Two-micro-segment intelligent identification and risk grading method and system for power transmission line
By acquiring topographic elevation data of transmission lines, extracting micro-topographic and meteorological parameters, and constructing a dynamic risk classification model, the problem of micro-topographic changes being ignored in traditional models is solved, and the accurate classification and improvement of risk levels are achieved.
Patent Information
- Application Number
- CN202511613873.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-06
- Publication Date
- 2026-01-30
AI Technical Summary
Traditional risk classification models use fixed or empirical thresholds, ignoring micro-topographical changes, which leads to a mismatch between risk level classification and actual disaster distribution, resulting in inaccurate output results.
By acquiring the topographic elevation data of transmission lines, extracting micro-topographic parameters such as curvature, slope, and topographic relief, and combining them with micro-meteorological parameters such as wind acceleration coefficient and icing coefficient, a comprehensive risk index is calculated, and a risk classification model with dynamic threshold division is constructed to accurately predict the risk level.
It achieves precise risk level classification, which is more in line with the actual operational risks of transmission lines and improves the accuracy of risk classification.
Smart Images
Figure CN121436680A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of transmission line risk analysis technology, and in particular to a method and system for intelligent identification and risk classification of two micro-segments for transmission lines. Background Technology
[0002] With the expansion of power transmission scale, more and more power grid lines pass through complex geographical environments and harsh climatic conditions. The meteorological factors caused by micro-topography change drastically within a small area, which significantly enhances the meteorological conditions in the area. This often leads to various disasters such as conductor breakage, transmission tower collapse and galloping, threatening the normal operation of power transmission lines.
[0003] Traditional risk classification models often use fixed or empirical thresholds, which are mostly subjectively set and often ignore the risks brought about by micro-topographical changes. This results in the risk level classification not matching the actual disaster distribution and inaccurate risk classification results. Summary of the Invention
[0004] The main purpose of this application is to provide a method and system for intelligent identification and risk classification of two micro-segments of transmission lines. It aims to solve the technical problem that the use of fixed thresholds and empirical thresholds in related technologies often ignores the risks brought about by micro-topographic changes, resulting in the risk level classification not matching the actual disaster distribution and the output risk classification results being inaccurate.
[0005] To achieve the above objectives, embodiments of this application provide a method for intelligent identification and risk classification of two micro-segments in transmission lines, including: The terrain elevation data of the area where the transmission line is located is obtained, and various micro-topographic parameters are extracted from the terrain elevation data, including curvature, slope and terrain undulation. Based on the micro-topographic parameters, the terrain type of the terrain area where the transmission line is located and the corresponding micro-meteorological parameters are determined. The dynamic division threshold is obtained by dynamically adjusting the terrain parameters corresponding to different terrain areas. The micro-meteorological parameters include wind acceleration coefficient and icing coefficient. Based on the micro-topographical parameters and the micro-meteorological parameters, a comprehensive risk index is calculated; Based on the aforementioned comprehensive risk index, a risk classification model is constructed; The risk level of each terrain region is predicted using the risk grading model and dynamic classification threshold to obtain the risk grading result.
[0006] In one possible implementation of this application, determining the terrain type of the terrain area where the transmission line is located and the corresponding micro-meteorological parameters based on the micro-topographic parameters includes: Based on the standard deviation of each micro-topographic parameter, a topographic feature pyramid corresponding to the micro-topographic parameter is constructed. The terrain feature pyramid is input into a preset deep learning model. Based on the preset deep learning model, the terrain feature pyramid is classified to obtain the terrain type of the terrain area where the transmission line is located. Based on preset standards and the terrain type, micro-meteorological parameters are determined.
[0007] In one possible implementation of this application, the step of classifying the terrain feature pyramid based on a preset deep learning model to obtain the terrain type of the terrain area where the transmission line is located includes: Based on the slope variance of each terrain region, the complexity index of each terrain region in the terrain feature pyramid is calculated. Based on the complexity index, the analysis scale of the preset deep learning model for different terrain regions is determined. When the complexity index is greater than a preset threshold, the analysis scale is a fine scale; otherwise, the analysis scale is a coarse scale. Based on the adaptive window mechanism in the preset deep learning model, the terrain feature pyramid is classified according to different analysis scales to obtain the terrain type of the terrain area where the transmission line is located.
[0008] In one possible implementation of this application, determining micro-meteorological parameters based on preset standards and the terrain type includes: The terrain-meteorological parameter comparison table corresponding to the preset standard is matched with the terrain type to obtain the matching result; Based on the matching results, the micro-meteorological parameters corresponding to the terrain type are determined.
