Wire icing early warning system based on terrain analysis

Abstract

The invention relates to the technical field of power system safety monitoring, in particular to a wire icing early warning system based on terrain analysis, which comprises a terrain analysis module, a microclimate simulation module, a meteorological data acquisition module, an icing risk assessment module, a real-time monitoring module and an early warning information generation module. Wherein the topographic analysis module is used for acquiring topographic feature data of a target area; the microclimate simulation module is used for performing microscale climate simulation on the target area to generate local meteorological parameters; the meteorological data acquisition module is used for acquiring real-time meteorological data of the target area; and the icing risk assessment module is used for constructing an icing risk assessment model. According to the method, accurate and real-time evaluation and early warning of the icing risk of the electric wire in the complex terrain area are realized by integrating the terrain features, the real-time meteorological data and the historical icing information and adopting the random forest algorithm and the microclimate simulation technology.

Description

Technical Field

[0001] The present invention relates to the technical field of power system safety monitoring, and in particular to a wire icing early warning system based on terrain analysis. Background Art

[0002] Icing on power lines is a common and serious natural disaster in power systems, especially in complex terrain areas such as mountains and hills. Icing will significantly increase the weight of power lines, causing increased sag and even line breakage due to insufficient mechanical strength, seriously threatening the safe and stable operation of power systems. Existing icing monitoring and early warning technologies mostly rely on data from single-point sensors or a small number of meteorological stations, and cannot fully reflect the spatial distribution characteristics of icing risks in the target area. In addition, meteorological conditions in complex terrain areas change rapidly, and traditional monitoring methods cannot quickly respond to real-time changes in the environment, resulting in untimely icing risk warnings or inaccurate prediction results.

[0003] At present, the existing technology has shortcomings in the following aspects: First, it fails to fully combine the complex terrain features to conduct a detailed analysis of the icing risk, and it is difficult to reflect the impact of the terrain in the region on the local meteorological conditions; second, there is a lack of high-precision microclimate simulation methods, and traditional methods are difficult to capture the fine distribution of meteorological elements such as wind speed, temperature and humidity; third, the risk assessment model is often constructed by ignoring the input and adjustment of dynamic real-time meteorological data, and the model prediction generalization ability is weak and the adaptability is insufficient. Therefore, it is of great practical significance and technical value to develop an early warning system that can combine terrain characteristics, real-time meteorological data and historical data to perform high-precision icing risk prediction. Summary of the invention

[0004] Based on the above objectives, the present invention provides a wire icing warning system based on terrain analysis.

[0005] A wire icing warning system based on terrain analysis includes a terrain analysis module, a microclimate simulation module, a meteorological data acquisition module, an icing risk assessment module, a real-time monitoring module and a warning information generation module; wherein: Terrain analysis module: used to obtain digital elevation model data of the target area, and calculate the slope, slope aspect, altitude and terrain roughness of the target area based on the digital elevation model data to form terrain feature data of the target area; Microclimate simulation module: Based on terrain feature data, computational fluid dynamics and numerical weather forecast models are used to simulate the micro-scale climate of the target area and generate local meteorological parameters, including refined wind field, temperature field and humidity field; Meteorological data acquisition module: used to deploy a distributed meteorological sensor array in the target area to collect real-time meteorological data of the target area, including temperature, humidity, precipitation and wind speed; Icing risk assessment module: Based on terrain feature data, local meteorological parameters and historical icing data, an icing risk assessment model is constructed using the random forest algorithm; Real-time monitoring module: receives real-time meteorological data provided by the meteorological data acquisition module, inputs the real-time meteorological data into the constructed icing risk assessment model, performs real-time risk prediction, and outputs the predicted value of ice thickness; Early warning information generation module: Generates early warning information for the target area based on the risk prediction results output by the real-time monitoring module.

[0006] Optionally, the terrain analysis module includes a data acquisition unit, a slope calculation unit, a slope aspect calculation unit, an altitude calculation unit, a terrain roughness calculation unit and a feature integration unit; wherein: Data acquisition unit: used to acquire digital elevation model data of the target area through satellite remote sensing data source, the digital elevation model data includes surface elevation information represented in the form of a grid, and perform format conversion and projection coordinate correction on the data to ensure data accuracy; Slope calculation unit: based on the elevation value change of the digital elevation model data, the slope value of each grid unit is calculated by differential operation, and the slope value represents the ratio of the elevation change with the horizontal distance; Aspect calculation unit: based on the digital elevation model data and the slope calculation result, the aspect value of each grid cell is calculated, and the aspect value represents the azimuth of the slope surface relative to the north direction; Elevation calculation unit: extract the elevation value of each grid cell in the digital elevation model data, and directly obtain the elevation distribution of each grid cell in the target area; Terrain roughness calculation unit: by analyzing the degree of change of local elevation in the digital elevation model data, the terrain roughness of each grid cell in the target area is calculated, and the terrain roughness is expressed as the standard deviation of the elevation change around the grid cell; Feature integration unit: used to integrate the calculation results of slope, aspect, altitude and terrain roughness to generate terrain feature data of the target area.

[0007] Optionally, the microclimate simulation module includes a grid division unit, a fluid dynamics solving unit, an energy transfer solving unit, a water vapor transport solving unit and an output generation unit; wherein: Meshing unit: used to mesh the target area according to terrain feature data and generate a three-dimensional mesh model for computational fluid dynamics simulation; Fluid dynamics solving unit: Based on the computational fluid dynamics model, the Reynolds-averaged Navier-Stokes equations are used to solve the fluid dynamics and simulate the airflow distribution in the target area. The calculation formula is: ,in, is the gas density, and The airflow is and The velocity component in the direction, For pressure, is the dynamic viscosity, is the volume force; through iterative calculation, the refined wind field distribution of the target area is obtained; Energy transfer solving unit: Based on the energy conservation equation in the numerical weather forecast model, the temperature field in the target area is simulated. The calculation formula is: ,in, is the temperature, is the specific heat capacity, is the thermal conductivity, is the heat source term; Water vapor transport solution unit: Based on the water vapor transport equation, simulate the humidity field distribution in the target area. The calculation formula is: ,in, For humidity, is the water vapor diffusion coefficient, is the water vapor source term; Output generation unit: Integrate the calculation results of the fluid dynamics solution unit, energy transfer solution unit and water vapor transport solution unit to generate local meteorological parameters of the target area and output them in the form of three-dimensional grid.

