A power cut risk prediction method and system based on micro-meteorological evolution trajectory

Abstract

This application discloses a method and system for predicting power outage risks based on micro-meteorological evolution trajectories, relating to the field of power system safety operation and prevention. The method includes: acquiring micro-meteorological monitoring data, micro-topographic data, and equipment characteristic data along transmission lines; constructing a spatiotemporally aligned multidimensional basic dataset; reconstructing the spatiotemporal evolution trajectory of continuously distributed micro-meteorological factors; extracting micro-meteorological modal components at different time scales and calculating the meteorological-equipment coupling factor in conjunction with equipment characteristics; based on the instantaneous hazard intensity of each micro-topographic unit, using a fractional integral operator to accumulate historical memories, generating a comprehensive risk index, thereby determining the power outage risk level and outputting physical disaster attribution early warning information. This improves the accuracy and interpretability of risk prediction.

Description

Technical Field

[0001] This application relates to the field of power system safety operation and prevention technology, and in particular to a method and system for predicting power outage risks based on micro-meteorological evolution trajectories. Background Technology

[0002] When power systems operate in complex micro-topographical areas such as mountainous regions, canyons, and coastal areas, they are highly susceptible to natural environmental factors such as wind, rain, and freezing, which can lead to line swaying, icing, and flashover, causing tripping or damage. These micro-topographical features often have a coupling effect with micro-meteorological conditions, such as the formation of localized strong winds, making the actual micro-meteorological conditions far more severe than those recorded by surrounding conventional weather stations.

[0003] However, existing power outage risk assessment methods mainly rely on traditional macro-meteorological station data, focusing primarily on extreme weather events such as typhoons and rainstorms. These methods are not very sensitive when dealing with complex terrain areas and struggle to effectively capture and characterize the gradual and cumulative risks posed to transmission lines by microclimate evolution under non-extreme weather conditions. Furthermore, existing microclimate early warning studies are mostly limited to static classifications based on single-point probabilities, lacking in-depth analysis of the spatial transmission and temporal evolution trajectories of microclimates along the transmission lines.

[0004] In addition, most of the data-driven prediction models currently introduced in the industry use "black box" algorithms such as deep neural networks. Although these models are capable of processing high-dimensional data, they seriously lack physical interpretability and cannot clearly give the specific mechanisms and risk attributions that cause the faults. This makes it difficult for power grid dispatchers and maintenance personnel to formulate targeted proactive defense and power supply service strategies based on the prediction results. Summary of the Invention

[0005] This application provides a method, system, storage medium, computer program product, and electronic device for predicting power outage risks based on micro-meteorological evolution trajectories, in order to at least solve the problem of insufficient identification and early warning capabilities for transmission line operation risks under complex terrain and meteorological conditions in the prior art.

[0006] In a first aspect, embodiments of this application provide a method for predicting power outage risk based on micro-meteorological evolution trajectories. The method includes: acquiring micro-meteorological monitoring data, micro-topographic data, and equipment characteristic data of the transmission line along its route; dividing the micro-topographic data into multiple micro-topographic units along the line; performing spatiotemporal alignment of the micro-meteorological monitoring data and the micro-topographic data using a time synchronization signal to construct a multidimensional basic dataset; reconstructing the spatiotemporal evolution trajectory of continuously distributed micro-meteorological factors along the transmission line based on the multidimensional basic dataset; the micro-meteorological factors including at least one of the following: wind speed, wind direction, temperature, humidity, air pressure, and rainfall; extracting micro-meteorological modal components from the spatiotemporal evolution trajectory, and aligning the micro-meteorological modal components with... The equipment characteristic data are matched and analyzed to calculate the coupling factor between the micro-meteorological factors and the equipment of the transmission line; wherein, the micro-meteorological modal components characterize the oscillation components of the micro-meteorological factors along the spatiotemporal evolution trajectory at different time scales; the instantaneous hazard intensity of the micro-meteorological factors in the spatiotemporal evolution trajectory on each micro-topographic unit is determined, and a fractional integral operator is used to perform a nonlinear accumulation operation based on a historical time window on the instantaneous hazard intensity to generate a comprehensive risk index with historical memory effect for each micro-topographic unit; based on the comprehensive risk index of each micro-topographic unit and the coupling factor, the power outage risk level of the transmission line is determined, and early warning information including the physical disaster attribution of the risk is generated.

[0007] Secondly, embodiments of this application provide a power outage risk prediction system based on micro-meteorological evolution trajectories. The system includes: a data acquisition unit for acquiring micro-meteorological monitoring data, micro-topographic data, and equipment characteristic data of the transmission line along the line, and dividing the micro-topographic data into multiple micro-topographic units along the line; a spatiotemporal alignment unit for performing spatiotemporal alignment of the micro-meteorological monitoring data and the micro-topographic data using a time synchronization signal to construct a multi-dimensional basic dataset; a trajectory reconstruction unit for reconstructing the spatiotemporal evolution trajectory of continuously distributed micro-meteorological factors along the transmission line based on the multi-dimensional basic dataset; the micro-meteorological factors include at least one of the following: wind speed, wind direction, temperature, humidity, air pressure, and rainfall; and a modal analysis unit for extracting micro-meteorological modal components from the spatiotemporal evolution trajectory and... The micro-meteorological modal components are matched and analyzed with the equipment characteristic data to calculate the coupling factor between the micro-meteorological factors and the equipment of the transmission line; wherein, the micro-meteorological modal components characterize the oscillation components of the micro-meteorological factors along the spatiotemporal evolution trajectory at different time scales; a risk accumulation unit is used to determine the instantaneous hazard intensity of the micro-meteorological factors in the spatiotemporal evolution trajectory on each of the micro-topographic units, and to perform a nonlinear accumulation operation based on a historical time window on the instantaneous hazard intensity using a fractional integral operator to generate a comprehensive risk index for each of the micro-topographic units with a historical memory effect; an early warning output unit is used to determine the power outage risk level of the transmission line based on the comprehensive risk index of each of the micro-topographic units and the coupling factor, and to generate early warning information including the physical causes of the risk disaster.

[0008] Thirdly, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the steps of the power outage risk prediction method based on micro-meteorological evolution trajectory according to any embodiment of the present application.

[0009] Fourthly, embodiments of this application provide a storage medium storing a computer program thereon, characterized in that, when the program is executed by a processor, it implements the steps of the power outage risk prediction method based on micro-meteorological evolution trajectory according to any embodiment of this application.

[0010] Fifthly, embodiments of this application provide a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps of the power outage risk prediction method based on micro-meteorological evolution trajectory according to any embodiment of this application.

[0011] The power outage risk prediction method and system based on micro-meteorological evolution trajectory provided in this application can achieve at least the following technical effects: (1) Instead of treating the micro-meteorological conditions along the transmission line as a series of isolated discrete observations, this paper combines micro-topographic unit division, spatiotemporal alignment of multi-source data, and continuous field reconstruction mechanisms to accurately construct the spatiotemporal evolution trajectory of micro-meteorological factors continuously distributed along the transmission line, and further extracts micro-meteorological modal components at different time scales. Based on this technical concept, the meteorological propagation trend and stage fluctuation characteristics in local complex micro-topography (such as canyon funnels or steep slopes) along the line can be continuously characterized, thereby transforming the assessment object of power outage risk from traditional single-point static macro-meteorological values ​​to a dynamic micro-physical process representation with high-resolution spatial extensibility and temporal evolution. As a result, it can more fully reflect the formation and evolution law of micro-meteorological disturbances under complex micro-topographic conditions, and establish a state expression basis that is more in line with the actual operation scenario for subsequent risk identification.

[0012] (2) Deep matching analysis is performed between micro-meteorological modal components and equipment characteristic data of transmission lines to construct coupling factors between micro-meteorological factors and line equipment. At the same time, combined with the instantaneous hazard intensity on each micro-topographic unit, a fractional integral operator is introduced to perform nonlinear accumulation of hazard intensity based on historical time windows, generating a comprehensive risk index with historical memory effect. Based on this technology chain, it is possible not only to keenly capture the instantaneous excitation or resonance response of current micro-meteorological high-frequency disturbances on line equipment (characterized by coupling factors), but also to accurately quantify the nonlinear cumulative damage to equipment caused by continuous and repetitive severe environmental effects (characterized by comprehensive risk index). Furthermore, since the determination of power outage risk level is subject to the combined dual constraints of comprehensive risk index and coupling factors, the final generated early warning information can directly output clear risk physical disaster attribution (such as icing overload or wind frequency resonance), thereby enhancing the interpretability and attribution ability of risk judgment results.