[0009] In one possible implementation of this application, after constructing the risk grading model based on the comprehensive risk index, the method further includes: Calculate the mean, standard deviation, skewness coefficient, and kurtosis coefficient of the comprehensive risk index corresponding to each topographic region; Based on the mean, standard deviation, skewness coefficient, and kurtosis coefficient, the dynamic partitioning threshold is calculated.
[0010] In one possible implementation of this application, the calculation of the dynamic partitioning threshold based on the mean, standard deviation, skewness coefficient, and kurtosis coefficient includes: A threshold function is constructed based on the mean and standard deviation, and a division threshold is generated based on the threshold function. The skewness coefficient and the kurtosis coefficient are used as adjustment coefficients, and the division threshold is dynamically adjusted based on the adjustment coefficients to obtain a dynamic division threshold.
[0011] In one possible implementation of this application, the calculation of the comprehensive risk index based on the micro-topographic parameters and the micro-meteorological parameters includes: The micro-topographical parameters and the micro-meteorological parameters are added together according to their corresponding weight values to obtain the comprehensive risk index.
[0012] In one possible implementation of this application, the step of predicting the risk level of each terrain region using the risk grading model and dynamic classification threshold to obtain the risk grading result includes: Based on the risk classification model, the comprehensive risk index and dynamic division threshold of each terrain region are processed to obtain the initial cluster centers; Based on a preset clustering algorithm, the initial cluster centers are iteratively optimized multiple times to obtain multiple first cluster centers; Based on the first cluster center that converges in the final iteration, the risk level of each terrain region is divided to obtain the risk classification result.
[0013] In one possible implementation of this application, the step of dividing the risk level of each terrain region based on the first cluster center that has finally converged in the final iteration, and obtaining the risk classification result, includes: Calculate the boundary thresholds for each risk level based on the first cluster center that converges in the final iteration; Based on the boundary threshold, the risk level of each terrain region is divided to obtain the risk classification result.
[0014] To achieve the above objectives, a two-segment intelligent identification and risk classification system for transmission lines is also provided. The system includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of the two-segment intelligent identification and risk classification method for transmission lines as described above.
[0015] This application provides a method and system for intelligent identification and risk classification of two micro-segments of transmission lines. Compared to related technologies that use fixed or empirical thresholds, which often ignore the risks posed by micro-topographic changes and lead to inaccurate risk classification results that do not match the actual disaster distribution, this application acquires topographic elevation data of the area where the transmission line is located and extracts various micro-topographic parameters from this data. These micro-topographic parameters include curvature, slope, and topographic relief. Based on these parameters, the topographic type of the area where the transmission line is located and the corresponding micro-meteorological parameters are determined. These micro-meteorological parameters include wind acceleration coefficient and icing. The coefficients, combined with micro-topographic and micro-meteorological parameters, are used to calculate a comprehensive risk index. Then, based on the comprehensive risk index, a risk grading model is constructed. This model, along with a dynamic threshold, is used to predict the risk level of each terrain region, yielding a risk grading result. The risk grading model, combining micro-topographic and micro-meteorological parameters, is used to dynamically categorize terrain risk levels. Since the dynamic threshold is dynamically adjusted based on the terrain parameters corresponding to different terrain regions, the predicted hazard level of each terrain region can better reflect the actual operational risks of transmission lines, thus obtaining an accurate risk grading result. Attached Figure Description
[0016] Figure 1 This is a flowchart illustrating the first embodiment of the intelligent identification and risk classification method for two micro-segments of transmission lines according to this application. Figure 2 This is a schematic diagram of the system architecture involved in the two-micro-segment intelligent identification and risk classification method for transmission lines in this application; Figure 3 This is a schematic diagram of the terrain elevation data involved in the two-segment intelligent identification and risk classification method for transmission lines in this application; Figure 4 This is a schematic diagram of the overall execution process of the two-micro-segment intelligent identification and risk classification method for transmission lines in this application; Figure 5 This is a schematic diagram illustrating the risk classification of transmission lines involved in the two-micro-segment intelligent identification and risk classification method for transmission lines in this application; Figure 6 This is a schematic diagram of the device structure of the hardware operating environment involved in the embodiments of this application. Detailed Implementation
[0017] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.
[0018] This application provides a method for intelligent identification and risk classification of two micro-segments in transmission lines. In the first embodiment of this method, referring to... Figure 1 The method includes: Step S10: Obtain the topographic elevation data of the terrain area where the transmission line is located, and extract various micro-topographic parameters from the topographic elevation data, including curvature, slope and topographic relief. Step S20: Based on the micro-topographic parameters, determine the terrain type of the terrain area where the transmission line is located and the corresponding micro-meteorological parameters. The dynamic division threshold is obtained by dynamically adjusting the terrain parameters corresponding to different terrain areas. The micro-meteorological parameters include wind acceleration coefficient and icing coefficient. Step S30: Calculate the comprehensive risk index based on the micro-topographical parameters and the micro-meteorological parameters; Step S40: Construct a risk grading model based on the comprehensive risk index; Step S50: Using the risk grading model and dynamic classification threshold, the risk level of each terrain region is predicted to obtain the risk grading result.