[0008] Optionally, the grid division unit includes: Data preprocessing: normalize the terrain feature data provided by the terrain analysis module and normalize the elevation values ​​to ensure that the input data is within a uniform scale; Terrain complexity analysis: Calculate the mesh refinement requirements in the target area based on the terrain roughness and slope; Initial grid generation: Based on the boundary data of the target area, the quadtree partitioning algorithm is used to generate an initial regular grid to divide the target area into a number of initial square units; Mesh refinement: Adaptive mesh refinement is used to adjust the initial mesh according to the results of terrain complexity analysis. For areas with greater roughness or higher slopes, the side length of the mesh unit is refined. For areas with gentle terrain changes, the resolution of the initial mesh unit is kept unchanged. Boundary optimization: Optimize the grid cell boundaries and generate unstructured grids using the Delaunay triangulation method to minimize the maximum angle of the triangle; Mesh verification and output: Verify the quality of mesh cells. Once the shape factors of all meshes meet the quality threshold requirements, the optimized mesh data will be output.

[0009] Optionally, the meteorological data acquisition module includes a sensor selection unit, a sensor layout unit, a data transmission unit, a sensor calibration unit and a data storage unit; wherein: Sensor selection unit: used to deploy distributed meteorological sensor arrays in the target area, each sensor array includes a temperature sensor, a humidity sensor, a precipitation meter and an anemometer; The sensor layout unit divides the target area based on the terrain feature data provided by the terrain analysis module and determines the layout positions of the sensors according to the following rules; Rule 1: Place high-density sensor arrays in areas with high slopes and high roughness; Rule 2: Deploy low-density sensor arrays in areas with flat terrain; Data transmission unit: used to transmit the collected meteorological data to the system's central data device in real time for standardization; Sensor calibration unit: used to periodically calibrate deployed sensors. Specifically, the collected temperature, humidity, precipitation and wind speed are compared with data from standard weather stations or known reference equipment to adjust the measurement deviation of the sensors to ensure the accuracy of the collected data. Data storage unit: used to store the collected meteorological data in different regions, and the meteorological data of each region is archived according to the timestamp.

[0010] Optionally, the icing risk assessment module includes a sample construction unit, a model training unit, a model verification unit, a parameter optimization unit and a model output unit; wherein: Sample construction unit: Based on historical ice cover data, the terrain feature data and local meteorological parameters of the target area are matched with the corresponding ice cover risk level labels to construct training sample sets and test sample sets; the training sample set is used for model training, and the test sample set is used for model evaluation; Model training unit: Use the random forest algorithm to train the training sample set, and classify the data by building multiple decision trees. Each decision tree takes a subset of the feature vector as input and the icing risk level as output. The outputs of all decision trees are voted on comprehensively to form an icing risk prediction model. The model verification unit verifies the trained risk assessment model using the test sample set, and calculates the model's performance indicators, including accuracy, error rate, and classification accuracy, by comparing the model's prediction results with the actual risk level labels; Parameter optimization unit: Use cross-validation method to optimize the hyperparameters of the icing risk assessment model, including the number of decision trees, maximum depth, and number of features used by each tree, to ensure the generalization ability of the model on different sample sets; at the same time, dynamically adjust the hyperparameters according to the performance indicators of the model on the test sample set; Model output unit: used to save the final optimized icing risk assessment model as a standardized model file for real-time monitoring module to call.

[0011] Optionally, the model training unit specifically includes: Training sample initialization: receiving the training sample set provided by the sample construction unit, each sample consists of an input feature vector and a corresponding icing risk level label; Decision Tree Construction: Random Forest generates multiple decision trees by repeating the following steps multiple times: Step 1: Randomly select a predetermined number of samples from the training sample set using the bootstrap sampling method to construct a training subset of a decision tree, and the remaining samples are used as the validation set; Step 2: Randomly select some features from all features as the candidate set of split features for the current decision tree. The number of selected features is usually less than the total number of features. Step 3: At each node, select the optimal split feature and split point from the candidate feature set to optimize the purity of the node after splitting; Decision tree training: Recursively split the training subset of each decision tree until any of the set conditions are met, including that all samples in the current node belong to the same category, the number of samples is lower than the preset threshold, and the decision tree reaches the preset maximum depth; Decision tree ensemble: All decision trees are combined into a random forest model. The prediction result of the random forest is the majority vote of the classification results of all decision trees, that is, the classification with the most occurrences is selected as the final prediction result. Model output: After training is completed, the random forest model is saved as an icing risk assessment model.