[0013] This technical solution constructs a closed-loop system based on the physical formation mechanism of micro-meteorological risks along transmission lines. It begins with spatiotemporal evolution trajectory reconstruction and multi-scale modal feature extraction, proceeds through coupled analysis of micro-meteorology and equipment frequency domains, and culminates in historical memory-based nonlinear risk accumulation and physical disaster attribution early warning. Employing a white-box closed-loop mechanism, power outage risks are no longer crudely characterized by a single probabilistic outcome indicator, but are comprehensively identified and hierarchically expressed within a unified physical framework of continuous evolution, coupled response, and spatiotemporal accumulation. This elevates power outage risk prediction from static description to dynamic, mechanistic characterization, and improves the accuracy, interpretability, and support for power grid dispatching and maintenance decisions in early warning assessments. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 A flowchart illustrating an example of a power outage risk prediction method based on micrometeorological evolution trajectories according to an embodiment of this application is shown; Figure 2 A flowchart illustrating an example of generating early warning information containing attribution of physical hazards according to an embodiment of this application is shown. Figure 3 A schematic diagram illustrating the system operation mechanism of an example of a power outage risk prediction method based on micrometeorological evolution trajectory according to an embodiment of this application is shown. Figure 4 A comparative experimental simulation diagram showing the results of different methods in the spatiotemporal risk thermodynamic evolution is presented. Figure 5 A comparative simulation diagram illustrating the precision-recall (PR) performance and overall Pareto front of different methods in a power outage risk prediction scenario is shown. Figure 6 A structural block diagram of an example of a power outage risk prediction system based on micrometeorological evolution trajectory according to an embodiment of this application is shown. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0017] It should be noted that in constructing power outage risk prediction systems, some scholars in related technologies are currently attempting to propose comprehensive weather-driven risk analysis frameworks that integrate and assess weather hazard indicators, grid topological vulnerability, and economic losses. These frameworks typically combine multivariate data such as numerical weather prediction, soil properties, and vegetation distribution, using algorithms like regression trees or extreme gradient boosters to model the probability of power outages caused by specific weather events such as storms and lightning. However, the hydrological and meteorological data relied upon by these assessment projects are often at a relatively coarse spatial resolution, typically around the kilometer level. While this coarse-grained data grid is applicable in plains areas, it severely masks the effects of steep valleys or narrow passes on local airflow, such as accumulation, acceleration, or deflection, when introduced into complex micro-topographical environments like mountains and canyons. This results in current models losing their ability to capture the fine-grained evolution of non-hazardous micro-meteorological conditions at the lowest data input stage.

[0018] In terms of data processing and fault mapping logic, some studies have introduced neural network models optimized by genetic algorithms to attempt to directly correlate meteorological indicators such as extreme precipitation and consecutive extreme temperatures with historical fault records of power lines. In this process, to cope with the vast amount of micrometeorological and fault samples, researchers often use weighted cost functions to address class imbalance in fault samples, or specifically optimize for statistical anomalies such as zero-inflation distribution and heteroscedasticity present in the samples, and perform spatial overlay analysis to classify risk levels. Although these data optimization methods have improved the statistical indicators of classification prediction to some extent, their underlying logic still treats environmental factors as isolated, discrete, static feature vectors for fitting.

[0019] Furthermore, even though some studies have begun to attempt to extract microclimate coupling coefficients using models such as logistic regression to quantitatively assess the contribution of specific micrometeorological factors to tripping probabilities, this attribution method based on pure statistics remains limited to independent solutions for single nodes or local areas. Currently, related technologies generally lack mathematical reconstruction of the continuous spatiotemporal evolution trajectories of multiple micrometeorological factors transmitted along long-distance transmission lines. Because these models sever the temporal evolution sequence of meteorological conditions from their spatial propagation paths, they fundamentally cannot accurately quantify the dynamic resonant shock waves between meteorological environmental changes and transmission physical equipment (such as conductor self-vibration and insulator moisture), nor can they accurately calculate the nonlinear cumulative damage boundaries with historical memory effects (such as continuous stress fatigue caused by long-term temperature and humidity accumulation). This lack of analysis of spatial continuity and temporal cumulative mechanisms ultimately makes it difficult to identify and accurately locate risk sources with localized multi-point coupling in complex terrain.

[0020] It should be understood that the above description of the relevant technologies is intended only to help the public better understand the inventive spirit and motivation of this application, and is not intended to limit this application. Furthermore, the technical solutions described in the above-mentioned relevant technologies are not prior art, and may also be undisclosed technical solutions, such as those under research or in the laboratory stage.

[0021] The technical solutions in this application, including the collection, storage, use, processing, transmission, provision, and disclosure of users' personal information, comply with relevant laws and regulations and do not violate public order and good morals.

[0022] Figure 1 A flowchart illustrating an example of a power outage risk prediction method based on micrometeorological evolution trajectories according to an embodiment of this application is shown.

[0023] Regarding the execution subject of the method in the embodiments of this application, it can be any controller or processor with computing or processing capabilities, such as a power grid dispatch analysis platform controller, a power transmission operation and maintenance management platform controller, or an edge computing platform controller. By calling computer program instructions pre-deployed in the memory, it performs spatiotemporal alignment, evolution trajectory reconstruction, modal component extraction, coupling analysis, risk accumulation calculation, and early warning information generation on the acquired micro-meteorological monitoring data, micro-topographic data, and power transmission line equipment characteristic data.

[0024] In some examples, it can be integrated into electronic devices or terminals through software, hardware, or a combination of both, and the types of terminals or electronic devices can be diverse, such as servers, industrial control computers, edge computing gateways, online monitoring terminals for power transmission lines, or other devices with data processing and analysis capabilities.

[0025] like Figure 1 As shown, in step S110, micro-meteorological monitoring data, micro-topographic data, and equipment characteristic data of the transmission line are acquired, and the micro-topographic data is divided into multiple micro-topographic units along the line.

[0026] Specifically, in the actual operating environment of complex power grids, long-distance transmission lines often cross special terrains such as towering mountains, canyons, and coastal areas. In the steps of this embodiment, firstly, various IoT sensor terminals deployed on transmission towers or line corridors collect real-time micro-meteorological monitoring data such as wind speed, wind direction, temperature, humidity, air pressure, and rainfall; simultaneously, micro-topographic data including local elevation, slope, canyon index, and other terrain features are obtained through a high-precision digital elevation model, and equipment characteristic data such as conductor cross-sectional area, insulator material, tower mechanical structure, and its inherent natural frequency are retrieved from the power grid asset and equipment management system.

[0027] Furthermore, to achieve refined grid management of transmission lines, in this embodiment, the geographical space where the entire continuous transmission line is located is divided into multiple micro-topographic units of appropriate scale based on the rate of change of geographical features such as terrain undulation, canyon orientation, or elevation difference. This overcomes the problems of insufficient spatial resolution and local terrain blind spots caused by traditional reliance on macro-meteorological stations, transforming the complex physical and geographical environment into a discrete spatial carrier that can be quantitatively analyzed point-by-point by computer, providing a solid spatial data foundation for subsequent refined risk location.

[0028] In step S120, the micro-meteorological monitoring data and micro-topographic data are spatiotemporally aligned by combining the time synchronization signal in order to construct a multi-dimensional basic dataset.

[0029] It should be noted that in the actual data collection process of the ubiquitous power Internet of Things, the sampling frequencies of various micro-meteorological sensors are often inconsistent, and there is an unavoidable network delay during data transmission, resulting in the acquired raw observation data being scattered along the time axis. Simultaneously, the dynamically changing meteorological flow field and the statically distributed micro-topographic units are also prone to spatial misalignment. Therefore, in this embodiment, a unified time synchronization signal is introduced as a global clock reference to perform time-series interpolation and resampling on the micro-meteorological monitoring data with discrete timestamps, ensuring strict synchronization along the time axis.

[0030] Subsequently, the time-synchronized micro-meteorological data and corresponding micro-topographic units were precisely matched and anchored in three-dimensional spatial coordinates, thereby constructing a multi-dimensional basic dataset that contains both a dynamic temporal evolution sequence and spatial three-dimensional distribution characteristics. This eliminates the misalignment errors in the spatiotemporal dimensions of multi-source heterogeneous meteorological and geographical data, providing an absolutely rigorous data input model for accurately capturing the flow patterns of micro-meteorological factors under complex terrain.

[0031] In step S130, based on the multidimensional basic dataset, the spatiotemporal evolution trajectory of micrometeorological factors continuously distributed along the transmission line is reconstructed. Here, micrometeorological factors include at least one of the following: wind speed, wind direction, temperature, humidity, air pressure, and rainfall.