[0019] This embodiment aims to: accurately classify the terrain type of the terrain area where the transmission line is located by combining micro-topographic parameters and dynamic classification thresholds, and dynamically classify the terrain risk level by combining micro-topographic parameters and micro-meteorological parameters, so that the predicted danger level of each terrain area is more in line with the actual operation risk of the transmission line, thereby obtaining an accurate risk classification result.
[0020] The specific steps are as follows: Step S10: Obtain the topographic elevation data of the terrain area where the transmission line is located, and extract various micro-topographic parameters from the topographic elevation data, including curvature, slope and topographic relief.
[0021] As an example, the two-micro segment intelligent identification and risk classification method for transmission lines can be applied to the two-micro segment intelligent identification and risk classification device for transmission lines, which belongs to the two-micro segment intelligent identification and risk classification equipment for transmission lines.
[0022] As an example, the two-micro-segment intelligent identification and risk classification method for transmission lines can also be applied to a two-micro-segment intelligent identification and risk classification system for transmission lines, as shown in the system architecture diagram below. Figure 2As shown, the system includes a model building module, a data reading and preprocessing module, a terrain identification module, and a risk classification module. The model building module is mainly used to build a risk classification model. The terrain identification module is mainly used to identify various terrain areas and determine the type of complex terrain. The data reading and preprocessing module is mainly used to obtain terrain elevation data of the area where the transmission line is located, output the elevation data as a fence file, and extract the slope, curvature, and terrain undulation of each calculation window. The risk classification module is mainly used to classify the risk of each terrain area by combining micro-topographic parameters and micro-meteorological parameters, and obtain the risk classification results.
[0023] As an example, topographic elevation data for different terrain regions, such as Figure 3 As shown, the terrain elevation data consists of the elevation data collected from various terrain areas where the transmission lines are located. The elevation data is saved as an elevation fence file using Global Mapper, and then the core parameters are preset using the Config function class using code software (such as Python), including DEM data path, risk classification standard, terrain parameter weight settings, calculation window size (3×3 for slope / curvature window, 5×5 for terrain relief window) and output file format.
[0024] As an example, after obtaining the terrain elevation data, it is necessary to preprocess the terrain elevation data. The elevation data file is read by the read_geotiff function to extract the elevation data matrix, spatial metadata (projection information, resolution, coordinate transformation parameters) and coordinate system. After the data extraction is completed, invalid values in the data are processed, converted to NaN (a missing value representation) and recorded to prepare for subsequent preprocessing; at the same time, basic data information is output to ensure data validity.
[0025] As an example, micro-topographic parameters mainly include slope, curvature, aspect, and topographic relief. These four parameters are used as the basic factors for risk assessment. Specifically, the calculation methods for the four micro-topographic parameters can be as follows: Slope calculation: Based on the Sobel operator, the surface tilt is calculated using the first derivative in the x / y directions. A 3×3 window convolution operation is then used, combined with DEM resolution, to convert the derivative into an actual slope value, reflecting the steepness of the terrain. The slope can be calculated in the following ways: It is the rate of change of elevation (z) in the x-direction (horizontal gradient); It is the rate of change of the elevation (z) coordinate in the y-direction (vertical gradient); Curvature Calculation: The degree of surface curvature is calculated using the second derivative. A combined index of Gaussian curvature and mean curvature is introduced, and the results are smoothed using Gaussian filtering. This distinguishes between concave, convex, and saddle-shaped landforms. Positive curvature represents convex terrain (such as ridges), and negative curvature represents concave terrain (such as valleys), reflecting differences in water flow convergence and soil stability. Curvature can be calculated in the following ways: : represent the first-order partial derivatives of x and y, respectively.
[0026] : denote the second-order partial derivatives of x and y, respectively.
[0027] Aspect calculation: Aspect (radians) calculated based on the first derivative: ; Convert to 0-360°; .
[0028] The method for calculating terrain undulation is as follows: based on the calculate_elevation_range function, the difference between the maximum and minimum elevation values within a local range is calculated through a 5×5 neighborhood rolling window to obtain the terrain undulation. The terrain undulation is used to reflect the severity of elevation changes within the region and to help evaluate the stability of tower site selection.
[0029] Step S20: Based on the micro-topographic parameters, determine the terrain type of the terrain area where the transmission line is located and the corresponding micro-meteorological parameters, including the wind acceleration coefficient and the icing coefficient.