[0012] Optionally, the parameter optimization unit includes: Initial parameter setting: Set the initial hyperparameters of the random forest algorithm, including the number of decision trees, the maximum depth, the number of features used by each tree, and the minimum number of node sample splits, as the initial parameter set for cross-validation; Cross-validation grouping: Divide the training sample set into In each iteration, one subset is selected as the validation set, and the rest subsets as training sets, and repeat times, making sure each subset is used as a validation set once; Model training and validation: In each iteration, the current hyperparameter settings are used to train The icing risk assessment model was trained using the training subsets, and the performance indicators of the model, including precision, recall, and F1 score, were calculated using the validation set. Performance Indicator Summary and Evaluation: Summary The verification results of the iterations are used to calculate the average performance index of the icing risk assessment model as a comprehensive evaluation index for the current hyperparameter settings; Dynamic adjustment of hyperparameters: Based on the average performance index of the validation set, the hyperparameter combination is adjusted, and the optimal hyperparameter combination is determined by grid search or random search method; Final hyperparameter determination: Select the hyperparameter combination with the best verification results and record the hyperparameter settings as the final parameters of the random forest model; Test sample performance verification: Use the test sample set to conduct final verification of the optimized icing risk assessment model, and calculate the performance indicators of the icing risk assessment model on the test sample set, including accuracy, error rate, and classification accuracy; if the test results do not meet the preset performance targets, the hyperparameters will be adjusted again and the above optimization steps will be repeated until the performance requirements are met; Parameter output: The optimized hyperparameters and the final verified performance indicators are recorded and saved as the configuration information file of the icing risk assessment model.

[0013] Optionally, the real-time monitoring module includes a data receiving unit, a data preprocessing unit, a feature matching unit and a model calling unit; wherein: Data receiving unit: used to receive real-time meteorological data provided by the meteorological data acquisition module, including temperature, humidity, precipitation and wind speed; Data preprocessing unit: used to preprocess the received real-time meteorological data, including data cleaning, missing value filling and outlier detection, to ensure the integrity of the input data; Feature matching unit: matches the preprocessed real-time meteorological data with the input features of the icing risk assessment model to ensure that the data format and dimension are consistent with the model requirements; Model calling unit: calls the icing risk assessment model constructed in the icing risk assessment module, inputs the matched real-time meteorological data into the model, performs real-time icing risk prediction, and outputs the risk prediction results.

[0014] Optionally, the warning information generation module includes a risk grading unit, a thickness analysis unit, a high-risk point positioning unit, an information integration unit and a warning information output unit; wherein: Risk grading unit: used to receive the risk prediction results output by the real-time monitoring module, and compare the ice risk value in the risk prediction results with the preset risk level threshold, and grade them into mild risk, moderate risk and severe risk to generate the risk level of the target area; Thickness analysis unit: Based on the predicted ice thickness value in the risk prediction results, combined with the meteorological conditions and terrain characteristics of the target area, generate ice thickness distribution maps for different areas to clarify the spatial variation of ice thickness; High-risk point positioning unit: By analyzing the risk level distribution and ice thickness prediction results, identify the areas with medium and above ice risk levels in the target area, determine the specific locations of high-risk points, and generate a list of high-risk points with corresponding geographic coordinates marked; Information integration unit: Integrate the data generated by the risk classification unit, thickness analysis unit and high-risk point positioning unit to generate complete warning information for the target area, including risk level distribution map, ice thickness distribution map and high-risk point list; Warning information output unit: outputs the integrated warning information in digital form for query by power maintenance departments and relevant monitoring personnel, and supports displaying warning information in the form of maps, text reports or tables.

[0015] Beneficial effects of the present invention: The present invention combines terrain feature data, real-time meteorological data and historical icing data, adopts an improved random forest algorithm for risk assessment, and uses computational fluid dynamics and numerical weather forecast models for microclimate simulation, thereby effectively solving the problem of insufficient accuracy in predicting icing risks in complex terrain areas. It also introduces refined terrain analysis and meteorological simulation technology to capture the fine distribution characteristics of meteorological elements such as wind speed, temperature and humidity in the target area, thereby achieving accurate assessment of icing risks.

[0016] The present invention, by comprehensively integrating historical data, terrain features and real-time environmental data, effectively improves the adaptability and generalization ability of the risk prediction model, provides a scientific basis for real-time monitoring and risk warning of icing of electric wires in the target area, and significantly reduces the risk of disconnection accidents and economic losses caused by icing of power lines. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1A schematic diagram of an electric wire ice warning system according to an embodiment of the present invention; Figure 2 Schematic diagram of a risk assessment module according to an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0020] It should be noted that the references to "one embodiment", "embodiment", "exemplary embodiments", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0021] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0022] like Figure 1-Figure 2 As shown, a wire ice warning system based on terrain analysis includes a terrain analysis module, a microclimate simulation module, a meteorological data acquisition module, an ice risk assessment module, a real-time monitoring module and a warning information generation module; wherein: Terrain analysis module: used to obtain the digital elevation model (DEM) data of the target area, and calculate the slope, slope aspect, altitude and terrain roughness of the target area based on the digital elevation model data to form the terrain feature data of the target area; Microclimate simulation module: Based on terrain feature data, computational fluid dynamics (CFD) and numerical weather forecast models are used to simulate the micro-scale climate of the target area and generate local meteorological parameters, including refined wind field, temperature field and humidity field; local meteorological parameters are used to build an icing risk assessment model; Meteorological data acquisition module: used to deploy a distributed meteorological sensor array in the target area to collect real-time meteorological data of the target area, including temperature, humidity, precipitation and wind speed; Icing risk assessment module: Based on terrain feature data, local meteorological parameters generated by the microclimate simulation module and historical icing data, an icing risk assessment model is constructed using the random forest algorithm; Real-time monitoring module: receives real-time meteorological data provided by the meteorological data acquisition module, inputs the real-time meteorological data into the constructed icing risk assessment model, performs real-time risk prediction, and outputs the predicted value of ice thickness; Early warning information generation module: Generates early warning information for the target area based on the risk prediction results output by the real-time monitoring module.

