A method and system for predicting the ice thickness of a power transmission line based on micro-topographic correction

Abstract

The application discloses a kind of transmission line icing thickness prediction method and system based on microtopography correction, belong to power grid disaster prevention and mitigation technical field.The method comprises: obtaining grid numerical weather prediction data and tower coordinates, calculate the initial icing thickness of coarse grid;Based on digital elevation model and water system distribution data extraction microtopographic feature parameters;According to the relief and slope, the prediction area is divided into plain icing area, low mountain and hilly icing area and high altitude icing area;Coarse grid is divided into fine grid, and microtopographic quantitative correction is executed for different icing areas, wherein low mountain and hilly icing area adopts the physical process correction based on atmospheric boundary layer dynamics flow state determination, to obtain fine grid icing thickness;Fine grid icing thickness is matched with tower coordinates, and the output single base tower icing thickness prediction result is output.The present application solves the problem that existing low resolution icing prediction technology cannot depict the icing spatial heterogeneity triggered by microtopography in complex terrain area.

Description

Technical Field

[0001] This invention relates to a method and system for predicting the icing thickness of transmission lines based on micro-topography correction, belonging to the field of power grid disaster prevention and mitigation technology. Background Technology

[0002] Icing on transmission lines is one of the major natural disasters threatening the safe operation of power grids. Existing icing thickness prediction technologies are usually based on gridded numerical weather forecasts with low spatial resolution. They use coarse grids of several kilometers square as the basic unit and employ mechanistic models or statistical regression models to calculate icing thickness. All towers within the same grid are assigned the same icing thickness value.

[0003] However, under complex terrain conditions, the actual variations in altitude, slope and aspect, and the distribution of ridges and valleys within the grid can cause airflow to exhibit drastically different flow patterns at different locations on the mountain, such as over-mountain and obstructed flow. This results in significant spatial differences in local wind speed, liquid water content, and ice thickness. For example, ridges may experience extremely heavy icing due to forced uplift and condensation, while leeward slopes may experience a sharp decrease in icing due to subsidence and warming, and obstructed areas at the foot of the mountain may even be almost ice-free.

[0004] Existing technologies use coarse grids as the smallest prediction unit, which cannot identify and quantify the refined icing distribution characteristics triggered by micro-topography. The prediction results deviate significantly from the actual situation on site, leading to inaccurate identification of high-risk towers and improper allocation of anti-icing resources. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method and system for predicting the icing thickness of transmission lines based on micro-topography correction, which solves the problem that existing low-resolution icing prediction technology cannot characterize the spatial heterogeneity of icing triggered by micro-topography in complex terrain areas.

[0006] The technical problem to be solved by this invention is achieved by the following technical solution:

[0007] A method and system for predicting icing thickness of transmission lines based on micro-topography correction, comprising:

[0008] Obtain gridded numerical weather forecast data and transmission line tower coordinates, and calculate the initial icing thickness of the coarse grid.

[0009] Based on digital elevation model data and water system distribution vector data, micro-topographic feature parameters are extracted;

[0010] Based on the topographic relief and slope, the predicted area is divided into plain icing zone, low mountain and hilly icing zone and high altitude icing zone.

[0011] The coarse grid is divided into multiple fine grids, and micro-topographic quantitative correction is performed for different icing areas. The icing area of ​​low mountains and hills is corrected by physical process based on atmospheric boundary layer dynamics flow regime determination to obtain the icing thickness of the fine grid.

[0012] By matching the fine-grid ice thickness with the tower coordinates, the predicted ice thickness of a single tower is output.

[0013] The present invention is further configured such that the physical process correction of the icy area in the low mountains and hills includes:

[0014] The topographic Froude number is calculated by using wind speed and atmospheric static stability parameters extracted from coarse-grid meteorological data, combined with the relative height of fine-grid data extracted from digital elevation models relative to an upstream reference point.

[0015] Based on the comparison between the Froude number and the preset critical threshold, the flow pattern of the airflow at the fine grid is determined to be either over a hill or blocked.

[0016] Based on the flow regime determination results, the effective wind speed and liquid water content are redistributed, and the fine grid icing thickness is calculated accordingly.

[0017] The present invention is further configured such that the preset critical threshold value ranges from 0.8 to 1.2, or is determined by the following statistical method:

[0018] The measured ice thickness at the towers in historical icing events and the corresponding gridded meteorological and topographic data are obtained. The areas where the measured ice thickness is lower than the preset ice-free threshold are marked as blocked flow, and the rest are marked as over-mountain flow. The Froude number corresponding to each event is calculated, and the upper quantile of the Froude number of the blocked flow sample is taken as the preset critical threshold.

[0019] The present invention is further configured such that when airflow is determined to be blocked, the following correction is triggered:

[0020] The upstream blockage impact distance is determined based on the relative height and Froude number.

[0021] For the fine grid located within the influence distance, an exponential correction is applied to its effective wind speed, which decreases with increasing depth into the blockage zone, so that the effective wind speed at the foot of the mountain is reduced to below a set percentage of the initial value.

[0022] The present invention is further configured such that: in the exponential law correction, the attenuation index is composed of the ratio of the horizontal distance of the fine grid from the upstream blockage starting point to the upstream blockage influence distance multiplied by an attenuation coefficient; the attenuation coefficient ranges from 1 to 3, and the attenuation coefficient is determined based on the Froude number through a preset monotonically decreasing function, and the smaller the Froude number, the larger the attenuation coefficient.

[0023] The present invention is further configured such that when the air mass is determined to be a mountain-crossing airflow and is in a saturated or near-saturated state before lifting, the following correction is triggered:

[0024] Elevation increments are extracted along the windward slope topography, and the cooling capacity of the air mass is calculated based on the wet adiabatic lapse rate.