[0030] As an example, a deep learning model is used to identify the micro-topographic parameters of various regions and determine the topographic type of each region. After determining the topographic type, micro-meteorological parameters are obtained through the meteorological parameter standards corresponding to each topographic type. Among them, the wind acceleration coefficient represents the wind speed amplification effect of micro-topography relative to open flat ground (its coefficient is 1.0), and the icing coefficient represents the degree of increase in icing thickness of micro-topography relative to the observation point of the standard meteorological station.
[0031] Step S20 includes: Step S21: Based on the standard deviation of each micro-topographic parameter, construct the topographic feature pyramid corresponding to the micro-topographic parameter.
[0032] As an example, constructing a terrain feature pyramid can be done by: constructing a Gaussian scale space, extracting terrain parameters such as slope and curvature in parallel at multiple different standard deviations, and then stacking the terrain parameters at all scales to form a pyramid-shaped data structure, thus obtaining the terrain feature pyramid, where: The bottom layer of the feature pyramid: high-resolution, richly detailed features, suitable for identifying small-scale terrain (such as erosion gullies and small collapses).
[0033] The top layer of the feature pyramid: low-resolution, highly generalized features, suitable for identifying large-scale terrain (such as mountains and large river valleys).
[0034] The middle layer of the feature pyramid: captures medium-scale features.
[0035] Step S22: Input the terrain feature pyramid into a preset deep learning model, and classify the terrain feature pyramid based on the preset deep learning model to obtain the terrain type of the terrain area where the transmission line is located.
[0036] As an example, after generating the terrain feature pyramid, it is input into a pre-defined deep learning model for classification. This model, such as U-Net, is well-suited for processing the input terrain feature pyramid and allows for inputting pyramids of different scales into different levels of the U-Net. The terrain feature pyramid addresses the "scale effect" problem in terrain analysis. Single-scale analysis may only be sensitive to terrain features of a specific size, while multi-scale pyramids ensure that regardless of the size of the target terrain unit, there is always a suitable scale to clearly capture it, thus accurately classifying the terrain type of the area where the transmission line is located.
[0037] As an example, when training the model, the model is trained using multi-scale feature data, and then the trained model is used to classify the terrain feature pyramid to obtain the terrain type of the area where the new power transmission line is located.
[0038] Step S22 includes: Based on the slope variance of each terrain region, the complexity index of each terrain region in the terrain feature pyramid is calculated.
[0039] Based on the complexity index, the analysis scale of the preset deep learning model for different terrain regions is determined. When the complexity index is greater than a preset threshold, the analysis scale is a fine scale; otherwise, the analysis scale is a coarse scale.
[0040] As an example, after determining the micro-topographic parameters, the slope data in the micro-topographic parameters is extracted, and the slope variance of each topographic region is calculated. Then, the complexity index of each topographic region in the topographic feature pyramid is calculated based on the slope variance. The complexity index is a normalized value with a value range of (0, 1).
[0041] As an example, the complexity index can also be based on the standard deviation of terrain elevation data, or by fusing multiple indicators of micro-topographic parameters to generate a comprehensive complexity index.
[0042] As an example, a complexity index is used to quantify the terrain complexity of each terrain region, thereby matching the corresponding analysis scale to different terrain regions. When the complexity index is high, the analysis scale is coarse, and vice versa. The preset threshold can be 0.6, 0.7, etc., without any specific limitation.
[0043] Based on the adaptive window mechanism in the preset deep learning model, the terrain feature pyramid is classified according to different analysis scales to obtain the terrain type of the terrain area where the transmission line is located.
[0044] As an example, by using an adaptive window mechanism in a pre-defined deep learning model, the terrain feature pyramid is classified according to different analysis scales (e.g., coarse or fine scales). The lower the complexity (the flatter the terrain), the closer it is to a coarse scale; the higher the complexity (the more rugged the terrain), the closer it is to a fine scale. In the process of terrain type classification, classification can be performed by setting thresholds. Threshold_ridge = μ_tpi + k * σ_tpi (k is an adjustable parameter, for example, 0.5). Threshold_valley = μ_tpi - k * σ_tpi, where μ_tpi represents the mean of TPI (micro-topographic parameter) within the window, and σ_tpi represents the standard deviation of TPI (micro-topographic parameter) within the window.
[0045] The TPI value of the point is compared with the dynamic threshold of its local region to determine its category.
[0046] When TPI_point > Threshold_ridge, the ridge is within a local area; When TPI_point < Threshold_valley, the valley is within a local range; Then, the final terrain type is obtained by using a threshold-based classification method.