[0032] Since the deployment of meteorological sensors is limited by the specific physical location of the towers, the aforementioned multidimensional basic dataset is essentially still a spatially discrete set of observation points. However, in micro-topographical areas such as mountainous regions or canyons, changes in airflow or temperature and humidity are not linear but are strongly constrained by the terrain, resulting in acceleration, swirling, or accumulation effects. Therefore, in the steps of this embodiment, mathematical methods such as nonlinear interpolation mapping or state estimation are used to fill the spatial meteorological information gaps between discrete monitoring nodes, and to extrapolate and reconstruct the discrete point data into a dynamic meteorological field that continuously spreads along the physical orientation of the transmission line.

[0033] In practice, this reconstruction process reproduces the dynamic physical process of the spread and evolution of meteorological disturbances such as local strong winds or cold waves along the topography of the power transmission corridor in the virtual data space. This overcomes the blind spot of "single-point isolated prediction" in traditional assessment methods, reproducing the "evolution and transmission trajectory" of micro-meteorological disasters along the power grid topology in virtual space, greatly enhancing the ability to continuously perceive the meteorological flow field in complex terrain areas.

[0034] In step S140, micrometeorological modal components are extracted from the spatiotemporal evolution trajectory, and these components are matched with equipment characteristic data to calculate the coupling factor between the micrometeorological factors and the equipment of the transmission line. Here, the micrometeorological modal components characterize the oscillation components of the micrometeorological factors at different time scales along the spatiotemporal evolution trajectory.

[0035] Here, signal mode decomposition is used to extract independent oscillating components (i.e., micro-meteorological mode components) representing different time scales from the complex evolution trajectory. For example, high-frequency wind speed fluctuation components are independently extracted for subsequent specialized assessments of their dynamic wind deflection impact on transmission lines. Based on this, the extracted high-frequency micro-meteorological mode components with specific oscillation frequencies are deeply frequency-domain matched with equipment characteristic data, including the inherent natural frequencies of the equipment and material sensitivity.

[0036] For example, when the instantaneous excitation frequency of high-frequency micro-meteorological disturbances approaches the mechanical natural frequency of power transmission equipment, the two will physically generate a strong resonance response, thereby calculating a coupling factor used to quantitatively characterize the intensity of this physical resonance response. This achieves a leap from "pure meteorological data statistical analysis" to "meteorological dynamics-power grid mechanical structure frequency domain coupling analysis," enabling precise capture of hidden equipment damage risks caused by micro-wind vibration, wind deflection, etc., under non-extreme weather conditions.

[0037] In step S150, the instantaneous hazard intensity of micro-meteorological factors in the spatiotemporal evolution trajectory on each micro-topographic unit is determined, and a nonlinear accumulation operation based on historical time windows is performed on the instantaneous hazard intensity using a fractional integral operator to generate a comprehensive risk index for each micro-topographic unit with historical memory effect.

[0038] It should be noted that when assessing the damage mechanism of meteorological environment on power transmission networks, in addition to instantaneous strong impacts, the cumulative effect of long-term adverse environment (such as severe icing caused by continuous high humidity and low temperature) is also a fatal factor leading to tower collapse and line breakage. Accordingly, in the steps of this embodiment, the instantaneous hazard intensity caused by micro-meteorological factors (such as current wind pressure intensity and current humidity saturation) on the line on a specific micro-topographic unit is first determined based on a physical model. Subsequently, breaking through the limitations of traditional integer-order integrals or simple linear summation, a fractional-order integral operator that can accurately express the historical memory effect of system state is introduced.

[0039] Specifically, this integral operation not only quantifies the risk at the current moment but also backtracks to a set historical time window, nonlinearly accumulating historical meteorological hazards with time decay weights into the current state. For example, even if several consecutive days of low temperature and high humidity do not reach the extreme value of destruction on a single day, the cumulative calculated ice thickness or the probability of flashover due to moisture on the insulator surface will show a nonlinear surge consistent with physical reality under the fractional integral operation. This greatly improves the sensitivity of the prediction model to capturing "gradual and cumulative" power grid disasters (such as icing overload and long-term stress fatigue), eliminating the underreporting phenomenon caused by "mild but persistent" micrometeorological conditions.

[0040] In step S160, based on the comprehensive risk index and coupling factor of each micro-topographic unit, the power outage risk level of the transmission line is determined, and early warning information containing the physical causes of the risk is generated.

[0041] Here, the system performs a multi-dimensional joint assessment of the "coupling factor" reflecting high-frequency dynamic impact calculated in step S140 and the "comprehensive risk index" reflecting long-term low-frequency cumulative damage calculated in step S150. By comparing these two types of indicators, which respectively characterize different physical damage mechanisms, with the system's preset equipment safety critical threshold, the system can accurately classify the power outage risk of each micro-topographic unit on the transmission line corridor (e.g., determine it as a safe level or a dangerous level).

[0042] In particular, based on which specific dimensional indicator triggered the danger threshold, the system can automatically identify the physical root cause of the high risk. For example, if the comprehensive risk indicator severely exceeds the standard, the risk is attributed to icing overload or continuous stress fatigue; if the coupling factor experiences a nonlinear jump exceeding the standard, it is directly attributed to wind deflection frequency resonance or conductor dynamic galloping. Thus, it not only clarifies "where the danger is and how great the danger level is" in the power grid space, but also completely breaks through the drawbacks of the "black box" early warning of traditional data-driven deep learning models by providing highly transparent physical disaster attribution. This provides highly interpretable and targeted decision support for dispatch centers and frontline maintenance personnel to take differentiated wind and vibration reduction or DC de-icing strategies in advance.

[0043] Regarding the operational details of constructing a multidimensional basic dataset through spatiotemporal alignment in step S120, in some examples of embodiments of this application, firstly, the network time protocol timestamp is used as a unified time synchronization signal to extract the time series labels of the micro-meteorological monitoring data collected by each sensor node, and spline interpolation algorithm is used to fill in the missing time series values ​​in the micro-meteorological monitoring data to generate time-aligned meteorological data.

[0044] Specifically, in the actual operation of the ubiquitous power Internet of Things (IoT), the raw observation data uploaded by micro-meteorological sensor nodes distributed along the power line often suffers from asynchronous timestamps and data frame loss (missing data) due to independent clock drift, network transmission congestion, or intermittent hardware hibernation. Therefore, this embodiment uses Network Time Protocol (NTP) or satellite timing signals as a globally unified clock reference to perform time-series calibration on all asynchronous data frames. For identified missing time-series values, this scheme abandons simple linear interpolation, which easily leads to discontinuities in physical derivatives, and instead employs a cubic spline interpolation algorithm for nonlinear reconstruction. For missing time-series intervals... Construct a piecewise smooth cubic polynomial function:

[0045] Equation (1) In the formula, Indicates time The fitted meteorological values, These are the polynomial coefficients obtained based on the boundary conditions of adjacent valid observation nodes. Calculating the missing time values ​​using this polynomial function ensures that the fitted micro-meteorological time series has continuous first and second derivatives at the microscale. This not only achieves strict temporal alignment of multi-source asynchronous data across the entire network but also guarantees that the rates of change of meteorological factors such as wind speed and temperature (e.g., wind acceleration, temperature gradient) conform to the smooth and gradual changes in fluid dynamics and thermodynamics in real nature, avoiding spurious high-frequency disturbances caused by data abrupt changes in subsequent spatiotemporal evolution analysis.

[0046] Then, the digital elevation model and digital surface model features of each micro-topographic unit are extracted, and the local slope and canyon index are calculated to construct a three-dimensional spatial topological mesh for each micro-topographic unit based on the local slope and canyon index.

[0047] In constructing spatial topology, the Digital Elevation Model (DEM) reflects the basic undulations of the bare surface, while the Digital Surface Model (DSM) further encompasses the aerodynamic roughness caused by vegetation canopy, buildings, etc. This embodiment extracts features from both models and calculates core parameters characterizing the strength of airflow constraint on micro-topography: local slope (characterizing airflow uplift resistance and windward effect) and canyon index (e.g., the ratio of valley depth to valley width, characterizing the intensity of the wind tunnel narrowing effect).

[0048] Based on the aforementioned physical parameters, the system performs three-dimensional spatial subdivision and meshing modeling of the micro-topographic units. In areas with steep slopes or high canyon indices and severe airflow distortion, the system adaptively increases the node density of the three-dimensional spatial topological mesh; while in gentler areas, the mesh density is appropriately reduced. This establishes an aerodynamic boundary model for the power transmission corridor that closely matches the real-world terrain features, allowing the mesh density to naturally match the intensity of spatial disturbances in the micro-meteorological field, effectively balancing computational efficiency with subsequent reconstruction accuracy.