[0025] The saturated vapor pressure difference is calculated based on the temperature change before and after the air mass cools, and the saturated vapor pressure difference is converted into the additional liquid water content, which is then superimposed on the environmental liquid water content of the corresponding fine grid.

[0026] The invention is further configured such that, after the air mass is determined to be a mountain-crossing airflow and crosses the ridge into the leeward slope, it also includes:

[0027] The temperature rise of the descending air mass along the leeward slope trajectory is calculated using the dry adiabatic lapse rate, and the fine grid temperature is updated accordingly.

[0028] When the updated temperature is below zero degrees and the relative humidity is below the saturation humidity of the ice surface, the ice surface sublimation process is initiated to reduce the ice thickness. The amount of ice mass reduction caused by sublimation is proportional to the difference between the saturation water vapor pressure of the ice surface and the actual water vapor pressure of the environment.

[0029] When the updated temperature is higher than or equal to zero degrees, a melting process is initiated to reduce the ice thickness. The amount of ice mass reduction caused by melting is proportional to the positive accumulated temperature.

[0030] The present invention is further configured such that, after completing the flow regime determination and the corresponding redistribution of wind speed and liquid water content, it also includes:

[0031] Obtain the characteristic diameter of the tower conductor and the median volume diameter of the supercooling water droplets;

[0032] The Stokes number is calculated based on the water droplet density, water droplet diameter, effective wind speed, aerodynamic viscosity, and wire diameter.

[0033] The local collision efficiency coefficient is determined based on the Stokes number. Only when the collision efficiency is greater than zero is the effective liquid water content of the fine mesh converted into the actual icing increment according to the collision efficiency coefficient.

[0034] The present invention is further configured such that the physical process correction of the low mountain and hilly icy area is performed within the following framework:

[0035] The initial icing thickness of the coarse grid is used as the background field icing thickness;

[0036] The initial physical prediction ice thickness of each fine grid is obtained by correcting through the physical process, and the area-weighted average of the initial physical prediction ice thickness of all fine grids in the coarse grid is calculated.

[0037] Using the initial icing thickness of the coarse grid as a baseline, the final icing thickness of each fine grid is determined incrementally. This increment is equal to the difference between the initial physical predicted icing thickness and the area-weighted average.

[0038] A transmission line icing thickness prediction system based on micro-topography correction, characterized in that it includes:

[0039] The data acquisition module is used to acquire gridded numerical weather forecast data and tower coordinates;

[0040] The initial icing calculation module is used to calculate the initial icing thickness of the coarse mesh;

[0041] The micro-topography extraction module is used to extract micro-topography feature parameters based on digital elevation model data and water system distribution vector data.

[0042] The region division module is used to classify icing areas based on terrain relief and slope.

[0043] The calculation module is modified to perform the aforementioned method for different icing areas to obtain the fine-grid icing thickness.

[0044] The matching output module is used to match the fine grid ice thickness with the tower coordinates and output the predicted ice thickness of a single tower.

[0045] The beneficial effects of this invention are:

[0046] 1. By dividing the predicted area into plain icing zone, low mountain and hilly icing zone and high altitude icing zone according to the topographic relief and slope, and performing differentiated micro-topographic quantitative correction for different icing zones, the correction strategy is matched with the dominant physical mechanism of the influence of various topographic features on icing, avoiding the problem of insufficient applicability of a single correction method under all-terrain conditions, and improving the rationality and universality of icing thickness prediction from the architectural level.

[0047] 2. By applying physical process correction based on atmospheric boundary layer dynamics to the icing area in low mountains and hills, the Froude number is used to determine the flow regime of airflow over mountains or blocking at various locations on the mountain. Based on this, the effective wind speed and liquid water content are redistributed based on physical mechanisms. This fundamentally describes the influence of micro-topography on the icing formation process, overcomes the shortcomings of traditional fixed coefficient methods or pure statistical methods that cannot reflect the real atmospheric dynamic-thermal processes, and significantly improves the accuracy of characterizing the spatial distribution differences of icing under complex terrain conditions.

[0048] 3. By incrementally fusing the physical prediction results of the fine grid with the initial ice thickness of the coarse grid, that is, using the initial ice thickness of the coarse grid as the background baseline value and the deviation obtained by the physical process correction as the spatial adjustment term, the relative pattern of ice spatial distribution revealed by the physical process correction is fully preserved, while ensuring that the prediction results are coordinated with the macro meteorological forecast conditions in terms of overall magnitude, thus achieving the organic unity of macro meteorological driving and micro topographic correction. Attached Figure Description

[0049] Figure 1 This is a flowchart of the main process of the method of the present invention.

[0050] Figure 2 This is an architecture diagram of the system of the present invention. Detailed Implementation

[0051] To facilitate a clear understanding of the technical means, creative features, objectives, and effects of this invention, the invention will be further described below in conjunction with specific illustrations.

[0052] Icing on transmission lines is one of the major natural disasters threatening the safe operation of power grids. Existing icing thickness prediction technologies are usually based on low spatial resolution meteorological numerical forecasts, using a 3km×3km grid as the basic unit, and employing mechanistic models or statistical regression models for calculation. All transmission line towers within the same grid are assigned the same icing thickness value. This approach ignores the crucial influence of micro-topographic factors such as elevation changes, slope and aspect differences, ridge and valley distribution, and proximity effects of water systems on icing formation and distribution within the grid.

[0053] Under complex terrain conditions, the foot of a mountain may be almost ice-free due to airflow obstruction, while ridges and saddles may be heavily iced due to forced uplift, and leeward slopes may experience a sharp decrease in icing due to subsidence and warming. Current technologies cannot identify and quantify the abnormal distribution of icing caused by these micro-topographical features, and the predicted results deviate significantly from the actual situation on site, leading to inaccurate identification of high-risk towers and improper allocation of anti-icing resources.

