at the location, and m(x_i) is the background wind field. The posterior mean includes: ； in, For in position The posterior mean of the residuals; σ_s² is the covariance matrix between the test points and the observation points; K_ss is the covariance matrix between the observation points; σ_s² is the variance of the observation error; I is the identity matrix; r is a vector composed of the residuals at all observation points. Posterior variance includes: ； in, For position Prior variance at the location; for The transpose of the matrix; For the residual field at position The posterior variance; The calibrated wind field values ​​include: ； in, The calibrated wind field value; For in position Background wind field; The total uncertainty for each grid point includes: ； in, Background wind field at location Uncertainty at the location; For in position The variance of the background wind field.

7. The method for calibrating local wind field deviation of transmission lines according to claim 1, characterized in that, The probability of exceeding the standard is shown in the following formula: p(x)=p(U(x)>T)=1 Φ((T μ(x)) / σ(x)); Where p(x) is the probability of exceeding the standard at geographical location x, U(x) is the actual wind field at location x; T is the preset equipment resilience threshold, μ(x) is the calibrated wind field value at geographical location x; σ(x) is the standard deviation of the uncertainty at geographical location x; and Φ is the cumulative distribution function of the standard normal distribution.

8. The method for calibrating local wind field deviation of transmission lines according to claim 1, characterized in that, After acquiring the multi-source wind field data, the following is also included: Spatiotemporal registration and quality control of multi-source wind field data are performed, specifically including: Set a synchronization time window to perform time registration between the point observation data and the wide-area observation data; A unified map projection is used to map the node positions of the point observation data to the grid points or wind vectors of the wide-area observation data to the same coordinate system; The observation data at the aforementioned locations are filtered using a hard threshold based on a preset wind speed range and tilt angle limit, and outlier data points are removed using the sliding median absolute deviation method. For missing data in the point observation data, the missing data is filled in using the neighborhood spatiotemporal interpolation method or the background wind field generated from the wide-area observation data.

9. A device for calibrating local wind field deviation in transmission lines, characterized in that, include: The data acquisition module acquires multi-source wind field data, including point observation data collected by ground-based fixed-point high-frequency wind measurement equipment and wide-area observation data collected by remote sensing wind field observation equipment. The model building module constructs an anisotropic covariance model that reflects the directional correlation of wind field and the non-stationary characteristics of terrain. The background wind field module generates a background wind field on the buffer zone grid on both sides of the transmission line using the wide-area observation data. The data calibration module uses the point observation data as high-weight observations, the background wind field as a priori, and the anisotropic covariance model to perform fusion and bias calibration on the background wind field, and outputs the calibrated wind field value and its uncertainty for each grid point in the buffer zone grid. The exceedance calculation module calculates the probability of the wind field along the transmission line exceeding the preset equipment toughness threshold based on the calibrated wind field value and its uncertainty. The risk level module generates a risk level map based on the probability of exceeding the standard.

10. A local wind field deviation calibration system for transmission lines, characterized in that, The system uses the transmission line local wind field deviation calibration device as described in claim 9.

Images