[0049] Furthermore, a spatial resampling algorithm based on Delaunay triangulation is adopted to map time-aligned meteorological data onto the grid nodes of the corresponding micro-topographic unit's three-dimensional spatial topology grid, so as to construct a multi-dimensional basic dataset that is strictly aligned in both time and spatial dimensions.

[0050] Since the actual deployed micro-meteorological sensor nodes are discrete and irregularly distributed in spatial coordinates, the standard three-dimensional spatial topology mesh constructed above cannot be directly applied. Therefore, this embodiment employs the Delaunay triangulation algorithm to construct an optimal spatial triangular mesh (or tetrahedral mesh) that avoids elongated distortion, using the discrete sensor nodes as vertices. Subsequently, for any target mesh node on the three-dimensional spatial topology mesh... The Delaunay subdivision cell to which the sensor belongs is located, and the observations of the sensor node are mapped to the target mesh node using barycentric interpolation. The mapping calculation formula is as follows:

[0051] Equation (2) In the formula, Represents the target mesh node At any moment Resampled meteorological data; The total number of vertex sensing nodes that form the Delaunay simplex (such as a triangle or tetrahedron) surrounding the grid node; Indicates the first Each vertex sensor node at time... Time-series aligned meteorological data; For target grid nodes Compared to the first Spatial geometric barycentric weight coordinates of vertices, and satisfying Therefore, by leveraging the geometric advantage of Delaunay partitioning to maximize the minimum angle, weight distortion in the spatial interpolation process is avoided. The originally disordered and discrete point-like sensor data is smoothly and rigorously resampled into a structured three-dimensional mesh, thus completing the fusion of heterogeneous data at the computer's underlying layer.

[0052] Through the embodiments of this application, a closed-loop underlying data governance system was established, encompassing one-dimensional high-order smooth fitting in the time domain, three-dimensional adaptive grid partitioning of physical topography, and centroid coordinate spatial mapping. This system enables the continuous, structured, and physically consistent reconstruction of original discrete observation data. Consequently, observational information with disordered distribution and local discontinuities can be transformed into a four-dimensional data tensor that is strictly aligned in both the time and spatial dimensions and conforms to the local geomorphological aerodynamic constraints. This provides standardized, high-fidelity, and physically meaningful foundational data support for subsequent micro-meteorological evolution analysis and risk prediction.

[0053] Regarding the implementation details of reconstructing the spatiotemporal evolution trajectory in step S130, in some examples of the embodiments of this application, for any... Each micro-meteorological factor is used to extract data sequences from the grid nodes of a three-dimensional spatial topological grid from a multi-dimensional basic dataset to construct a state vector. By combining local slope and canyon index as micro-topographical environmental constraints, nonlinear state transition equations and observation equations are established: Equation (3) Equation (4) In the formula, To characterize the nonlinear transfer function of the dynamic evolution of micrometeorological states under topographic constraints, For nonlinear observation functions, For process noise, To observe the noise, The sampling time interval, For a moment The observation vector; Indicates time The state vector; Indicates time The state vector.

[0054] In practical implementation, due to the highly complex evolution of microclimates in mountainous or canyon environments, which is far from a simple linear translation, the state-space equations constructed in this embodiment deeply integrate the physical constraints of the geographical environment at the underlying logic. Among them, the state vector... Represents the actual micro-meteorological conditions at the grid nodes (such as the actual physical wind field or temperature field); nonlinear transfer function Then, the "local slope" and "canyon index" are substituted as boundary constraint parameters into the fluid or thermodynamic evolution model. For example, in areas with a large canyon index, This will naturally characterize a stronger acceleration effect in the narrow channel. Nonlinear observation function This establishes a connection between the actual physical state and the actual sensor observations. The mapping relationship between them. Process noise. Used to absorb physical uncertainties such as unmodeled stochastic meteorological turbulence, while observation noise This encompasses the inherent measurement errors of the sensor. Therefore, by establishing a rigorous nonlinear state-space model, pure data monitoring is elevated to a dynamic system expression that conforms to the laws of fluid mechanics and environmental physics, thus establishing a mathematical framework for subsequent noise removal and extraction of true meteorological evolution patterns.

[0055] Then, for the nonlinear state transition equation and observation equation, the unscented Kalman filter algorithm is used to construct a Sigma point set to pair the state vector. Lossless propagation of nonlinear probability distributions is performed to estimate the posterior mean and covariance at each time step, thereby obtaining a discrete meteorological field distribution with temporal smoothness characteristics.

[0056] When dealing with highly nonlinear meteorological state transitions, the first-order Taylor expansion used in traditional Extended Kalman Filter (EKF) introduces a significant truncation error. Therefore, this embodiment employs Unscented Kalman Filter (UKF), which directly approximates the probability distribution of the state vector rather than approximating the nonlinear function itself. Specifically, let the state vector... The dimension is The system is based on its prior mean. Covariance Matrix Deterministically construct Sigma points:

[0057] , , Equation (5) In the formula, Describes the constructed Sigma point set. To fine-tune the scaling parameters for controlling the distribution range of Sigma points, The first square root of a matrix represents the square root of the matrix. List.

[0058] Subsequently, these Sigma point sets were directly substituted into the nonlinear equations. and The algorithm performs lossless nonlinear transfer and recalculates the posterior mean and covariance based on the transferred point set using weighted recalculation. Thus, UKF achieves at least second- or even third-order Taylor series approximation accuracy, solving the filtering divergence problem caused by nonlinear distortion of meteorological variables under complex micro-topographic constraints. Through this algorithm, the system successfully filters out high-frequency measurement noise and random turbulence interference from sensors, outputting a highly time-smooth discrete meteorological field distribution that closely approximates the actual physical state.

[0059] Furthermore, a manifold learning algorithm based on Laplace eigenmaps is adopted to map the discrete meteorological field distribution to a low-dimensional embedding space to extract the meteorological feature flow axis that fits the topographic orientation characterized by local slope and canyon index. Based on the meteorological feature flow axis, spatial continuous interpolation is performed on the three-dimensional spatial topological grid of each micro-topographic unit to reconstruct the spatiotemporal evolution trajectory of the micro-meteorological factors continuously distributed along the transmission line.

[0060] It should be noted that, despite the UKF time-series smoothing, the meteorological field is still spatially discrete high-dimensional data distributed across various grid nodes. Due to the strict constraints imposed by landforms such as mountains and valleys on the spatial diffusion of airflow, temperature, and humidity, the high-dimensional meteorological data actually resides in a low-dimensional manifold space dominated by topography.

[0061] To address this, this embodiment introduces Laplacian eigenmaps. First, an adjacency weight matrix is ​​constructed based on the spatial connectivity of the grid nodes. Degree matrix And calculate the graph Laplacian matrix. Subsequently, the generalized eigenvalue problem is solved:

[0062] Equation (6) In the formula, For feature vectors, These are the corresponding eigenvalues. After discarding zero eigenvalues, the eigenvectors corresponding to the smallest number of non-zero eigenvalues ​​are selected. This allows for the dimensionality reduction and mapping of discrete meteorological fields to a low-dimensional embedding space. These feature vectors, in a physical sense, represent the "meteorological feature flow axis" that conforms to the local terrain orientation (such as the direction of canyon extension). In the final spatial interpolation reconstruction, the system no longer uses isotropic Euclidean distance interpolation (which would lead to erroneous meteorological smoothing across ridges), but instead strictly performs anisotropic interpolation along the extracted "meteorological feature flow axis." Thus, manifold learning reveals the meteorological spatial topology hidden beneath the complex appearance, ensuring that the reconstructed continuous spatiotemporal evolution trajectory absolutely obeys the physical boundary constraints of the underlying micro-topography.

[0063] This application's embodiments address the strong nonlinearity and spatial topological complexity of micrometeorological data by constructing a refined processing algorithm system that integrates state-space equation constraints, unscented transformation temporal smoothing, and Laplace manifold spatial reconstruction. This technical solution effectively mitigates the shortcomings in representing topographic physical features and the biases in nonlinear evolution representation caused by traditional static spatial interpolation methods. Based on this system, the system can suppress multi-source heterogeneous noise interference while performing constrained dimensionality reduction and continuous reconstruction of complex high-dimensional meteorological fields distorted by topography, thereby obtaining a spatiotemporal evolution trajectory consistent with actual aerodynamic boundary characteristics. This provides high-fidelity basic evolutionary data support with temporal continuity and spatial topographic adaptability for subsequent extraction of equipment resonance factors and nonlinear cumulative damage indicators, thus enhancing the model's ability to identify and reconstruct locally hidden micrometeorological hazard sources.