[0054] like Figure 1 As shown, this invention provides a method for predicting the icing thickness of transmission lines based on micro-topography correction, comprising the following steps:

[0055] First, obtain gridded numerical weather forecast data and the coordinates of transmission line towers, and then calculate the initial icing thickness of the coarse grid.

[0056] Numerical weather forecast data can be sourced from the output products of global forecasting systems or regional scale numerical weather prediction models, with a typical spatial resolution of 3km × 3km. Extracted meteorological elements include near-surface temperature, relative humidity, wind speed, precipitation, and temperature profiles of multiple isobaric surfaces. Tower coordinates include the latitude, longitude, and altitude of each tower. The initial icing thickness of the coarse grid can be calculated using a CRREL mechanism model, a BP neural network model, or a statistical regression model trained on historical data. Taking the BP neural network model as an example, the input layer includes four nodes: temperature, humidity, wind speed, and precipitation; the hidden layer has several nodes; and the output layer is the icing thickness. The model training samples are taken from measured meteorological and icing thickness data of winter icing events over the years.

[0057] After obtaining the initial ice thickness of the coarse grid, micro-topographic feature parameters, including altitude, topographic relief, slope, aspect, ridgeline, valleyline, saddle, and distance to water system, are extracted based on digital elevation model data and water system distribution vector data.

[0058] Digital elevation model (DEM) data, with a preferred spatial resolution of 30m, can be obtained from publicly available data sources such as the Space Shuttle Radar Topographic Mapping Mission or the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) Global Digital Elevation Model. Water system distribution vector data, including linear and areal features of rivers and lakes, can be extracted from national basic geographic information data. Topographic relief is calculated by taking a window of a certain radius centered on the target grid and calculating the difference between the maximum and minimum elevations within that window.

[0059] Slope calculation employs a fitted surface method, performing least-squares surface fitting on local windows of the DEM and using the surface gradient as the slope value. Slope aspect is measured clockwise with true north as 0 degrees. Ridge and valley lines are extracted using a DEM-based hydrological analysis method. Watersheds and dividing lines are determined through depression filling, flow direction calculation, and runoff accumulation calculation; the dividing line is the ridge line, and the watershed line is the valley line. Saddle identification is determined by identifying local DEM values ​​that are minimum in a certain direction and maximum in the vertical direction. The river system distance is the horizontal distance from the center of the fine grid to the nearest river system element.

[0060] Next, based on topographic relief and slope thresholds, the predicted area is divided into plain icing zones, low mountain and hilly icing zones, and high-altitude icing zones. A typical setting for the topographic relief threshold is as follows: areas with relief less than 50m and slope less than 5 degrees are classified as plain icing zones; areas with relief greater than 200m are classified as high-altitude icing zones; and the remaining areas are classified as low mountain and hilly icing zones. The topographic relief threshold can be appropriately adjusted according to the overall topographic characteristics of the actual predicted area, or the optimal threshold can be determined by statistically analyzing the icing distribution characteristics under different topographic conditions in historical icing events.

[0061] After completing the regional division, the coarse grid is divided into multiple fine grids for different icing areas and micro-topographic quantitative correction is performed. The resolution of the fine grid is consistent with the resolution of the DEM data, usually 30m×30m. Each 3km×3km coarse grid is divided into 100 rows×100 columns, totaling 10,000 fine grids.

[0062] The following sections explain the correction methods for the three types of icing zones.

[0063] In plain icing areas, a water system influence correction is applied. This correction reduces the ice thickness based on the distance between the fine grid and the water system, using a distance attenuation function. Because the terrain of plains is flat, the spatial variation in ice thickness mainly stems from changes in water vapor conditions near water systems. The closer to a river or lake, the more abundant the local water vapor supply, and the easier it is for ice to form and grow. Therefore, the water system correction reflects this: the closer to the water system, the thicker the ice, gradually attenuating to the background value with increasing distance. The distance attenuation function can take the form of linear attenuation, exponential attenuation, or piecewise functions.

[0064] For example, when the distance between the fine grid and the water system is less than a certain threshold, the ice thickness is multiplied by an enhancement factor greater than 1; as the distance exceeds the threshold, the enhancement factor gradually decreases to 1. It is important to note that the above enhancement correction is only activated when the fine grid is located on the windward side of the line where the prevailing wind direction is blowing from the water system. When the fine grid is located on the leeward side of the water system, the water body's replenishment effect on local water vapor is not significant, and in this case, the correction is not activated or a weakening correction is activated. Specific parameters can be calibrated by comparing ice accumulation observation data from plain areas.

[0065] In high-altitude icing areas, an altitude logarithmic relationship correction was adopted. This correction utilizes the logarithmic function relationship between altitude and icing thickness to incrementally correct the icing thickness of fine grids. Icing observation data show that in higher altitude areas, as altitude increases, temperature decreases, cloud and fog frequency increases, and icing thickness shows an altitude-related increasing trend. The logarithmic relationship can fit this change characteristic well.

[0066] The general form of the correction formula is: the corrected thickness equals the original thickness multiplied by a correction factor proportional to the logarithm of altitude, i.e., R_corrected = R_original × [1 + β·ln(H / H0)], where H is the fine grid altitude, H0 is the reference altitude, and β is the empirical fitting coefficient. The specific coefficient of the correction factor is determined by least-squares fitting based on the climate background and historical icing observation data of the prediction area. The logarithmic altitude correction is applicable to high-altitude areas with undulations exceeding 200m. These areas typically correspond to the main body of mountain ranges, where the wind speed and water vapor differences caused by the terrain itself are largely reflected in the large-scale meteorological field, and the remaining influences are mainly dominated by altitude.

[0067] The low mountain and hilly icing area is the region with the most complex topographic influence and the most significant spatial differences in icing, and it is also the key target of this invention for optimization and correction. This invention adopts physical process correction based on atmospheric boundary layer dynamics flow regime determination. Its core lies in using the dimensionless parameter of topographic Froude number to determine whether the airflow will cross the mountain or be blocked and flow around it when it encounters a mountain. Based on different flow regimes, corresponding wind speed corrections and liquid water content redistribution are triggered, thereby characterizing the influence of micro-topography on icing from a physical essence.