[0047] Step S23: Determine micro-meteorological parameters based on preset standards and the terrain type.
[0048] Step S23 includes: The terrain-meteorological parameter comparison table corresponding to the preset standard is matched with the terrain type to obtain the matching result.
[0049] Based on the matching results, the micro-meteorological parameters corresponding to the terrain type are determined.
[0050] As an example, micrometeorological parameters can be determined by looking up tables. Specifically, the preset standard can be a topographic-meteorological parameter comparison table corresponding to Chinese standards. This table is then matched with the relevant table, and the wind acceleration coefficient and icing coefficient corresponding to each terrain type are determined based on the matching results. Taking the wind acceleration coefficient as an example: Emphasizing mountain ridges and cliffs: The wind acceleration coefficient is greatest at the mountain top, decreasing downwards along the windward and leeward slopes.
[0051] Mountain passes and mountain passes: Based on the ratio of their width to height, a fixed amplification factor (such as 1.2-1.5) is assigned, and then the wind acceleration coefficient under different terrains is determined.
[0052] The icing coefficients for various terrain types are as follows: Windward slope: Especially in areas with abundant moisture, the windward slope has the highest coefficient (e.g., 1.2-1.4).
[0053] Passes and mountaintops: High wind speeds lead to a higher impact rate of supercooled water droplets, resulting in increased icing (e.g., 1.1-1.3).
[0054] Leeward slopes and valleys: may be lighter (e.g., 0.9-1.0).
[0055] Step S30: Calculate the comprehensive risk index based on the micro-topographical parameters and the micro-meteorological parameters.
[0056] As an example, the comprehensive risk index for each topographic region is calculated using micro-topographic parameters and micro-meteorological parameters. It is not simply a matter of adding up the weights. The comprehensive risk index is used to reflect the interaction between various topographic factors.
[0057] Step S30 includes: The micro-topographical parameters and the micro-meteorological parameters are added together according to their corresponding weight values to obtain the comprehensive risk index.
[0058] As an example, each factor (such as slope, wind acceleration coefficient, etc.) in the micro-topographic parameters and micro-meteorological parameters is uniformly normalized to the interval [0, 1], where 0 represents no risk and 1 represents extremely high risk, to obtain the normalized risk value of each parameter. The normalization function itself can be nonlinear (such as S-shaped function, exponential function) to reflect the critical point effect.
[0059] As an example, the comprehensive risk index can be calculated as follows: CRI = 1-∏(1-W i * R i)^(1+α* Interaction_ij) Among them, CRI: Comprehensive Risk Index, W i The weight of the i-th factor, R i： The normalized risk value of the i-th factor, ∏: multiplication symbol, α * Interaction_ij: represents the interaction coefficient of other factors j on factor i (which can be determined through data mining).
[0060] The multiplication formula means that if any factor has an extremely high risk (close to 1), the overall risk will increase sharply. This is consistent with the "weakest link effect" and helps the model effectively capture the main problems.
[0061] Step S40: Construct a risk grading model based on the comprehensive risk index.
[0062] As an example, in the process of constructing a risk classification model, a comprehensive risk index is calculated, a dynamic classification threshold is preset, and a risk classification model is established to classify the risk levels of various terrain regions.
[0063] Step S50: Using the risk grading model and dynamic classification threshold, the risk level of each terrain region is predicted to obtain the risk grading result.
[0064] As an example, after establishing the risk classification model, the risk level of different terrain areas can be predicted using the model and the dynamic classification threshold, thereby realizing the visual classification of micro-topography and micro-meteorological risk levels.
[0065] Before step S50, the following steps are also included: Calculate the mean, standard deviation, skewness coefficient, and kurtosis coefficient of the comprehensive risk index corresponding to each terrain region.
[0066] Based on the mean, standard deviation, skewness coefficient, and kurtosis coefficient, the dynamic partitioning threshold is calculated.
[0067] As an example, for the comprehensive risk index corresponding to each type of terrain region, the mean, standard deviation, skewness coefficient, and kurtosis coefficient are calculated. The skewness coefficient is used to describe the asymmetry of the data distribution, and the kurtosis coefficient is used to describe the steepness of the data distribution.
[0068] As an example, a dynamic classification threshold is calculated using the mean, standard deviation, skewness coefficient, and kurtosis coefficient. The risk level of each terrain region is determined by the dynamic classification threshold of the comprehensive risk index for each terrain region.
[0069] As an example, the dynamic classification threshold is a threshold that is dynamically adjusted based on the terrain parameters of different terrain regions. The corresponding terrain parameter thresholds are different for different regions. If a fixed threshold is used for judgment, there will be cases where the risk classification is inaccurate. Therefore, the classification / segmentation thresholds of each terrain are dynamically adjusted through local statistical characteristics. For example, the curvature threshold for ridge identification is automatically increased in high-altitude areas, and the curvature threshold is decreased in low-altitude areas. The adaptability of the deep learning model to different landforms is adjusted by using dynamic classification thresholds.