[0064] Regarding the implementation details of extracting micro-meteorological modal components from the spatiotemporal evolution trajectory in step S140, in some examples of embodiments of this application, firstly, for the micro-meteorological factor function in the spatiotemporal evolution trajectory... A Gaussian white noise sequence with a defined standard deviation is introduced and ensemble averaging is performed. The ensemble empirical mode decomposition algorithm is then used to analyze the micrometeorological factor function. Multi-scale signal decomposition is performed to eliminate mode aliasing of complex meteorological signals, and micro-meteorological factor functions are decomposed. Decomposed into multiple eigenmode functions and residual terms of different frequencies: Equation (7) In the formula, Represents the spatial topological coordinates of the transmission line. Indicates the first Micrometeorological factors in spatial topological coordinates and time coordinates Microclimate factor function; The total number of layers of the intrinsic mode functions obtained from the decomposition. This represents the residual trend term reflecting the long-term evolution of meteorological patterns. Indicates the first Each intrinsic mode function characterizes the oscillation component of micrometeorological factors along their spatiotemporal evolution trajectory at different time scales, serving as the corresponding micrometeorological mode components.

[0065] In practice, the real micro-topographical meteorological environment is extremely complex, often containing intermittent high-frequency disturbances such as sudden gusts and turbulence. If traditional empirical mode decomposition (EMD) is used, such intermittent signals are very likely to cause "mode aliasing", that is, frequencies of different physical scales are mixed into the same mode component, or the same physical frequency is split into different modes.

[0066] To address this issue, this embodiment employs the Ensemble Empirical Mode Decomposition (EEMD) algorithm. This involves artificially introducing uniformly distributed Gaussian white noise across the entire frequency domain into the original meteorological signal, forcing the signal to automatically map to a reference frequency band corresponding to its characteristic time scale during decomposition. Subsequently, leveraging the zero statistical mean of Gaussian white noise, different noises are added multiple times and ensemble averaging is performed to cancel out the introduced background noise. This effectively reduces the aliasing effects in complex meteorological signals, enabling the decomposition and identification of meteorological components at different time scales. It allows for clearer separation of high-frequency disturbance components, mid-frequency fluctuation components, and low-frequency trend components, providing a fundamental signal with clear physical meaning for subsequent equipment resonance factor extraction and related response analysis.

[0067] Then, for each micrometeorological modal component, the corresponding analytic signal is constructed by Hilbert transform, the instantaneous phase derivative and amplitude square of the analytic signal are calculated, and the instantaneous phase derivative is used as the instantaneous excitation frequency of the micrometeorological modal component, and the amplitude square is used as the energy density distribution of the micrometeorological modal component.

[0068] In some implementations, to achieve transient feature extraction with extremely high temporal resolution, for each extracted intrinsic mode function... The Hilbert transform is used to shift the phase by 90 degrees, thereby constructing the corresponding analytic signal in the complex plane. : Equation (8) In the formula, The imaginary unit; Represents the Hilbert transform operator; The instantaneous amplitude function representing the analytic signal; This represents the instantaneous phase function of an analytic signal.

[0069] Based on the aforementioned complex plane analytic signal, rigorous mathematical differentiation and transformation of the instantaneous characteristics of micro-meteorological modal components can be performed. Specifically, for the instantaneous phase function... Regarding time variables The instantaneous excitation frequency can be obtained by taking the first derivative (and normalizing it). Meanwhile, for instantaneous amplitude Squaring the distribution yields the energy density distribution, which characterizes the degree of local energy concentration. :

[0070] Equation (9) Equation (10) In the formula, Indicates the first The first micrometeorological factor The micro-meteorological modal components are in the coordinates and time The instantaneous excitation frequency; This represents the corresponding energy density distribution.

[0071] Traditional Fourier transforms can only provide a global average frequency, failing to pinpoint the exact time of transient meteorological impacts. However, by introducing Hilbert transforms to construct analytic signals, the precise frequency and energy magnitude of each micro-meteorological peak or trough at any given moment can be characterized with extreme time resolution at the original sampling rate. This allows meteorological time series to be further characterized as frequency and energy parameters reflecting local mechanical excitation characteristics, achieving a fine-grained characterization of the dynamic features during micro-meteorological transient disturbances. This provides a more physically meaningful analytical basis for identifying transient disaster-causing processes such as short-term light wind-induced high-frequency conductor vibrations.

[0072] This application's embodiments construct a joint processing mechanism combining ensemble empirical mode decomposition and Hilbert analytic signal analysis to address the time-frequency structural characteristics of micrometeorological signals. Based on this technical solution, raw time-series meteorological data can be decomposed into multiple modal components with different time scales and frequency characteristics. Furthermore, the instantaneous frequency and energy distribution characteristics corresponding to each modal component are extracted, resulting in a multi-dimensional physical feature characterization result containing time, frequency, and energy information. This enhances the ability to separate and identify complex disturbance components in the micrometeorological environment and provides a more physically meaningful mathematical representation and data support for subsequent analysis of the resonance response relationship and galloping coupling characteristics between meteorological flow fields and power transmission equipment in specific frequency bands.

[0073] Regarding the implementation details of calculating the coupling factor in step S140, in some examples of embodiments of this application, the natural frequency of the equipment of the transmission line and the material sensitivity coefficient characterizing the sensitivity of the equipment of the transmission line to different micro-meteorological factors are extracted from the equipment characteristic data.

[0074] In practice, the mechanical structure of transmission lines (such as conductors of different spans, insulator strings of specific types, or towering transmission towers) all have their inherent mechanical natural frequencies. At the same time, different equipment materials and structures have significantly different susceptibility to different meteorological factors (for example, conductors are highly sensitive to high-frequency wind speed disturbances and are easily induced by light wind vibrations; while insulators are more sensitive to long-term changes in humidity and temperature).

[0075] Therefore, the steps in this embodiment accurately extract the natural frequency of the corresponding target equipment from the static equipment asset ledger. Furthermore, for different types of microclimate factors, corresponding material sensitivity coefficients were calibrated. This establishes a precise parameter mapping between the "environmental incentives" on the meteorological side and the "physical bearing properties" on the power grid side, providing basic structural parameters for subsequent calculations on whether destructive physical coupling occurs between the two.

[0076] Then, for each micro-topography unit, the first Extracting the first micro-topography unit. Spatial topological coordinates of each micro-topographic unit And based on spatial topological coordinates The energy density distribution and instantaneous excitation frequency extracted at the point are used to calculate the coupling factor between the micro-meteorological factors and the equipment of the transmission line, using the following formula: Equation (11) In the formula, Represents spatial topological coordinates The first The coupling factor between micro-meteorological factors on a micro-topographic unit and the equipment of the power transmission line; This represents the total number of micrometeorological factors. This indicates the corresponding number extracted from the device characteristic data. Material sensitivity coefficients for individual microclimate factors; Indicates the first The first micrometeorological factor The micro-meteorological modal components in spatial topological coordinates Time variables of integrator and integral The energy density distribution; To characterize the exponential term of nonlinear strengthening and ; This represents the natural frequency of the transmission line equipment extracted from the equipment characteristic data; Indicates the first The first micrometeorological factor The micro-meteorological modal components in spatial topological coordinates Time variables of integrator and integral The instantaneous excitation frequency; The length of the observation time window; To prevent the adjustment term of the positive constant with a denominator of zero. When the instantaneous excitation frequency... Approaching the natural frequency At that time, coupling factor Nonlinear jumps are generated to quantitatively characterize the physical resonance intensity.

[0077] Equation (11) aims to quantitatively characterize the disaster-causing intensity of physical resonance in line equipment induced by micro-meteorological transient disturbances. The numerator of the formula... For time window The meteorological excitation energy within the region was integrated, and an exponential term was introduced. Its function is to nonlinearly amplify and enhance high-energy excitation pulses (such as instantaneous strong gusts of wind), highlighting the weight of destructive energy peaks. The denominator of the formula... The integral of the absolute frequency difference between the instantaneous excitation frequency of the meteorological mode and the natural frequency of the equipment was calculated in real time. In physical mechanics, destructive resonance is not caused by an instantaneous frequency coincidence, but requires the excitation frequency to be within a certain time window. It continuously "approaches" its inherent frequency.

[0078] When the instantaneous excitation frequency of micro-meteorological factors (such as the vortex shedding frequency of wind) Continuously approaching the natural frequency When this happens, the integral term will rapidly approach its minimum value (due to the minimum constant). (Overflow prevention). At this point, under the influence of constant or nonlinearly amplified energy molecules, the coupling factor of the output is calculated. This will produce a sharp nonlinear jump. Therefore, by designing a piecewise integral ratio, the model can keenly capture the "frequency domain resonance" phenomenon between the meteorological flow field and the physical equipment, directly translating the originally abstract meteorological data fluctuations into the intensity of resonant shock waves with clear mechanical destructive significance.