[0068] When performing corrections for icing areas in low mountains and hills, near-surface wind speed and atmospheric static stability parameters are first extracted from coarse-grid meteorological data. These parameters characterize the degree to which the air layer inhibits or promotes vertical motion and are derived from the potential temperature gradient in the vertical direction. Specifically, the calculation method involves extracting the temperature and potential temperature values ​​of adjacent layers from the multi-layer isobaric surface temperature data in numerical weather prediction products. Potential temperature is the temperature a gas parcel would have if it were moved to 1000 hPa standard pressure via a dry adiabatic process.

[0069] The square of the Brent-Vesara frequency N is equal to the gravitational acceleration divided by the average potential temperature and then multiplied by the potential temperature gradient, i.e. The larger the N value, the more stable the air layer, and the more likely the airflow is to circumvent the mountain rather than cross it; the smaller the N value, the less stable the air layer, and the easier it is for the airflow to cross the ridge.

[0070] Simultaneously, the relative height of the fine grid relative to the upstream reference point is extracted from the digital elevation model. The upstream reference point is defined as the lowest elevation point on the windward side boundary of the coarse grid. This location represents the initial elevation at which the air mass begins to encounter topographic obstruction. Calculating the relative height based on this reference point accurately reflects the obstructive effect of topography on airflow. For any fine grid, let its elevation be Z, and the elevation of the upstream reference point be Zref. Then, the relative height h = Z - Zref. When h is negative, it is taken as zero, indicating that the fine grid is lower than the reference point and can be considered to have no topographic influence.

[0071] Based on this, the topographic Froude number Fr is calculated, defined as Fr = U / (N·h), where U is the wind speed. The physical meaning of the Froude number is the ratio of air kinetic energy to the potential energy required to overcome the resistance of stable stratification and climb over ridges. When the Froude number is large, the air kinetic energy is sufficient to overcome the inhibitory effect of stable stratification and cross the ridge; when the Froude number is small, the air kinetic energy is insufficient to climb, and the airflow is forced to bypass the mountain or stagnate in front of it.

[0072] The calculated Froude number is compared with the preset critical threshold Fr_c. The critical threshold ranges from 0.8 to 1.2, which corresponds to the critical transition range where the Froude number is about 1 in atmospheric dynamics theory. The theoretical value is 1, but in the actual atmosphere, there are factors such as turbulent mixing and boundary layer inhomogeneity. The transition between blocking and over mountains is gradually completed within a transition range. Therefore, setting 0.8 to 1.2 as the transition range can more closely reflect the actual atmospheric behavior.

[0073] The critical threshold can also be determined based on the statistical values ​​of Froude numbers when blocking and mountain crossing actually occur in historical icing events. The specific statistical method is as follows: obtain the measured ice thickness at the tower and its corresponding grid meteorological and topographic data in historical icing events, mark the area where the tower with the measured ice thickness is lower than the preset ice-free threshold as the blocking flow state, and mark the rest as the mountain crossing flow state, calculate the Froude number corresponding to each event, and take the upper quantile of the Froude number of the blocking flow state sample as the preset critical threshold applicable to the local area.

[0074] If the calculated Fr is less than the critical threshold, the airflow at that fine grid is determined to be a mountain-circling blockage airflow; otherwise, it is determined to be a mountain-crossing airflow. Based on the flow pattern determination result, the following different correction mechanisms are triggered respectively.

[0075] When obstructed airflow is detected, obstructed airflow correction is triggered. First, the upstream obstruction influence distance L_b is determined based on the relative height and Froude number. The obstruction influence distance refers to the range in which the wind speed begins to decrease significantly after the airflow is blocked in front of the mountain. This distance is directly proportional to the relative height and inversely proportional to the Froude number, i.e., L_b=γ·h / Fr, where γ is an empirical coefficient, usually ranging from 0.5 to 2.0, preferably 1.0.

[0076] For the fine grid located within the influence distance, a correction is applied to its effective wind speed, which gradually increases in attenuation as the depth of entry into the blockage zone increases. The specific correction of the effective wind speed follows an exponential law: U_eff=U·exp(-k·x / L_b), where x is the horizontal distance of the fine grid from the upstream blockage starting point, and k is the attenuation coefficient, which ranges from 1 to 3.

[0077] To make the correction more physically grounded, the attenuation coefficient k is determined based on the Froude number using a preset monotonically decreasing function. The smaller the Froude number, the larger the attenuation coefficient; for example, k = 3.0 × (1 - Fr) + 1.0 can be used. Through this correction, the effective wind speed in the foothills' stagnation area can be attenuated to less than 5% of the initial value, reasonably quantifying the low wind speed conditions in the blocking area at the foot of the mountain. Due to the extremely low wind speed, supercooled water droplets are difficult to effectively reach the conductor surface, resulting in very little or no icing in this area.

[0078] When it is determined to be a cross-mountain airflow and the air mass is in a saturated or near-saturated state (i.e., relative humidity reaches 95% or more) before it rises, the correction of the liquid water content on the windward slope is triggered, and the elevation increment is extracted along the windward slope topography. Let the elevation increment from the upstream reference point to the current fine grid be Δz.