[0023] The terrain analysis module includes a data acquisition unit, a slope calculation unit, a slope aspect calculation unit, an altitude calculation unit, a terrain roughness calculation unit and a feature integration unit; wherein: Data acquisition unit: used to obtain digital elevation model (DEM) data of the target area through satellite remote sensing data source. The digital elevation model data includes surface elevation information represented in the form of a grid, and performs format conversion and projection coordinate correction on the data to ensure data accuracy; Slope calculation unit: Based on the elevation value change of the digital elevation model data, the slope value of each grid unit is calculated using differential operation. The slope value represents the ratio of the elevation change with the horizontal distance. The calculation formula of the slope value is: ,in, is the slope value, Indicates that the digital elevation model data is in The horizontal elevation change rate is calculated as follows: ; Indicates that the digital elevation model data is in The horizontal elevation change rate is calculated as follows: ; is the elevation difference; and are the horizontal distances of the grid cells respectively; Aspect calculation unit: Based on the digital elevation model data and slope calculation results, the aspect value of each grid cell is calculated. The aspect value represents the azimuth of the slope surface relative to the north direction. The calculation formula of the aspect value is: ,in, is the slope value, and the result is adjusted to to range, and Respectively represent the digital elevation model data in and The rate of change of elevation in the horizontal direction; Elevation calculation unit: extract the elevation value of each grid cell in the digital elevation model data , directly obtain the altitude distribution of each grid cell in the target area without complicated calculations, where Indicates the elevation value of a grid cell; Terrain roughness calculation unit: By analyzing the degree of change in local elevation in the digital elevation model data, the terrain roughness of each grid cell in the target area is calculated. The terrain roughness is expressed as the standard deviation of the elevation change around the grid cell. The calculation formula for roughness is: ,in, is the terrain roughness, is the elevation value of a grid cell, is the average elevation around the grid cell, is the number of surrounding grid cells; Feature integration unit: used to integrate the calculation results of slope, aspect, altitude and terrain roughness to generate terrain feature data of the target area and provide input for subsequent modules; through the collaborative work of the above units, the terrain analysis module can accurately and efficiently extract the terrain feature data of the target area and provide basic data support for the microclimate simulation module and icing risk assessment module.

[0024] The microclimate simulation module includes a grid division unit, a fluid dynamics solution unit, an energy transfer solution unit, a water vapor transport solution unit, and an output generation unit; among which: Meshing unit: used to mesh the target area according to the terrain feature data and generate a three-dimensional mesh model for computational fluid dynamics simulation. The mesh model is based on the terrain characteristics of the target area to ensure mesh refinement in complex terrain areas. Fluid dynamics solving unit: Based on the computational fluid dynamics model, the Reynolds-averaged Navier-Stokes (RANS) equations are used to solve the fluid dynamics and simulate the airflow distribution in the target area. The calculation formula is: ,in, is the gas density, and The airflow is and The velocity component in the direction, For pressure, is the dynamic viscosity, is the volume force; through iterative calculation, the refined wind field distribution of the target area is obtained; Energy transfer solving unit: Based on the energy conservation equation in the numerical weather forecast model, the temperature field in the target area is simulated. The calculation formula is: ,in, is the temperature, is the specific heat capacity, is the thermal conductivity, is the heat source term, and the refined temperature field distribution of the target area is obtained by calculation; Water vapor transport solution unit: Based on the water vapor transport equation, simulate the humidity field distribution in the target area. The calculation formula is: ,in, For humidity, is the water vapor diffusion coefficient, is the water vapor source term, and the calculation result is used to describe the spatial variation of humidity; Output generation unit: Integrate the calculation results of the fluid dynamics solution unit, energy transfer solution unit and water vapor transport solution unit to generate local meteorological parameters of the target area, including refined wind field, temperature field and humidity field, and output them in the form of a three-dimensional grid; through the collaborative work of the above units, the microclimate simulation module can fully combine the terrain feature data and real-time meteorological data of the target area, accurately simulate the microclimate changes under complex terrain, generate high-precision local meteorological parameters, and provide reliable data support for icing risk assessment.

[0025] The mesh division units include: Data preprocessing: normalize the terrain feature data provided by the terrain analysis module and normalize the elevation values ​​to ensure that the input data is within a uniform scale; Terrain complexity analysis: Based on the terrain roughness and slope of the target area, the mesh refinement requirements in the area are calculated. The density of the mesh division is determined by the weighted sum of the terrain roughness and slope, and the proportional coefficient is used to adjust the mesh resolution of the target area. Initial grid generation: Based on the boundary data of the target area, the quadtree partitioning algorithm is used to generate the initial regular grid to divide the target area into a number of initial square units. The side length of the initial grid unit is determined according to the area of ​​the region and the number of grid units. Mesh refinement: Adaptive mesh refinement is used to adjust the initial mesh according to the results of terrain complexity analysis. For areas with greater roughness or higher slopes, the side length of the mesh unit is refined. For areas with gentle terrain changes, the resolution of the initial mesh unit is kept unchanged. Boundary optimization: To ensure the smoothness of the mesh boundary and the calculation accuracy, the mesh cell boundary is optimized and the Delaunay triangulation method is used to generate an unstructured mesh to minimize the maximum angle of the triangle and ensure that the geometry of the mesh is close to uniform. Mesh verification and output: Verify the quality of the mesh units. Once the shape factors of all meshes meet the quality threshold requirements, the optimized mesh data will be output and provided to the fluid dynamics solution unit.