[0079] This application's embodiments extend power grid meteorological risk prediction beyond traditional meteorological statistical analysis to the analysis of the coupling relationship between environmental dynamic characteristics and equipment mechanical response. Based on a mathematical model integrating nonlinear time-window energy integration and dynamic frequency difference tracking, the correlation characteristics between micro-meteorological fluctuations and the dynamic response of transmission lines can be characterized, thereby identifying resonance risks that may induce light wind vibrations or strong wind galloping of transmission lines. This enhances the early warning model's ability to characterize the formation mechanism of related risks, improves the model's predictive ability for sudden and concealed line dynamic faults under complex micro-topographic conditions, and provides a more physically meaningful analytical basis for risk prevention.

[0080] Regarding the implementation details of generating the comprehensive risk index in step S150, in some examples of the embodiments of this application, firstly, for the micro-meteorological factors in the spatiotemporal evolution trajectory, nonlinear hazard transformation functions corresponding to wind pressure-induced disasters, icing and freezing, and insulation dampness are constructed respectively to calculate the first... On the first micro-topographic unit Micrometeorological factors in the integral time variable Instantaneous hazard intensity .

[0081] In practice, the original micrometeorological observations (such as wind speed, temperature, and humidity values) are not directly equivalent to the physical disaster intensity of transmission lines. The damage mechanisms of different meteorological factors on power grid equipment exhibit significant nonlinear characteristics. For example, the wind pressure load exerted on conductors by wind speed is usually proportional to the square of the wind speed; while the icing intensity is nonlinearly coupled and controlled by whether the local temperature crosses the freezing point and the ambient humidity.

[0082] In some implementations, taking wind pressure-induced disasters and icing / freezing as examples, a nonlinear hazard transformation function can be constructed using the following technical logic: Equation (12) Equation (13) In the formula, and These represent the calculated wind pressure and the instantaneous hazard intensity of icing, respectively. , and Representing the integration time variable respectively At the moment Instantaneous wind speed, temperature, and relative humidity on individual micro-topographic units; The critical temperature threshold (e.g., 0°C) for triggering the icing phase transition; and This is to establish a hazard conversion coefficient corresponding to a specific physical disaster-causing mechanism. Therefore, by introducing a nonlinear hazard conversion mechanism, abstract meteorological state indicators are precisely mapped to equivalent physical destructive stresses directly acting on power grid equipment, thereby significantly improving the engineering practical significance and physical rigor of the underlying risk assessment data.

[0083] Then, a fractional integral accumulation model based on historical memory decay weights is used to perform multi-factor joint integral fusion on each instantaneous hazard intensity to calculate the comprehensive risk index: Equation (14) In the formula, Indicates the first Each micro-topographic unit at the current assessment time Comprehensive risk indicators; Indicates the first Micrometeorological factors in the integral time variable For the The instantaneous hazard intensity generated by a micro-topographic unit; Indicates the first Disaster-causing weights of micrometeorological factors; Denotes the order parameter of the fractional integral operator and It is used to control the memory decay rate of physical fatigue or icing damage to transmission lines caused by historical meteorological events; It is a gamma function; The time decay kernel function characterizes the historical memory effect.

[0084] This study aims to overcome the limitation of traditional integer-order calculus in accurately describing the material memory degradation effect of a system. Traditional Markov models without aftereffects (focusing only on the current transient) or simple integer-order linear accumulation (assuming that the damage never recovers) cannot realistically simulate real physical catastrophe processes.

[0085] In equation (14), firstly, through the addition of terms... This method achieves weighted fusion of the intensities of various micrometeorological disasters at a specific time. In particular, an attenuation kernel function is introduced into the external integral operation. This endows the power grid risk model with the physical property of "memory decay". Specifically: for the time elapsed since the current moment... The more recent historical meteorological events (i.e. The closer The larger the value of this kernel function, the more severe the impact of recent severe weather on the current equipment status; while for distant historical events, the impact varies with time distance. It exhibits power-law nonlinear decay. Fractional-order parameter. The smaller the value, the faster the memory fades (e.g., rapid drying and recovery after brief exposure to moisture). When the value approaches 1, it indicates that the damage memory is deep and difficult to heal itself (such as irreversible metal fatigue accumulation caused by long-term high-frequency wind deviation). Thus, this fractional integral model accurately reproduces the complex dynamic evolution process in the real physical world, such as "slow accumulation of ice and natural melting and dissipation" or "stress fatigue and microscopic release of stress in metal components", improving the accuracy of the prediction model in capturing the destructive boundary of long-period, slow-changing meteorological disasters.

[0086] This application presents a comprehensive risk assessment technical solution that addresses the nonlinear response characteristics and time-series cumulative damage characteristics in the evolution of micro-meteorological disasters. Based on this solution, the instantaneous and historical cumulative effects of meteorological disaster-causing factors can be considered simultaneously during risk assessment. Furthermore, fractional calculus methods can be used to characterize the state evolution of power transmission equipment under continuous influence in complex environments. This enables a unified risk characterization of multidimensional heterogeneous meteorological disaster-causing factors and enhances the ability to depict the gradual attenuation process of power transmission equipment's carrying capacity. Consequently, it provides physically meaningful data support and analytical basis for long-term composite risk assessment in complex micro-topographical areas.

[0087] In some examples of embodiments of this application, after generating a comprehensive risk index with historical memory effect for each micro-topographic unit, spatial adjacency coupling smoothing processing can also be performed on the comprehensive risk index.

[0088] First, a one-dimensional spatial topological coordinate axis is established along the transmission line, and the comprehensive risk index of each micro-topographic unit is mapped onto the one-dimensional spatial topological coordinate axis to construct a continuously distributed spatial risk field along the line.

[0089] In practice, a transmission line is a linear infrastructure consisting of continuous conductors in terms of physical topology. Although the preceding steps discretize the space into multiple independent micro-topographic units to characterize the complex micro-topographic constraints, in the actual physical world, meteorological disaster effects (such as conductor icing accumulation and wind galloping) are not strictly isolated and stop within the virtual boundary of a certain micro-topographic unit. Instead, they can cause mechanical tension transmission or disaster spread along the taut conductor to adjacent spans.

[0090] In this embodiment, firstly, a one-dimensional spatial topological coordinate axis is established along the actual physical route of the transmission line (for example, the physical arc length along the cable route is used as the coordinate metric, with the substation at the beginning of the transmission line as the origin). Then, the comprehensive risk index of each discrete micro-topographic unit is projected onto this one-dimensional spatial topological coordinate axis based on its actual geographic center location. Through smoothing transition methods such as spatial interpolation, the discrete risk assessment values ​​are reconstructed into a continuously distributed spatial risk field along the line. This completely eliminates the physical fragmentation caused by artificial grid division in the mathematical model, establishing a continuous and unified one-dimensional spatial carrier for subsequent simulations of the mechanical and environmental coupling transmission mechanism of disasters between adjacent spans.

[0091] Then, spatial adjacency coupling smoothing calculation is performed on the spatial risk field using a spatial kernel function based on distance decay to obtain a comprehensive risk index that incorporates spatial spread effects and has undergone spatial smoothing: Equation (15) In the formula, Represents the position of the topological coordinate axis in one-dimensional space. Location, Current Assessment Time The comprehensive risk index after spatial smoothing; This represents the spatial risk field mapped from the comprehensive risk indicators of each micro-topographic unit in the coordinates of the integral variable. The value at that location; The pre-defined spatial coupling influence radius; The normalized distance decay kernel function is used to characterize the degree of spatial coupling and transmission of meteorological disasters between adjacent micro-topographic units; Let be the coordinates of the integral variable along the one-dimensional spatial topological coordinate axes.

[0092] It should be noted that in actual power transmission projects, if a conductor experiences severe icing or strong wind galloping, the resulting abnormal mechanical tension will inevitably be transmitted to adjacent spans through the insulator strings and fittings. In Equation (15), a one-dimensional convolutional smoothing model is constructed to quantitatively characterize this "spatial coupling transmission effect".

[0093] Specifically, in equation (15), the integration interval The effective physical radius of influence for the spatial spread of disasters is strictly limited. This not only aligns with the local attenuation principle of mechanical transmission (i.e., the distance at the far end is unaffected by local galloping), but also significantly reduces the computational complexity of the unbounded integral across the entire line. Simultaneously, the normalized distance attenuation kernel function... (For example, a Gaussian kernel function can be used in implementation) here acts as a weight scheduler for spatial coupling: when the neighborhood coordinates The closer to the current assessment position When the kernel function value is large, it indicates a stronger transmission threat from high-risk locations in adjacent areas to the current location; conversely, as spatial distance increases... With the increase of , the coupling effect exhibits a smooth nonlinear decay. Through the above spatial kernel function integration operation, the model can keenly identify and correct abnormal jumps in risk values ​​caused by "hard grid boundaries", so that a certain high-risk micro-topographic unit can reasonably spill over the risk and affect its adjacent safe units. It better simulates the chain reaction of real power grid disasters such as "local severe icing leading to tension imbalance of several adjacent spans".