[0070] The step of calculating the dynamic partitioning threshold based on the mean, standard deviation, skewness coefficient, and kurtosis coefficient includes: A threshold function is constructed based on the mean and standard deviation, and a division threshold is generated based on the threshold function.
[0071] As an example, when calculating the dynamic partitioning threshold, a threshold function is first constructed based on the mean and standard deviation to obtain a partitioning threshold, which is then used as the base threshold.
[0072] The skewness coefficient and the kurtosis coefficient are used as adjustment coefficients, and the division threshold is dynamically adjusted based on the adjustment coefficients to obtain a dynamic division threshold.
[0073] As an example, based on the skewness and kurtosis coefficients of each terrain region, adjustment coefficients are used to dynamically adjust the classification thresholds. Specifically, the adjustment method could be: Threshold setting (assuming the distribution is close to normal): T5 = μ+1.5σ, the dividing line between extremely high risk and extreme risk; T4 = μ+0.5σ; T3 = μ, the dividing line between medium and high risk can be referenced from the mean; T2 = μ-0.5σ; T1 = μ-1.5σ, the dividing line between low risk and very low risk, where μ represents the mean and σ represents the standard deviation.
[0074] Introducing a skewness coefficient (γ1) for nonlinear adjustment: Positive skewness (γ1 > 0) means that high-risk tails require more refined differentiation.
[0075] T5 = T5 + α*γ1*σ, where α is a gain coefficient (e.g., 0.3). In the case of positive skewness, the T5 threshold shifts to the right, tightening the standard for "extreme risk".
[0076] T4 = T4 + β*γ1*σ, β < α, which makes the interval (T5-T4) larger.
[0077] For negative skewness (γ1 < 0), the threshold on the low-risk side is adjusted accordingly.
[0078] Introducing kurtosis coefficient (β2) for tail adjustment: A thick tail with sharp peaks (β2 > 3) means more extreme values, but we need to ensure that the highest rank only captures the most extreme ones.
[0079] T5 = T5 + γ*(β2 - 3) * σ, where γ is a small positive coefficient that fine-tunes T5 to make it higher.
[0080] For low peak values (β2 < 3), a slight reduction in T5 can be considered.
[0081] Thus, dynamic thresholds for classifying five risk levels are obtained.
[0082] Step S50 includes: Step S51: Based on the risk classification model, process the comprehensive risk index and dynamic division threshold of each terrain region to obtain the initial cluster center.
[0083] As an example, we first calculate the initial statistic: the mean ( ), standard deviation ( ), skewness coefficient ( ), kurtosis coefficient ( ).
[0084] Mean: , where x i Let represent the comprehensive risk index of the i-th terrain region, and n represent the number of terrain regions.
[0085] Standard deviation: Skewness coefficient: Initial cluster centers were then generated based on statistical parameters. , where C k This represents the k-th initial cluster center.
[0086] Step S52: Based on the preset clustering algorithm, the initial cluster centers are iteratively optimized multiple times to obtain multiple first cluster centers.
[0087] As an example, the pre-defined clustering algorithm could be the K-means algorithm. It should be noted that the cluster centers are: (SK), among which, Indicates the indicator vector.
[0088] in This is a skewed operation function that adjusts the distribution of cluster centers after determining the initial cluster centers to adapt to the skewed characteristics of the data.
[0089] The objective function is Then, iterative calculations are performed to assign each sample to the nearest center, and the center of each cluster is recalculated to obtain multiple first cluster centers.
[0090] Step S53: Based on the first cluster center that has finally converged in the final iteration, the risk level of each terrain region is divided to obtain the risk classification result.
[0091] As an example, by adding a gradient variation coefficient constraint to the traditional K-means algorithm, we first calculate the gradient of adjacent cluster centers: .
[0092] Calculate the mean and standard deviation of the gradient: Gradient coefficient of variation: in and These are the mean and standard deviation of the gradient between adjacent cluster centers, respectively.
[0093] If the gradient variation coefficient is greater than the threshold (taken as 0.3), the position of the cluster center is adjusted, and then a large number of iterations are performed. The calculation is terminated when the center change is less than the tolerance or the gradient variation coefficient is less than the threshold for two consecutive iterations. Based on the final cluster center (6 categories), the risk classification result is output.
[0094] The risk classification result, obtained by dividing the risk levels of each terrain region based on the first cluster center that has finally converged in the final iteration, includes: Calculate the boundary thresholds for each risk level based on the first cluster center that converges in the final iteration.