[0026] The specific steps for meshing the target area are as follows: First, based on the terrain roughness of the target area and slope , calculate the mesh refinement requirements in the area; the calculation formula for the density of mesh division is: ,in, is the grid density, is the scale factor, which is used to adjust the grid resolution of the target area. is the terrain roughness, is the slope; Then, based on the boundary data of the target area , divide the target area into several initial square units, the side length of the initial square unit According to the area and the number of grid cells OK, the formula is: ; Then, according to the results of terrain complexity analysis, the initial grid is adjusted using the adaptive grid refinement method; Large or slope In higher areas, refine the mesh element side length , the formula is: ,in, is the refinement factor, adjust the refinement strength according to the needs; Subsequently, in order to ensure the smoothness of the mesh boundary and the calculation accuracy, the mesh unit boundary is optimized, and the Delaunay triangulation method is used to generate an unstructured mesh to minimize the maximum angle of the triangle and ensure that the geometry of the mesh is close to uniform. The optimization objective function is: ,in, is the interior angle of each triangle; Finally, verify the quality of the mesh elements and ensure that all mesh shape factors The following conditions are met: ,in, is the grid quality threshold, which is generally 0.3; and the final optimized grid data is output and provided to the fluid dynamics solving unit.

[0027] The meteorological data acquisition module includes a sensor selection unit, a sensor layout unit, a data transmission unit, a sensor calibration unit and a data storage unit; wherein: Sensor selection unit: used to deploy distributed meteorological sensor arrays in the target area. Each sensor array includes temperature sensors, humidity sensors, precipitation meters and anemometers. The sensors are weather-resistant and have long-distance data transmission capabilities, suitable for complex terrain and extreme climate environments. The sensor layout unit divides the target area based on the terrain feature data provided by the terrain analysis module and determines the layout positions of the sensors according to the following rules; Rule 1: Deploy high-density sensor arrays in areas with high slopes and high roughness to ensure data coverage accuracy; Rule 2: Arrange low-density sensor arrays in areas with flat terrain to optimize the allocation of sensor resources; the spacing is determined based on the complexity of the terrain and the area of ​​the target area to ensure that there are no blind spots in the coverage area; Data transmission unit: used to transmit the collected meteorological data to the system's central data device in real time for standardization; Sensor calibration unit: used to periodically calibrate deployed sensors. Specifically, the collected temperature, humidity, precipitation and wind speed are compared with data from standard weather stations or known reference equipment to adjust the measurement deviation of the sensors to ensure the accuracy of the collected data. Data storage unit: used to store the collected meteorological data in different regions. The meteorological data of each region is archived according to the timestamp to facilitate the subsequent calling of the icing risk assessment module and the microclimate simulation module. Through the collaborative work of the above units, the meteorological data acquisition module can scientifically arrange the distributed meteorological sensor array according to the terrain characteristics and monitoring needs of the target area, realize full coverage and high-precision acquisition of meteorological data in the target area, and provide reliable real-time meteorological data support for the real-time monitoring module and the microclimate simulation module.

[0028] The icing risk assessment module includes a sample construction unit, a model training unit, a model verification unit, a parameter optimization unit and a model output unit; among which: Sample construction unit: Based on historical ice cover data, the terrain feature data and local meteorological parameters of the target area are matched with the corresponding ice cover risk level labels to construct training sample sets and test sample sets; the training sample set is used for model training, and the test sample set is used for model evaluation; Model training unit: Use the random forest algorithm to train the training sample set, and classify the data by building multiple decision trees. Each decision tree takes a subset of the feature vector as input and the icing risk level as output. The outputs of all decision trees are voted on comprehensively to form an icing risk prediction model. The model verification unit verifies the trained risk assessment model using the test sample set, and calculates the model's performance indicators, including accuracy, error rate, and classification accuracy, by comparing the model's prediction results with the actual risk level labels; Parameter optimization unit: Use cross-validation method to optimize the hyperparameters of the icing risk assessment model, including the number of decision trees, maximum depth, and number of features used by each tree, to ensure the generalization ability of the model on different sample sets; at the same time, dynamically adjust the hyperparameters according to the performance indicators of the model on the test sample set; Model output unit: used to save the final optimized icing risk assessment model as a standardized model file for the real-time monitoring module to call, so as to facilitate the subsequent real-time risk prediction; through the collaborative work of the above units, the icing risk assessment module can use terrain feature data, local meteorological parameters and historical icing data to build an accurate and reliable random forest algorithm model, providing a scientific basis for the prediction of icing risks of power lines in the target area.

[0029] The model training unit specifically includes: Training sample initialization: receiving the training sample set provided by the sample construction unit, each sample consists of an input feature vector and a corresponding icing risk level label; Decision Tree Construction: Random Forest generates multiple decision trees by repeating the following steps multiple times: Step 1: Randomly select a predetermined number of samples (duplicates are allowed) from the training sample set using the bootstrap sampling method to construct a training subset of a decision tree, and the remaining samples are used as the validation set; Step 2: Randomly select some features from all features as the candidate set of split features for the current decision tree. The number of selected features is usually less than the total number of features. Step 3: At each node, select the optimal split feature and split point from the candidate feature set to optimize the purity of the node after splitting; Decision tree training: Recursively split the training subset of each decision tree until any of the set conditions are met, including that all samples in the current node belong to the same category, the number of samples is lower than the preset threshold, and the decision tree reaches the preset maximum depth; Decision tree ensemble: All decision trees are combined into a random forest model. The prediction result of the random forest is the majority vote of the classification results of all decision trees, that is, the classification with the most occurrences is selected as the final prediction result. Model output: After training is completed, the random forest model is saved as an icing risk assessment model.

[0030] The specific calculation steps for training the training sample set are as follows: First, the training sample set provided by the sample construction unit is received, and each sample consists of an input feature vector and the corresponding icing risk level label Composition, of which Contains data on terrain characteristics of the target area (such as slope, aspect, altitude, and terrain roughness) and local meteorological parameters (such as wind speed, temperature, and humidity). is a classification label indicating the ice risk level; Then, random forest generates multiple decision trees by repeating the following steps multiple times ; Next, the training subset of each decision tree is recursively split until one of the following conditions is met: all samples in the current node belong to the same category; the number of samples is lower than the preset threshold; the decision tree reaches the preset maximum depth; Finally, all decision trees A random forest model is formed. The prediction result of the random forest is the majority vote of the classification results of all decision trees. The model expression is: ,in, is the predicted icing risk level, is an indicator function, which takes the value of 1 when the condition is true, otherwise it is 0. Through the above steps, the model training unit uses the random forest algorithm to train and generate an icing risk prediction model that comprehensively considers terrain characteristics and local meteorological parameters. The model can accurately classify different icing risk levels in the multi-dimensional feature space, providing a reliable basis for subsequent risk prediction.