[0094] This application's embodiments construct a risk characterization mechanism that combines one-dimensional continuous risk field reconstruction with kernel function spatial neighborhood smoothing, targeting the risk transmission and derivation characteristics under the linear topology of transmission lines. This mechanism can continuously model the discrete risk states of each segment along the transmission line and, combined with the distance decay kernel function, smoothly characterize the spatial influence relationship between adjacent segments. This not only reflects the risk level of locally high-risk segments but also further depicts the range and degree of their derivational influence on surrounding adjacent segments. Consequently, it enhances the continuity and physical consistency of the comprehensive risk index in spatial distribution, providing a more holistic risk assessment basis for power grid dispatching departments to conduct multi-segment collaborative defense along the transmission line.

[0095] Figure 2 A flowchart illustrating an example of generating early warning information containing attribution of physical hazards in a method according to an embodiment of this application is shown.

[0096] like Figure 2 As shown, in step S210, based on the equipment characteristic data of the transmission line, a matching critical cumulative damage threshold and critical resonant shock wave threshold are set.

[0097] In practice, due to differences in construction standards, service life, and materials, the physical limits of equipment in different sections of the power transmission network vary in their ability to withstand extreme environments. In this embodiment, equipment characteristic data (such as conductor ultimate tensile strength, insulator mechanical failure load, and tower structure structural damping ratio) from the underlying ledger are directly accessed. Through material mechanics and dynamics deduction, differentiated safety red lines are tailored for equipment in each section along the line. Specifically, the critical cumulative damage threshold corresponds to the physical boundary where irreversible plastic deformation or overload fracture occurs (used to prevent static damage); while the critical resonant shock wave threshold corresponds to the critical point where the meteorological excitation energy absorbed by the system exceeds its own structural damping dissipation capacity (used to prevent dynamic damage). Therefore, by combining the characteristic differences of equipment in different sections along the transmission line, corresponding risk threshold boundaries can be determined, matching the risk assessment benchmark with the actual load-bearing capacity of the equipment, and enabling differentiated assessment of different types of failure risks. This helps improve the precision and physical rationality of early warning judgments, providing a basis for subsequent risk prevention and control by section and equipment.

[0098] In step S220, for each micro-topographic unit, a comprehensive risk index after spatial smoothing is extracted at the corresponding position of the micro-topographic unit on the one-dimensional spatial topological coordinate axis.

[0099] In step S230, the comprehensive risk index after spatial smoothing is compared with the critical cumulative damage threshold, and the coupling factor is compared with the critical resonant shock threshold.

[0100] In some implementations, to achieve rigorous logical branching and attribution determination, this embodiment transforms the above comparison operation into a priority physical state determination function. In practical engineering, conductor galloping caused by frequency domain resonance is often sudden, highly destructive, and prone to causing phase-to-phase short-circuit tripping, and its urgency is usually higher than that of slowly progressive icing overload. Therefore, the following joint state determination logic model can be constructed at the algorithm's underlying layer:

[0101] In the formula, and They represent the first The power outage risk level and physical cause of the risk are output by each micro-topography unit; and These represent the extracted coupling factor and the comprehensive risk index after spatial smoothing, respectively. and These represent the preset critical resonant shock threshold and critical cumulative damage threshold, respectively.

[0102] In step S240, if the comprehensive risk index after spatial smoothing is less than the critical cumulative damage threshold and the coupling factor is less than the critical resonant shock threshold, the power outage risk level of the corresponding micro-topography unit is determined to be safe.

[0103] In step S250, if the comprehensive risk index after spatial smoothing is greater than or equal to the critical cumulative damage threshold and the coupling factor is lower than the critical resonant shock threshold, the power outage risk level of the corresponding micro-topography unit is determined to be dangerous, and the physical cause of the risk is determined to be icing overload or continuous stress fatigue.

[0104] In step S260, if the coupling factor is greater than or equal to the critical resonant shock wave threshold, the power outage risk level of the corresponding micro-topographic unit is determined to be dangerous, and the physical cause of the risk is determined to be wind deflection frequency resonance or conductor galloping.

[0105] Specifically, when the system enters the first branch (i.e., the coupling factor exceeds the limit), it determines that the transient excitations such as high-frequency eddy current shedding in the current micro-meteorology have already or are about to resonate with the conductor at the same frequency. At this time, it not only outputs the danger level but also accurately identifies the cause as "wind deflection frequency resonance or conductor galloping." When the system enters the second branch (i.e., only the comprehensive risk index exceeds the limit), it indicates that although the system has not experienced severe resonance, the nonlinear integral accumulation of the previous long-term low temperature and high humidity environment has brought the equipment load close to the yield limit. At this time, it identifies the cause as "icing overload or continuous stress fatigue." When the system enters the third branch, it indicates that the current transient impact and historical accumulation are both within the equipment's elastic recovery capacity and safety margin, and it is judged to be at the safe level. Thus, through two-dimensional index constraints and logic tree branches, it perfectly avoids the "false alarms" and "missed alarms" that are easily caused by single indicators, and directly outputs the underlying mechanical root cause of the disaster using a "white box" mechanism.

[0106] In step S270, by combining the power outage risk level and physical cause of the risk of each micro-topography unit, a multi-dimensional risk map covering the transmission line and including spatial location and disaster mechanism is generated.

[0107] In step S280, early warning information containing differentiated scheduling and control strategies for high-risk areas is generated based on the multi-dimensional risk map.

[0108] Here, the ultimate purpose of early warning is to guide defense, using the aforementioned spatial topological coordinates. The rating and attribution results are rendered into an intuitive multi-dimensional risk map. In this map, not only are the extreme risk anchor points along the line highlighted with different color spectra, but the underlying disaster-causing mechanism is also presented intuitively through dynamic legends (such as "vibration ripples" representing dancing risk and "frost superposition" representing overload risk).

[0109] Furthermore, the system can generate corresponding visualized risk distribution information based on risk classification and disaster attribution results, and output differentiated scheduling suggestions or operation and maintenance strategies based on different risk types. For example, when an icing overload risk is identified, corresponding de-icing scheduling suggestions or load adjustment suggestions can be generated; when a local wind polarization risk is identified, corresponding inspection suggestions or operation control suggestions can be generated. This enables the transformation of risk assessment results into scheduling and operation and maintenance suggestions, improving the support capability of early warning information for transmission line defense decisions.

[0110] This application's embodiments construct a decision-making and response support mechanism for power outage risk prediction and introduce a two-dimensional risk assessment model based on physical limit thresholds. By distinguishing and assessing transient resonance impact risks and historical cumulative damage risks, risk classification and assessment can be achieved for micro-topographical scenarios, and corresponding disaster-causing mechanism analysis results can be provided. Based on this, early warning output results combining risk level, disaster attribution, spatial positioning information, and differentiated response suggestions can be generated, thereby improving the pertinence and physical interpretability of early warning results and providing data support for differentiated defense in power grid dispatching and operation and maintenance.

[0111] Figure 3 The diagram illustrates the system operation mechanism of an example of a power outage risk prediction method based on micro-meteorological evolution trajectory according to an embodiment of this application. The system operation mechanism is divided into four hierarchical core architectures from left to right: input data area, core processing and model area, scheduling and control / constraint area, and output and result area.

[0112] like Figure 3 As shown, in the input data area, the system acquires micro-meteorological monitoring data, 3D micro-topography modeling data, and line operation and fault data (which are mainly used to extract equipment characteristic ledger data of transmission lines in specific implementation). These heterogeneous data eliminate misalignment errors in time sampling and spatial distribution through the internal multi-source data spatiotemporal alignment mechanism, thereby constructing a strictly aligned high-quality data foundation for subsequent in-depth feature extraction.

[0113] In the core processing and modeling area, the system executes three key algorithmic logics in parallel. Specifically, the "Micro-meteorological trajectory reconstruction and multi-scale decomposition" module is used to reconstruct the continuous spatiotemporal evolution flow field under the physical constraints of complex micro-topography, and to extract micro-meteorological modal components in specific frequency bands; the "Micro-meteorological-equipment coupling factor calculation" module is responsible for dynamically matching the extracted meteorological excitation energy with the equipment's natural frequency to quantify the transient physical resonance intensity; and the "Fractional-order risk fusion and integration" module uses fractional-order calculus operators to perform nonlinear accumulation of hazard intensity based on historical time windows to generate a comprehensive risk index characterizing the progressive damage of equipment.