[0095] Based on the boundary threshold, the risk level of each terrain region is divided to obtain the risk classification result.
[0096] As an example, in the process of classifying risk levels, a dynamic classification threshold can be used to classify the comprehensive risk corresponding to each terrain region, thereby determining the risk level of each terrain region.
[0097] It could also be that the final cluster centers are: And satisfy < < < < < ,in, This represents the first type of cluster center, and so on.
[0098] Statistical characteristics of each cluster: The mean, standard deviation, and sample size within each risk level (cluster) are output simultaneously for subsequent analysis.
[0099] Boundary threshold: Calculate the midpoint of adjacent cluster centers as a dynamic threshold for defining risk ranges. T1 = ( + ) / 2; T2 = ( + ) / 2; ... T5 = ( + ) / 2; Based on these dynamic thresholds, all risk indices are ultimately classified to obtain the risk classification results.
[0100] In this embodiment, the overall implementation process is illustrated as follows: Figure 4 As shown, the schematic diagram of transmission line risk classification is composed of... Figure 5 As shown.
[0101] This application provides a method and system for intelligent identification and risk classification of two micro-segments of transmission lines. Compared to related technologies that use fixed or empirical thresholds, which often ignore the risks posed by micro-topographic changes and lead to inaccurate risk classification results that do not match the actual disaster distribution, this application acquires topographic elevation data of the area where the transmission line is located and extracts various micro-topographic parameters from this data. These micro-topographic parameters include curvature, slope, and topographic relief. Based on these parameters, the topographic type of the area where the transmission line is located and the corresponding micro-meteorological parameters are determined. Meteorological parameters, including wind acceleration coefficient and icing coefficient, are combined with micro-topographical parameters and micro-meteorological parameters to calculate a comprehensive risk index. Based on this comprehensive risk index, a risk grading model is constructed to predict the risk level of various terrain regions, yielding risk grading results. Combining micro-topographical parameters and dynamic classification thresholds, the terrain types of the areas where transmission lines are located are accurately classified. Furthermore, based on the micro-topographical and micro-meteorological parameters, the terrain risk level is dynamically classified, ensuring that the predicted hazard level of each terrain region more closely matches the actual operational risks of the transmission lines, thus obtaining accurate risk grading results.
[0102] This application also provides a two-segment intelligent identification and risk classification system for transmission lines. The system includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of the two-segment intelligent identification and risk classification method for transmission lines as described above.
[0103] Reference Figure 6 , Figure 6 This is a schematic diagram of the device structure of the hardware operating environment involved in the embodiments of this application.
[0104] like Figure 6 As shown, the two-segment intelligent identification and risk classification device for transmission lines may include: a processor 1001, a memory 1005, and a communication bus 1002. The communication bus 1002 is used to realize the connection and communication between the processor 1001 and the memory 1005.
[0105] Optionally, the two-segment intelligent identification and risk classification device for transmission lines may also include a user interface, a network interface, a camera, RF (Radio Frequency) circuitry, sensors, a WiFi module, etc. The user interface may include a display screen and an input submodule such as a keyboard; optional user interfaces may also include standard wired or wireless interfaces. The network interface may include standard wired or wireless interfaces (such as a Wi-Fi interface).
[0106] Those skilled in the art will understand that Figure 6 The two-micro segment intelligent identification and risk classification device structure shown in the figure does not constitute a limitation on the two-micro segment intelligent identification and risk classification device for transmission lines. It may include more or fewer components than shown, or combine certain components, or have different component arrangements.
[0107] like Figure 6 As shown, the memory 1005, serving as a storage medium, may include an operating system, a network communication module, and a two-segment intelligent identification and risk classification program for transmission lines. The operating system is a program that manages and controls the hardware and software resources of the two-segment intelligent identification and risk classification device for transmission lines, supporting the operation of the two-segment intelligent identification and risk classification program for transmission lines, as well as other software and / or programs. The network communication module is used to enable communication between the various components within the memory 1005, as well as communication with other hardware and software in the two-segment intelligent identification and risk classification system for transmission lines.
[0108] exist Figure 6In the two-segment intelligent identification and risk classification device for transmission lines shown, the processor 1001 is used to execute the two-segment intelligent identification and risk classification program for transmission lines stored in the memory 1005, and implement the steps of the two-segment intelligent identification and risk classification method for transmission lines described in any of the above claims.
[0109] The specific implementation method of the two-micro segment intelligent identification and risk classification device for transmission lines in this application is basically the same as the embodiments of the two-micro segment intelligent identification and risk classification method for transmission lines described above, and will not be repeated here.
[0110] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.
[0111] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.