[0031] The parameter optimization unit includes: Initial parameter setting: Set the initial hyperparameters of the random forest algorithm, including the number of decision trees, the maximum depth, the number of features used by each tree, and the minimum number of node sample splits, as the initial parameter set for cross-validation; Cross-validation grouping: Divide the training sample set into In each iteration, one subset is selected as the validation set, and the rest subsets as training sets, and repeat times, making sure each subset is used as a validation set once; Model training and validation: In each iteration, the current hyperparameter settings are used to train The icing risk assessment model was trained using the training subsets, and the performance indicators of the model, including precision, recall, and F1 score, were calculated using the validation set. Performance Indicator Summary and Evaluation: Summary The verification results of the iterations are used to calculate the average performance index of the icing risk assessment model as a comprehensive evaluation index for the current hyperparameter settings; for example, for the calculation of accuracy: ; Dynamic adjustment of hyperparameters: Based on the average performance index of the validation set, the hyperparameter combination is adjusted, and the optimal hyperparameter combination is determined by grid search or random search method. For hyperparameter combinations with poor validation results, the number of decision trees is adjusted first (increasing decision trees to improve model stability) or the maximum depth of the tree is reduced (to prevent overfitting). For combinations with good validation results, the feature usage ratio of each tree is gradually refined to optimize model performance. Final hyperparameter determination: Select the hyperparameter combination with the best verification results and record the hyperparameter settings as the final parameters of the random forest model; Test sample performance verification: Use the test sample set to conduct final verification of the optimized icing risk assessment model, and calculate the performance indicators of the icing risk assessment model on the test sample set, including accuracy, error rate, and classification accuracy; if the test results do not meet the preset performance targets, the hyperparameters will be adjusted again and the above optimization steps will be repeated until the performance requirements are met; Parameter output: The optimized hyperparameters and the final verified performance indicators are recorded and saved as the configuration information file of the icing risk assessment model to facilitate subsequent model calls and performance tracking.

[0032] The real-time monitoring module includes a data receiving unit, a data preprocessing unit, a feature matching unit and a model calling unit; wherein: Data receiving unit: used to receive real-time meteorological data provided by the meteorological data acquisition module, including temperature, humidity, precipitation and wind speed; Data preprocessing unit: used to preprocess the received real-time meteorological data, including data cleaning, missing value filling and outlier detection, to ensure the integrity of the input data; Feature matching unit: matches the preprocessed real-time meteorological data with the input features of the icing risk assessment model to ensure that the data format and dimension are consistent with the model requirements; Model calling unit: calls the icing risk assessment model constructed in the icing risk assessment module, inputs the matched real-time meteorological data into the model, performs real-time icing risk prediction, and outputs the risk prediction results; through the collaborative work of the above units, the real-time monitoring module can efficiently input real-time meteorological data into the constructed icing risk assessment model, perform real-time risk prediction, and provide accurate risk assessment results for the early warning information generation module.

[0033] The warning information generation module includes a risk classification unit, a thickness analysis unit, a high-risk point positioning unit, an information integration unit and a warning information output unit; wherein: Risk grading unit: used to receive the risk prediction results output by the real-time monitoring module, and compare the ice risk value in the risk prediction results with the preset risk level threshold, and grade them into mild risk, moderate risk and severe risk to generate the risk level of the target area; Thickness analysis unit: Based on the predicted ice thickness value in the risk prediction results, combined with the meteorological conditions and terrain characteristics of the target area, generate ice thickness distribution maps for different areas to clarify the spatial variation of ice thickness; High-risk point positioning unit: By analyzing the risk level distribution and ice thickness prediction results, identify the areas with medium and above ice risk levels in the target area, determine the specific locations of high-risk points, and generate a list of high-risk points with corresponding geographic coordinates marked; Information integration unit: Integrate the data generated by the risk classification unit, thickness analysis unit and high-risk point positioning unit to generate complete warning information for the target area, including risk level distribution map, ice thickness distribution map and high-risk point list; Warning information output unit: outputs the integrated warning information in digital form for query by the power maintenance department and relevant monitoring personnel, and supports the display of warning information in the form of maps, text reports or tables; through the collaborative work of the above units, the warning information generation module can generate warning information including ice thickness prediction values, risk levels and high-risk point locations based on the risk prediction results output by the real-time monitoring module, providing accurate decision support for power line maintenance and emergency response.

[0034] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0035] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A wire ice warning system based on terrain analysis, characterized in that: It includes terrain analysis module, microclimate simulation module, meteorological data acquisition module, icing risk assessment module, real-time monitoring module and warning information generation module; among which: Terrain analysis module: used to obtain digital elevation model data of the target area, and calculate the slope, slope aspect, altitude and terrain roughness of the target area based on the digital elevation model data to form terrain feature data of the target area; Microclimate simulation module: Based on terrain feature data, computational fluid dynamics and numerical weather forecast models are used to simulate the micro-scale climate of the target area and generate local meteorological parameters, including refined wind field, temperature field and humidity field; Meteorological data acquisition module: used to deploy a distributed meteorological sensor array in the target area to collect real-time meteorological data of the target area, including temperature, humidity, precipitation and wind speed; Icing risk assessment module: Based on terrain feature data, local meteorological parameters and historical icing data, an icing risk assessment model is constructed using the random forest algorithm; Real-time monitoring module: receives real-time meteorological data provided by the meteorological data acquisition module, inputs the real-time meteorological data into the constructed icing risk assessment model, performs real-time risk prediction, and outputs the predicted value of ice thickness; Early warning information generation module: Generates early warning information for the target area based on the risk prediction results output by the real-time monitoring module.