[0079] The cooling amount of an air mass is calculated using the wet adiabatic lapse rate, Γ_m, which is the rate of temperature change with altitude when saturated moist air rises and expands and cools. A typical value is 6 to 7 degrees Celsius per kilometer, with the specific value determined by referring to a table based on the initial temperature of the air mass. The cooling amount of the air mass is ΔT = Γ_m·Δz. The saturated vapor pressure difference is calculated based on the temperature change before and after the air mass cools. Before the air mass rises, it has an initial temperature T. Because it is in a near-saturated state, its initial vapor pressure e is approximately equal to the saturated vapor pressure e_s(T) at temperature T. After rising and cooling, the temperature becomes T - ΔT, and the corresponding saturated vapor pressure is e_s(T - ΔT). The supersaturated vapor pressure difference Δe = e_s(T) - e_s(T - ΔT). The vapor pressure difference is converted into the increased liquid water content using the ideal gas law: ΔLWC = (Δe·M_w) / (R·T_m), where M_w is the molar mass of water, taken as 18.015 g / mol, R is the universal gas constant, taken as 8.314 joules / mol Kelvin, and T_m is the average temperature of the lifting process. This increased liquid water content is then superimposed on the corresponding fine-grid environmental liquid water content as input for subsequent icing calculations.

[0080] This correction quantitatively incorporates the increased liquid water supply on the windward slope due to forced uplift into the icing calculation, explaining the significantly heavier icing on the ridge and upper windward slope compared to other areas.

[0081] After the air mass crosses the ridge and enters the leeward slope, leeward slope attenuation correction is required. The air mass's descent temperature is calculated along its leeward slope trajectory using the dry adiabatic lapse rate, and the fine grid temperature is updated. The dry adiabatic lapse rate Γ_d is the rate of temperature change with altitude during the compression and warming of unsaturated dry air during descent, and is taken as 9.8 degrees Celsius per kilometer. The air mass temperature T_ridge at the ridge is determined by the temperature after cooling on the windward slope. The temperature of each fine grid on the leeward slope is T_lee = T_ridge + Γ_d · Δz_down, where Δz_down is the elevation difference from the ridge to that fine grid.

[0082] When the updated temperature is below zero degrees and the relative humidity is below the ice surface saturation humidity, it indicates that the ambient water vapor pressure is below the ice surface saturation water vapor pressure, satisfying the thermodynamic conditions for sublimation. At this time, the ice surface sublimation rate function is activated to reduce the ice thickness.

[0083] The ice surface sublimation rate function is: ΔM_sub=K_sub×(e_i(T)-e_a)×Δt, where K_sub is the sublimation mass transfer coefficient, the value of which can be obtained through laboratory measurements or literature review, with a typical value in the range of 1×10. -9 Up to 1×10 - 8 kg·m -2 ·s -1 ·Pa -1 Order of magnitude (corresponding to a reduction of approximately 10 from the original dimensional parameters) 5 The equivalent value after multiplication), e_i(T) is the saturated water vapor pressure on the ice surface corresponding to temperature T, which can be calculated using the Magnus ice surface formula e_i(T)=6.112×exp[22.46×T / (272.62+T)] (T is in °C), e_a is the actual water vapor pressure in the environment, and Δt is the time step.

[0084] When the updated temperature is higher than or equal to zero degrees Celsius, a melting rate function is activated to reduce the icing thickness. The melting rate function is: ΔM_melt = k_m × max(T - T_0, 0) × Δt, where k_m is the melting coefficient, typically ranging from 0.01 to 0.1 mm·h. -1 ·℃ -1 T_0 = 0℃, Δt is the time step, and the relationship between mass reduction and icing thickness reduction is ΔR = ΔM / (ρ_i × A), where ρ_i is the icing density and A is the surface area per unit length of conductor. Through the above sub-mechanism-based leeward slope reduction correction, the rapid melting or sublimation process of icing on the leeward slope due to the foehn effect is reasonably described.

[0085] After determining the flow regime and the corresponding redistribution of wind speed and liquid water content, the influence of conductor micro-collision efficiency on icing formation is further considered. The characteristic diameter of the tower conductor and the median volume diameter of the supercooling water droplets are obtained. The characteristic diameter of the conductor is the outer diameter of the transmission line conductor, an engineering design parameter, with typical values ​​between 15mm and 40mm.

[0086] The median volume diameter of supercooled water droplets refers to the diameter value corresponding to 50% of the cumulative volume of water droplet diameter distribution in clouds and fog. It characterizes the average scale of cloud and fog droplet spectrum. This parameter can be extracted from the microphysical quantity output of numerical weather prediction models. If the operational numerical weather prediction system does not provide microphysical quantity output, the typical median volume diameter value of icing clouds and fog is taken. This typical value is between 15 μm and 30 μm, preferably 20 μm.

[0087] Calculate the Stokes number Stk. The Stokes number is a dimensionless number describing the ratio of inertial force to viscous drag when supercooled water droplets move with the airflow. It is defined as Stk = (ρ_w·d_p) / (ρ_w·d_p) 2·U_eff) / (18·μ·D), where ρ_w is the water droplet density, d_p is the water droplet diameter, U_eff is the local effective wind speed, μ is the aerodynamic viscosity, and D is the conductor diameter. The coefficient 18 in the denominator is a component of the Stokes number standard definition.

[0088] The local collision efficiency coefficient is determined based on the Stokes number. Only when the collision efficiency is greater than zero is the effective liquid water content of the fine grid converted into the actual icing increment according to the collision efficiency coefficient. When the Stokes number is small, the water droplets have small inertia and tend to flow around the conductor with the airflow without colliding with it, and the collision efficiency is zero. When the Stokes number is large, the water droplets have large inertia and tend to deviate from the streamline and collide with the surface of the conductor.

[0089] The conversion relationship is: the increase in icing thickness within the time step ΔR = (β·U_eff·LWC_eff·Δt) / (ρ_i·π), where β is the collision efficiency coefficient and ρ_i is the icing density, usually taken as 900 kg / m³. 3 π is the mathematical constant pi. The collision efficiency coefficient β can be determined based on the Stokes number using the following piecewise function: when Stk < 0.125, β = 0; when 0.125 ≤ Stk < 1.1, β = 0.466·(lnStk). 2 +0.883·lnStk+0.848; when Stk≥1.1, β=[Stk / (Stk+π / 2)] 2 .