[0112] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, 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 is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.
[0113] The above are merely preferred embodiments of this application and do not limit the scope of this application. Any equivalent structural or procedural transformations made based on the description and drawings of this application, or direct or indirect applications in other related technical fields, are similarly included within the scope of protection of this application.
[0114] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0115] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.
Claims
1. A method for two-micro-section intelligent identification and risk grading for power transmission lines, characterized in that, The method comprises: obtaining terrain elevation data of a terrain area where a power transmission line is located, and extracting a plurality of micro-terrain parameters in the terrain elevation data, the micro-terrain parameters including curvature, slope and terrain relief degree; based on the micro-terrain parameters, determining the terrain type of the terrain area where the power transmission line is located and the micro-meteorological parameters corresponding to the terrain type, the micro-meteorological parameters including wind acceleration coefficient and icing coefficient; based on the micro-terrain parameters and the micro-meteorological parameters, a comprehensive risk index is calculated; based on the comprehensive risk index, a risk classification model is constructed; through the risk classification model and the dynamic division threshold, the risk level of each terrain area is predicted and processed to obtain a risk classification result, and the dynamic division threshold is obtained based on the corresponding terrain parameters of different terrain areas.
2. The method for two-section intelligent identification and risk grading of power transmission lines of claim 1, wherein, Based on the micro-terrain parameters, the terrain type of the terrain area where the power transmission line is located and the micro-meteorological parameters corresponding to the terrain type are determined, comprising: based on the standard deviation of each micro-terrain parameter, a terrain feature pyramid corresponding to the micro-terrain parameter is constructed; input the terrain feature pyramid into a preset deep learning model, and based on the preset deep learning model, the terrain feature pyramid is classified to obtain the terrain type of the terrain area where the power transmission line is located; based on a preset standard and the terrain type, the micro-meteorological parameters are determined.
3. The method for two-section intelligent identification and risk grading of power transmission lines of claim 2, wherein, Based on the preset deep learning model, the terrain feature pyramid is classified to obtain the terrain type of the terrain area where the power transmission line is located, comprising: based on the slope variance of each terrain area, the complexity index of each terrain area in the terrain feature pyramid is calculated; based on the complexity index, the analysis scale of the preset deep learning model for different terrain areas is determined, wherein when the complexity index is greater than a preset threshold, the analysis scale is fine, otherwise, the analysis scale is coarse; based on the adaptive window mechanism in the preset deep learning model, the terrain feature pyramid is classified according to different analysis scales to obtain the terrain type of the terrain area where the power transmission line is located.
4. The method for two-section intelligent identification and risk grading of power transmission lines of claim 2, wherein, Based on the preset standard and the terrain type, the micro-meteorological parameters are determined, comprising: matching the terrain-meteorological parameter table corresponding to the preset standard with the terrain type to obtain a matching result; based on the matching result, the micro-meteorological parameters corresponding to the terrain type are determined.
5. The method for two-section intelligent identification and risk grading of power transmission lines of claim 1, wherein, Based on the micro-terrain parameters and the micro-meteorological parameters, a comprehensive risk index is calculated, comprising: adding the micro-terrain parameters and the micro-meteorological parameters according to the corresponding weight values to obtain a comprehensive risk index.
6. The method for two-section intelligent identification and risk grading of power transmission lines of claim 5, wherein, After constructing the risk classification model based on the comprehensive risk index, it further comprises: calculating the mean, standard deviation, skewness coefficient and kurtosis coefficient corresponding to the comprehensive risk index of each terrain area; based on the mean, standard deviation, skewness coefficient and kurtosis coefficient, a dynamic division threshold is calculated.
7. The method for two-section intelligent identification and risk grading of power transmission lines of claim 6, wherein, Based on the mean, standard deviation, skewness coefficient and kurtosis coefficient, a dynamic division threshold is calculated, comprising: construct a threshold function based on the mean and the standard deviation, and generate a division threshold based on the threshold function; take the skewness coefficient and the kurtosis coefficient as adjustment coefficients, and dynamically adjust the division threshold based on the adjustment coefficients to obtain a dynamic division threshold.
8. The method for transmission line oriented two section intelligent identification and risk grading as claimed in claim 1 wherein, The risk classification result includes: The risk classification result includes: The risk classification result includes: The risk classification result includes:
9. The method for two-section intelligent identification and risk grading of power transmission lines of claim 8, wherein, The risk classification result includes: The risk classification result includes: The risk classification result includes:
10. A two-section intelligent identification and risk grading system for power transmission lines, characterized in that, The system includes a memory, a processor, and a computer program stored in the memory and running on the processor, and when the processor executes the computer program, the steps of the method of any one of claims 1-9 are implemented.