2. The wire ice warning system based on terrain analysis according to claim 1 is characterized in that: The terrain analysis module includes a data acquisition unit, a slope calculation unit, a slope aspect calculation unit, an altitude calculation unit, a terrain roughness calculation unit and a feature integration unit; wherein: Data acquisition unit: used to acquire digital elevation model data of the target area through satellite remote sensing data source, the digital elevation model data includes surface elevation information represented in the form of a grid, and perform format conversion and projection coordinate correction on the data to ensure data accuracy; Slope calculation unit: based on the elevation value change of the digital elevation model data, the slope value of each grid unit is calculated by differential operation, and the slope value represents the ratio of the elevation change with the horizontal distance; Aspect calculation unit: based on the digital elevation model data and the slope calculation result, the aspect value of each grid cell is calculated, and the aspect value represents the azimuth of the slope surface relative to the north direction; Elevation calculation unit: extract the elevation value of each grid cell in the digital elevation model data, and directly obtain the elevation distribution of each grid cell in the target area; Terrain roughness calculation unit: by analyzing the degree of change of local elevation in the digital elevation model data, the terrain roughness of each grid cell in the target area is calculated, and the terrain roughness is expressed as the standard deviation of the elevation change around the grid cell; Feature integration unit: used to integrate the calculation results of slope, aspect, altitude and terrain roughness to generate terrain feature data of the target area.

3. The wire ice warning system based on terrain analysis according to claim 1 is characterized in that: The microclimate simulation module includes a grid division unit, a fluid dynamics solving unit, an energy transfer solving unit, a water vapor transport solving unit and an output generation unit; wherein: Meshing unit: used to mesh the target area according to terrain feature data and generate a three-dimensional mesh model for computational fluid dynamics simulation; Fluid dynamics solving unit: Based on the computational fluid dynamics model, the Reynolds-averaged Navier-Stokes equations are used to solve the fluid dynamics and simulate the airflow distribution in the target area. The calculation formula is: ,in, is the gas density, and The airflow is and The velocity component in the direction, For pressure, is the dynamic viscosity, is the volume force; through iterative calculation, the refined wind field distribution of the target area is obtained; Energy transfer solving unit: Based on the energy conservation equation in the numerical weather forecast model, the temperature field in the target area is simulated. The calculation formula is: ,in, is the temperature, is the specific heat capacity, is the thermal conductivity, is the heat source term; Water vapor transport solution unit: Based on the water vapor transport equation, simulate the humidity field distribution in the target area. The calculation formula is: ,in, For humidity, is the water vapor diffusion coefficient, is the water vapor source term; Output generation unit: Integrate the calculation results of the fluid dynamics solution unit, energy transfer solution unit and water vapor transport solution unit to generate local meteorological parameters of the target area and output them in the form of three-dimensional grid.

4. The wire ice warning system based on terrain analysis according to claim 3 is characterized in that: The grid division unit comprises: Data preprocessing: normalize the terrain feature data provided by the terrain analysis module and normalize the elevation values ​​to ensure that the input data is within a uniform scale; Terrain complexity analysis: Calculate the mesh refinement requirements in the target area based on the terrain roughness and slope; Initial grid generation: Based on the boundary data of the target area, the quadtree partitioning algorithm is used to generate an initial regular grid to divide the target area into a number of initial square units; Mesh refinement: Adaptive mesh refinement is used to adjust the initial mesh according to the results of terrain complexity analysis. For areas with greater roughness or higher slopes, the side length of the mesh unit is refined. For areas with gentle terrain changes, the resolution of the initial mesh unit is kept unchanged. Boundary optimization: Optimize the grid cell boundaries and generate unstructured grids using the Delaunay triangulation method to minimize the maximum angle of the triangle; Mesh verification and output: Verify the quality of mesh cells. Once the shape factors of all meshes meet the quality threshold requirements, the optimized mesh data will be output.

5. The wire ice warning system based on terrain analysis according to claim 1 is characterized in that: The meteorological data acquisition module includes a sensor selection unit, a sensor layout unit, a data transmission unit, a sensor calibration unit and a data storage unit; wherein: Sensor selection unit: used to deploy distributed meteorological sensor arrays in the target area, each sensor array includes a temperature sensor, a humidity sensor, a precipitation meter and an anemometer; The sensor layout unit divides the target area based on the terrain feature data provided by the terrain analysis module and determines the layout positions of the sensors according to the following rules; Rule 1: Place high-density sensor arrays in areas with high slopes and high roughness; Rule 2: Deploy low-density sensor arrays in areas with flat terrain; Data transmission unit: used to transmit the collected meteorological data to the system's central data device in real time for standardization; Sensor calibration unit: used to periodically calibrate deployed sensors. Specifically, the collected temperature, humidity, precipitation and wind speed are compared with data from standard weather stations or known reference equipment to adjust the measurement deviation of the sensors to ensure the accuracy of the collected data. Data storage unit: used to store the collected meteorological data in different regions, and the meteorological data of each region is archived according to the timestamp.