[0090] The form of the above collision efficiency function is derived from the classical collision theory of Langmuir and Blodgett and the engineering expression of the Makkonen icing model.

[0091] To ensure that the fine-grid ice thickness obtained by the above physical process correction can reflect the spatial distribution differences caused by micro-topography and maintain consistency with the overall trend of coarse-grid weather forecasts, this invention adopts an incremental correction framework to determine the final thickness of ice in low mountain and hilly areas.

[0092] Specifically, the initial icing thickness R_init of the coarse grid is used as the background icing thickness, representing the average icing state at that coarse grid scale. Through the aforementioned physical process correction, the initial physical prediction icing thickness R_phys is calculated for each fine grid. This value already includes all physical effects of flow regime determination, liquid water content redistribution, collision efficiency conversion, and leeward slope attenuation. The area-weighted average of R_phys for all fine grids within the coarse grid is then calculated as R_avg.

[0093] Subsequently, the final icing thickness of each fine grid is calculated incrementally: R_final = R_init + (R_phys - R_avg). In this framework, R_phys - R_avg represents the deviation of the fine grid from the average state of the coarse grid revealed by the physical process correction, while R_init provides a reliable background baseline value. Thus, the relative pattern of icing spatial distribution characterized by the physical process correction is fully preserved, while the background value is determined by the statistical or mechanistic model, avoiding unreasonable amplification or reduction caused by forced scaling due to a mismatch in magnitude between the physical and statistical models. When the difference between R_avg and R_init is too large (e.g., exceeding ±50%), parameter calibration or model verification processes can be further triggered, or a weighted fusion mechanism can be introduced for a smooth transition to enhance the robustness of the system.

[0094] Finally, the calculated ice thickness of the fine grid after all corrections is matched with the tower coordinates to output the predicted ice thickness for a single tower. The matching method involves locating the fine grid containing the tower based on its latitude and longitude coordinates, and assigning the ice thickness value of that fine grid to the tower. If a tower spans multiple fine grids, the average or maximum value of the fine grids within the coverage area is used. The output results may include the ice thickness value, icing level, and corresponding warning level for each tower.

[0095] like Figure 2 As shown, the present invention also provides a transmission line icing thickness prediction system based on micro-topography correction. This system includes a data acquisition module, an initial icing calculation module, a micro-topography extraction module, a region division module, a correction calculation module, and a matching output module, wherein:

[0096] The data acquisition module is used to acquire gridded numerical weather forecast data and tower coordinates. This module obtains gridded numerical forecast products from the weather forecast data server or local storage through a network interface or local file reading method, extracts the meteorological element field of the forecast area within the specified forecast period, and obtains transmission line tower coordinate data from the power grid production management system or geographic information system. It also has data format conversion and coordinate system integration functions.

[0097] The initial icing calculation module is used to calculate the initial icing thickness of the coarse mesh. It embeds the CRREL mechanism model, BP neural network model, or other icing thickness calculation models and supports model parameter configuration and model selection functions.

[0098] The micro-topography extraction module is used to extract micro-topographic feature parameters based on digital elevation model data and water system distribution vector data. It integrates GIS spatial analysis functions and outputs raster data of micro-topographic feature parameters, including altitude, topographic relief, slope, aspect, ridge line, valley line, saddle location, and distance to water system.

[0099] The region division module is used to classify icing areas according to terrain relief and slope, dividing each grid cell in the prediction area into plain icing areas, low mountain and hilly icing areas, or high altitude icing areas. The threshold parameters can be configured by the user.

[0100] The correction calculation module is the core calculation unit of the system. It integrates all the algorithms for water system correction in plain areas, logarithmic elevation correction in high-altitude areas, and refined correction based on flow regime determination in low-mountain and hilly areas using Froude number. Based on the classification results provided by the region division module, it performs fine-grid subdivision on each coarse grid. For each fine grid, it calls the corresponding correction algorithm subroutine according to the type of icing area it belongs to. After completing all correction calculations, it outputs the final icing thickness raster data for each fine grid.

[0101] The matching output module is used to match the fine grid ice thickness with the tower coordinates and output the predicted ice thickness of a single tower. It supports multiple formats such as text tables, geographic information layers, and statistical charts.

[0102] The implementation process and effects of the present invention are illustrated below through a specific embodiment. The embodiment selects a typical icing section of a 220kV transmission line in Xuancheng, Anhui Province. This section belongs to a low mountain and hilly icing area, where severe icing of the ridges and saddles is likely to occur when cold air moves southward in winter, while the foot of the mountain is almost ice-free.

[0103] First, meteorological forecast data for the 3km × 3km coarse grid in this section were obtained: near-surface temperature -2.6℃, wind speed at 10m height 4.8m / s, relative humidity 94%, and precipitation 1.2mm / h. Simultaneously, isobaric surface temperature data at 850hPa and 925hPa were acquired, and the potential temperature difference and height difference were calculated, yielding a static stability parameter N of approximately 0.018 per square second. The initial icing thickness R_init for this coarse grid was calculated to be 12.4mm using a BP neural network model. Using 30m resolution DEM data, topographic parameters were extracted centered on this coarse grid, calculating a topographic relief of 310m, an average slope of 16.3 degrees, and an altitude range of 580m to 770m. The topographic relief of 310m exceeds the 200m threshold, the average slope of 16.3 degrees, and the maximum slope exceeding 25 degrees, consistent with the topographic characteristics of a low-mountain and hilly icy area. Therefore, this coarse grid was classified as a low-mountain and hilly icy area.

[0104] Next, a physical process correction based on the Froude number flow regime determination is performed. The lowest elevation point on the windward side boundary of the coarse grid is selected as the upstream reference point, with an elevation of 580m. For a fine grid at the ridge, with an elevation of 770m and a relative height h of 190m, the wind speed U is set to 4.8m / s, and the static stability parameter N is set to 0.018 seconds squared. The Froude number Fr is calculated to be approximately 1.40, which is greater than the critical threshold of 1.0. This indicates an over-mountain airflow, triggering a correction for windward slope adiabatic uplift.

[0105] The elevation increment Δz from the reference point to the ridge along the windward slope is 190m, the wetted adiabatic lapse rate is 6.5℃ / km, and the air mass cooling amount ΔT is 6.5×0.19=1.235℃. Before uplift, the air mass temperature was -2.6℃, close to saturation; after uplift and cooling, the temperature is approximately -3.835℃.

[0106] Using the Magnus formula, the saturated vapor pressure at -2.6℃ is approximately 5.09 hPa, and the saturated vapor pressure at -3.835℃ is approximately 4.54 hPa. The supersaturated vapor pressure difference Δe = 0.55 hPa = 55 Pa.

[0107] By transforming the ideal gas law, the newly added liquid water content ΔLWC = (Δe × M_w) / (R × T_m) = (55 × 0.018015) / (8.314 × 269.5) ≈ 4.42 × 10 -4 kg / m 3 That is, approximately 0.44 g / m 3 .

[0108] After incorporating the ambient liquid water content, the effective liquid water content (LWC_eff) of the fine-grid ridge is approximately 0.65 g / m². 3 (Assuming the background environmental value is approximately 0.21 g / m³) 3 ).

[0109] Next, the ice thickness increment of the fine mesh within a time step Δt = 1 h is calculated. The wire diameter D is taken as 30 mm, the median volume diameter d_p of the supercooled water droplet is taken as 20 μm, and the aerodynamic viscosity μ is taken as 1.72 × 10⁻⁶. -5 Pa·s. The effective wind speed U_eff is maintained at 4.8 m / s.

[0110] Calculate the Stokes number Stk = (ρ_w × d_p) 2 ×U_eff) / (18×μ×D)=(1000×(20×10 -6 ) 2 ×4.8) / (18×1.72×10 -5 ×30×10 -3 ). numerator = 1000 × 4 × 10-10 ×4.8=1.92×10 -6 The denominator is 18 × 1.72 × 10⁻⁶. -5 ×0.03=9.29×10 -6 Stk≈0.207.

[0111] Based on the piecewise function of collision efficiency, lnStk≈-1.575, β=0.466×(lnStk) 2 +0.883×lnStk+0.848≈0.614. The density of the icing layer ρ_i is taken as 900 kg / m³. 3 .

[0112] The ice thickness increment within this time step is ΔR = (β × U_eff × LWC_eff × Δt) / (ρ_i × π) = (0.614 × 4.8 × 0.65 × 10^6 ppm). -3 (×3600) / (900×3.1416)≈2.44×10 -3 m = 2.44 mm. Assuming the icing process lasts for 6 hours and the weather conditions remain stable, the initial physical prediction of the icing thickness of the fine grid on the ridge is R_phys = 2.44 × 6 ≈ 14.6 mm.

[0113] For a fine grid on the leeward slope, the elevation difference behind the ridge is about 200m, and the dry adiabatic warming is about 1.96℃. The air mass temperature at the ridge is -3.8℃, and after the warming on the leeward slope, the temperature T_lee is about -1.84℃. At this time, the relative humidity of the air mass drops to about 65% due to the compression of the descending air mass, and the ambient water vapor pressure e_a is lower than the ice surface saturated water vapor pressure e_i (-1.84℃) (about 5.30hPa=530Pa).

[0114] To initiate the sublimation process, set K_sub = 5 × 10 -10 kg·m -2 ·s -1 ·Pa -1 The time step Δt is set to 6h = 21600s.

[0115] ΔM_sub=K_sub×(e_i-e_a)×Δt=5×10 -10 ×(530-344.5)×21600≈2.0×10 - 3 kg / m 2 .

[0116] The reduction in icing thickness caused by sublimation is ΔR_sub = ΔM_sub / ρ_i ≈ 2.2 × 10⁻⁶ -6With m≈0.002mm, the leeward slope itself has very limited icing growth (due to reduced wind speed and poor water vapor conditions), and its initial physical prediction of icing thickness R_phys is about 0.05mm.

[0117] The R_phys values ​​for all 10,000 fine grids within the coarse grid were calculated using the same method, yielding an area-weighted average value R_avg of approximately 5.2 mm. Using an incremental correction framework, the initial ice thickness of the coarse grid was R_init = 12.4 mm. The final ice thickness of the fine grid at the ridge was R_final = 12.4 + (14.6 - 5.2) = 21.8 mm.

[0118] The final icing thickness of the fine grid on the leeward slope is R_final = 12.4 + (0.05 - 5.2) = 7.25 mm. In the blocking area at the foot of the mountain, due to extremely low effective wind speed, the R_phys value of the fine grid is almost zero, resulting in a final icing thickness of R_final = 12.4 + (0 - 5.2) = 7.2 mm.

[0119] Comparing the above predictions with measured data, the measured ice cover in this section was approximately 22.5 mm on the ridge, approximately 8.0 mm on the leeward slope, and approximately 6.8 mm at the foot of the mountain. The relative error of the method of this invention is within 10%, and it accurately captures the spatial distribution pattern of heavy ice on the ridge, and relatively light ice cover on the leeward slope and at the foot of the mountain.

[0120] In contrast, if micro-topographical correction is not used and the entire coarse grid is uniformly assigned a value of 12.4 mm, it will be completely unable to reflect the difference between the ridge and the foot of the mountain. If the traditional fixed conversion factor method is used for total back calculation, the predicted value of the ridge is often over-inflated to more than 30 mm, and the leeward slope is reduced to less than 5 mm, which deviates significantly from the actual distribution.