6. The wire ice warning system based on terrain analysis according to claim 1 is characterized in that: The icing risk assessment module includes a sample construction unit, a model training unit, a model verification unit, a parameter optimization unit and a model output unit; wherein: Sample construction unit: Based on historical ice cover data, the terrain feature data and local meteorological parameters of the target area are matched with the corresponding ice cover risk level labels to construct training sample sets and test sample sets; the training sample set is used for model training, and the test sample set is used for model evaluation; Model training unit: Use the random forest algorithm to train the training sample set, and classify the data by building multiple decision trees. Each decision tree takes a subset of the feature vector as input and the icing risk level as output. The outputs of all decision trees are voted on comprehensively to form an icing risk prediction model. The model verification unit verifies the trained risk assessment model using the test sample set, and calculates the model's performance indicators, including accuracy, error rate, and classification accuracy, by comparing the model's prediction results with the actual risk level labels; Parameter optimization unit: Use cross-validation method to optimize the hyperparameters of the icing risk assessment model, including the number of decision trees, maximum depth, and number of features used by each tree, to ensure the generalization ability of the model on different sample sets; at the same time, dynamically adjust the hyperparameters according to the performance indicators of the model on the test sample set; Model output unit: used to save the final optimized icing risk assessment model as a standardized model file for real-time monitoring module to call.

7. The wire ice warning system based on terrain analysis according to claim 6 is characterized in that: The model training unit specifically includes: Training sample initialization: receiving the training sample set provided by the sample construction unit, each sample consists of an input feature vector and a corresponding icing risk level label; Decision Tree Construction: Random Forest generates multiple decision trees by repeating the following steps multiple times: Step 1: Randomly select a predetermined number of samples from the training sample set using the bootstrap sampling method to construct a training subset of a decision tree, and the remaining samples are used as the validation set; Step 2: Randomly select some features from all features as the candidate set of split features for the current decision tree. The number of selected features is usually less than the total number of features. Step 3: At each node, select the optimal split feature and split point from the candidate feature set to optimize the purity of the node after splitting; Decision tree training: Recursively split the training subset of each decision tree until any of the set conditions are met, including that all samples in the current node belong to the same category, the number of samples is lower than the preset threshold, and the decision tree reaches the preset maximum depth; Decision tree ensemble: All decision trees are combined into a random forest model. The prediction result of the random forest is the majority vote of the classification results of all decision trees, that is, the classification with the most occurrences is selected as the final prediction result. Model output: After training is completed, the random forest model is saved as an icing risk assessment model.

8. The wire ice warning system based on terrain analysis according to claim 7 is characterized in that: The parameter optimization unit comprises: Initial parameter setting: Set the initial hyperparameters of the random forest algorithm, including the number of decision trees, the maximum depth, the number of features used by each tree, and the minimum number of node sample splits, as the initial parameter set for cross-validation; Cross-validation grouping: Divide the training sample set into In each iteration, one subset is selected as the validation set, and the rest subsets as training sets, and repeat times, making sure each subset is used as a validation set once; Model training and validation: In each iteration, the current hyperparameter settings are used to train The icing risk assessment model was trained using the training subsets, and the performance indicators of the model, including precision, recall, and F1 score, were calculated using the validation set. Performance Indicator Summary and Evaluation: Summary The verification results of the iterations are used to calculate the average performance index of the icing risk assessment model as a comprehensive evaluation index for the current hyperparameter settings; Dynamic adjustment of hyperparameters: Based on the average performance index of the validation set, the hyperparameter combination is adjusted, and the optimal hyperparameter combination is determined by grid search or random search method; Final hyperparameter determination: Select the hyperparameter combination with the best verification results and record the hyperparameter settings as the final parameters of the random forest model; Test sample performance verification: Use the test sample set to conduct final verification of the optimized icing risk assessment model, and calculate the performance indicators of the icing risk assessment model on the test sample set, including accuracy, error rate, and classification accuracy; if the test results do not meet the preset performance targets, the hyperparameters will be adjusted again and the above optimization steps will be repeated until the performance requirements are met; Parameter output: The optimized hyperparameters and the final verified performance indicators are recorded and saved as the configuration information file of the icing risk assessment model.

9. The wire ice warning system based on terrain analysis according to claim 1 is characterized in that: The real-time monitoring module includes a data receiving unit, a data preprocessing unit, a feature matching unit and a model calling unit; wherein: Data receiving unit: used to receive real-time meteorological data provided by the meteorological data acquisition module, including temperature, humidity, precipitation and wind speed; Data preprocessing unit: used to preprocess the received real-time meteorological data, including data cleaning, missing value filling and outlier detection, to ensure the integrity of the input data; Feature matching unit: matches the preprocessed real-time meteorological data with the input features of the icing risk assessment model to ensure that the data format and dimension are consistent with the model requirements; Model calling unit: calls the icing risk assessment model constructed in the icing risk assessment module, inputs the matched real-time meteorological data into the model, performs real-time icing risk prediction, and outputs the risk prediction results.

10. The wire ice warning system based on terrain analysis according to claim 1, characterized in that: The warning information generation module includes a risk classification unit, a thickness analysis unit, a high-risk point positioning unit, an information integration unit and a warning information output unit; wherein: Risk grading unit: used to receive the risk prediction results output by the real-time monitoring module, and compare the ice risk value in the risk prediction results with the preset risk level threshold, and grade them into mild risk, moderate risk and severe risk to generate the risk level of the target area; Thickness analysis unit: Based on the predicted ice thickness value in the risk prediction results, combined with the meteorological conditions and terrain characteristics of the target area, generate ice thickness distribution maps for different areas to clarify the spatial variation of ice thickness; High-risk point positioning unit: By analyzing the risk level distribution and ice thickness prediction results, identify the areas with medium and above ice risk levels in the target area, determine the specific locations of high-risk points, and generate a list of high-risk points with corresponding geographic coordinates marked; Information integration unit: Integrate the data generated by the risk classification unit, thickness analysis unit and high-risk point positioning unit to generate complete warning information for the target area, including risk level distribution map, ice thickness distribution map and high-risk point list; Warning information output unit: outputs the integrated warning information in digital form for query by power maintenance departments and relevant monitoring personnel, and supports displaying warning information in the form of maps, text reports or tables.

Images