[0121] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments, and various changes and modifications can be made without departing from the spirit and scope of the invention, all of which fall within the scope of protection claimed by the present invention. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for predicting icing thickness of transmission lines based on micro-topography correction, characterized in that, include: Obtain gridded numerical weather forecast data and transmission line tower coordinates, and calculate the initial icing thickness of the coarse grid. Based on digital elevation model data and water system distribution vector data, micro-topographic feature parameters are extracted; Based on the topographic relief and slope, the predicted area is divided into plain icing zone, low mountain and hilly icing zone and high altitude icing zone. The coarse grid is divided into multiple fine grids, and micro-topographic quantitative correction is performed for different icing areas. The icing area of ​​low mountains and hills is corrected by physical process based on atmospheric boundary layer dynamics flow regime determination to obtain the icing thickness of the fine grid. By matching the fine-grid ice thickness with the tower coordinates, the predicted ice thickness of a single tower is output.

2. The method according to claim 1, characterized in that, The physical process corrections for the icy-covered low-mountain and hilly areas include: The topographic Froude number is calculated by using wind speed and atmospheric static stability parameters extracted from coarse-grid meteorological data, combined with the relative height of fine-grid data extracted from digital elevation models relative to an upstream reference point. Based on the comparison between the Froude number and the preset critical threshold, the flow pattern of the airflow at the fine grid is determined to be either over a hill or blocked. Based on the flow regime determination results, the effective wind speed and liquid water content are redistributed, and the fine grid icing thickness is calculated accordingly.

3. The method according to claim 2, characterized in that, The preset critical threshold value ranges from 0.8 to 1.2, or can be determined using the following statistical methods: The measured ice thickness at the towers in historical icing events and the corresponding gridded meteorological and topographic data are obtained. The areas where the measured ice thickness is lower than the preset ice-free threshold are marked as blocked flow, and the rest are marked as over-mountain flow. The Froude number corresponding to each event is calculated, and the upper quantile of the Froude number of the blocked flow sample is taken as the preset critical threshold.

4. The method according to claim 2, characterized in that, When airflow is determined to be blocked, the following correction is triggered: The upstream blockage impact distance is determined based on the relative height and Froude number. For the fine grid located within the influence distance, an exponential correction is applied to its effective wind speed, which decreases with increasing depth into the blockage zone, so that the effective wind speed at the foot of the mountain is reduced to below a set percentage of the initial value.

5. The method according to claim 4, characterized in that, In the exponential correction, the attenuation exponent is composed of the ratio of the horizontal distance of the fine grid from the upstream blockage starting point to the distance of the upstream blockage influence multiplied by an attenuation coefficient; the attenuation coefficient ranges from 1 to 3, and the attenuation coefficient is determined based on the Froude number through a preset monotonically decreasing function, and the smaller the Froude number, the larger the attenuation coefficient.

6. The method according to claim 2, characterized in that, When the air mass is determined to be a mountain-crossing airflow and is in a saturated or near-saturated state before rising, the following correction is triggered: Elevation increments are extracted along the windward slope topography, and the cooling capacity of the air mass is calculated based on the wet adiabatic lapse rate. The saturated vapor pressure difference is calculated based on the temperature change before and after the air mass cools, and the saturated vapor pressure difference is converted into the additional liquid water content, which is then superimposed on the environmental liquid water content of the corresponding fine grid.

7. The method according to claim 6, characterized in that, Once it is determined to be a mountain-crossing airflow and the air mass crosses the ridge and enters the leeward slope, it also includes: The temperature rise of the descending air mass along the leeward slope trajectory is calculated using the dry adiabatic lapse rate, and the fine grid temperature is updated accordingly. When the updated temperature is below zero degrees and the relative humidity is below the saturation humidity of the ice surface, the ice surface sublimation process is initiated to reduce the ice thickness. The amount of ice mass reduction caused by sublimation is proportional to the difference between the saturation water vapor pressure of the ice surface and the actual water vapor pressure of the environment. When the updated temperature is higher than or equal to zero degrees, a melting process is initiated to reduce the ice thickness. The amount of ice mass reduction caused by melting is proportional to the positive accumulated temperature.

8. The method according to claim 2, characterized in that, After completing the flow regime determination and the corresponding redistribution of wind speed and liquid water content, the following steps are also included: Obtain the characteristic diameter of the tower conductor and the median volume diameter of the supercooling water droplets; The Stokes number is calculated based on the water droplet density, water droplet diameter, effective wind speed, aerodynamic viscosity, and wire diameter. The local collision efficiency coefficient is determined based on the Stokes number. Only when the collision efficiency is greater than zero is the effective liquid water content of the fine mesh converted into the actual icing increment according to the collision efficiency coefficient.

9. The method according to claim 1, characterized in that, The physical process correction for the icy areas in the low mountains and hills is performed within the following framework: The initial icing thickness of the coarse grid is used as the background field icing thickness; The initial physical prediction ice thickness of each fine grid is obtained by correcting through the physical process, and the area-weighted average of the initial physical prediction ice thickness of all fine grids in the coarse grid is calculated. Using the initial icing thickness of the coarse grid as a baseline, the final icing thickness of each fine grid is determined incrementally. This increment is equal to the difference between the initial physical predicted icing thickness and the area-weighted average.

10. A transmission line icing thickness prediction system based on micro-topography correction, characterized in that, include: The data acquisition module is used to acquire gridded numerical weather forecast data and tower coordinates; The initial icing calculation module is used to calculate the initial icing thickness of the coarse mesh; The micro-topography extraction module is used to extract micro-topography feature parameters based on digital elevation model data and water system distribution vector data. The region division module is used to classify icing areas based on terrain relief and slope. A correction calculation module is used to perform the method as described in any one of claims 1 to 9 for different icing areas to obtain the fine-grid icing thickness; The matching output module is used to match the fine grid ice thickness with the tower coordinates and output the predicted ice thickness of a single tower.

Images