[0114] As multidimensional features and indicators are passed down, the mechanism enters the scheduling and control / constraint area. The "risk classification and critical frequency determination" module here constitutes the core of the system's two-dimensional, dual-control decision-making. By comparing the calculated comprehensive risk indicators and coupling factors with the physical critical cumulative damage threshold and critical resonant shock wave threshold, a rigorous control logic and strategy generation are triggered. Finally, in the output and result area, the system renders and generates a "multidimensional risk map and early warning list" covering precise spatial positioning and underlying disaster-causing mechanisms, and directly outputs "proactive operation and maintenance and scheduling control strategies" for high-risk areas to the power grid dispatching terminal. This achieves a mechanistic closed loop from meteorological bottom-level perception to top-level power grid business execution.

[0115] To verify the effectiveness and engineering applicability of the proposed method based on the fusion of micrometeorological evolution trajectory and fractional risk, this embodiment conducted detailed offline simulation and comparative experiments based on historical operation data and micrometeorological monitoring data of real transmission lines.

[0116] During the data preparation phase, two consecutive years of micro-meteorological high-frequency monitoring data (sampling rate, for example, 5 minutes) and synchronous operating status and fault waveform data of the line were selected. Considering that real power grid outage faults are typical of extremely unbalanced samples (i.e., the number of samples with normal and stable operation is far greater than the number of fault samples), this experiment uses a sliding window sampling mechanism to construct a time-series sample set. In the data preprocessing phase, synthetic minority class oversampling (SMOTE) technique is used to appropriately enhance the minority class samples containing fault precursors to ensure the objectivity and generalization ability of subsequent model evaluation.

[0117] To highlight the technical advantages of this application based on physical mechanisms and spatiotemporal evolution modeling, three representative benchmark models were set up for comparison: first, a traditional logistic regression risk assessment model (Macro-LR) based on macro meteorological station data, used to compare the advantages of micro-topography reconstruction; second, a gradient boosting tree classification model (LightGBM) that relies only on single-point discrete micro-meteorological features, used to compare the advantages of continuous spatial field evolution; and third, an end-to-end "black box" deep learning network (CNN-LSTM) that directly inputs high-dimensional time-series meteorological features and outputs fault probabilities, used to compare the advantages of physical disaster attribution and white box mechanisms.

[0118] In selecting evaluation metrics, given the highly imbalanced statistical characteristics of the sample, this experiment abandoned the conventional accuracy metric, which is prone to distorting the evaluation, and instead adopted a more rigorous approach to assessing imbalanced data. Score, Area under the PR curve (AUC-PR), Mean advance warning time ( The core evaluation system includes false alarm rate (FAR).

[0119] Furthermore, to match the core mathematical operator in the aforementioned embodiments, the preset spatial coupling influence radius is set in the core parameter settings of the method in this application as follows: The order parameter of the fractional integral operator is set to This parameter configuration can better match the historical memory decay rate and spatial tension spread characteristics of icing and wind deflection disasters on mountain lines, thereby maximizing the restoration of the real physical scenario of disaster evolution.

[0120] Figure 4 This diagram illustrates a comparative simulation of different methods for assessing the spatiotemporal thermal evolution of risk. The simulation extracted the risk response process of a transmission line under typical conditions of strong convection and cooling weather in a canyon. The horizontal axis of the heatmap represents the time series (in minutes), the vertical axis represents the spatial topological mileage of the transmission line (in terms of specific tower nodes), and the color intensity of the heatmap characterizes the strength of the power outage risk index.

[0121] like Figure 4 In the left part of the image, in the traditional baseline model (Macro-LR) heat map, because the model relies heavily on smooth observation data from macro weather stations, it fails to characterize the spatial evolution under micro-topographic constraints. As a result, the spatial risk distribution of the entire line presents a uniform and relatively lagging low-risk state (the overall background is light-colored), making it impossible to accurately locate the specific tower nodes where a sudden change in risk occurs.

[0122] In contrast, Figure 4 The heatmap of the model in this application (which integrates Unscented Kalman Filtering (UKF) and Fractional Integral Algorithm) on the right side clearly and continuously reconstructs the micro-spatiotemporal ascent and migration trajectory of local micro-meteorological disaster sources (such as local crosswinds or cooling centers) along the canyon topography through the highlighted diagonal bands (shown by the dashed lines in the figure). In particular, thanks to the quantification of the memory accumulation effect of historical environmental effects by fractional integrals, when the intensity of local physical effects approaches the critical coupling threshold (…), Figure 4 The Chinese logo is When the risk index of the model in this application is at a certain value, it exhibits an extremely significant nonlinear jump characteristic.

[0123] Further as Figure 4 As shown in the warning indicator on the right side, compared to the sudden, unpredictable alarms of deep learning "black box" models, the mechanism of this application enables the system to issue progressive, highly sensitive warnings for specific micro-topographic units (such as tower location 12 in the figure). This prediction mechanism achieves an advance warning time of approximately 15 to 20 minutes compared to traditional benchmark models, thus intuitively and powerfully verifying the significant technical advantages of this invention in terms of precise spatial positioning of meteorological disaster sources, tracking of nonlinear disaster evolution, and active defense support under complex micro-topography at a macroscopic visualization level.

[0124] Figure 5 This paper presents a comparative simulation diagram of the precision-recall (PR) performance and overall Pareto front of different methods in a power outage risk prediction scenario. The simulation results are presented in a biplot format.

[0125] like Figure 5 The left subfigure (a) shows the precision-recall (PR) comparison curves of each model under different decision thresholds. The curve distribution shows that the proposed method (the method in this paper) and the CNN-LSTM deep learning model significantly outperform the traditional Macro-LR and LightGBM benchmark models. Specifically, the area under the PR curve (AUC-PR) of the proposed method reaches 0.89, a significant improvement over traditional benchmark models, achieving extremely high risk prediction precision while maintaining high recall. In particular, compared to the uninterpretable "black box" defect of the deep learning CNN-LSTM model, the proposed method, due to its deep integration of the physical and mechanical mechanisms of micro-meteorology and equipment characteristics at the underlying level, can clearly output specific physical causes of disaster when outputting high-risk warnings (such as...). Figure 5 The “wind deflection frequency resonance” indicated in the bubble on the right breaks the limitation of traditional algorithms that can only output fuzzy probabilities, providing highly transparent and interpretable decision support for front-line power grid dispatchers to formulate targeted defense strategies.

[0126] like Figure 5 The right-hand subplot (b) constructs a graph with the false alarm rate (FAR) as the horizontal axis and the average early warning time (AER) as the horizontal axis. The Pareto front scatter bubble plot with the vertical axis () represents the computation time of each algorithm's single inference. This plot objectively reveals the physical boundaries and performance trade-offs of the proposed method in practical engineering applications. Regarding computational overhead, although the proposed mechanism introduces high-resolution spatiotemporal evolution trajectory reconstruction (manifold learning) and fractional-order nonlinear tensor integration, its single inference time (approximately 1.2 seconds) is higher than the lightweight LightGBM (approximately 0.05 seconds) which relies on single-point discrete features, it fully meets the timeliness requirements for proactive maintenance and power grid dispatching scenarios with minute-level responses. Regarding sensor spatial density dependence, sensitivity simulations show that the proposed method has certain requirements for the continuity of the physical distribution of the monitoring network. When sensor nodes along the line are damaged or the spacing between them exceeds 3 km, the interpolation error of manifold learning in reconstructing the nonlinear micrometeorological boundary will increase, leading to an increase in the system's false alarm rate (FAR) of approximately 8.5%.

[0127] It should be noted that, given the high risk identification sensitivity and strong physical mechanism characterization capabilities of the power outage risk prediction system proposed in this application, it can be preferentially deployed in important transmission channels and key monitoring sections with complex micro-topographical conditions and high risks of micro-meteorological disasters, such as canyon sections prone to micro-wind vibration or sections with high incidence of local icing. By adopting a differentiated deployment approach, the system can fully leverage its refined risk early warning capabilities while also considering the consumption of underlying computing resources, the deployment cost of IoT sensing devices, and the requirements for ensuring the safe operation of the power grid.

[0128] This application proposes a power outage risk prediction method based on micrometeorological evolution trajectories. Compared with traditional prediction methods based on macrometeorology or "black box" neural networks, it exhibits significant advantages in terms of white-box approach and mechanism-based analysis. Specifically, it breaks through the limitations of traditional discrete observations by using unscented Kalman filtering and manifold learning algorithms to accurately reconstruct the spatiotemporal continuous evolution trajectory of micrometeorological factors on a three-dimensional topological grid in complex terrain areas. Furthermore, it extracts micrometeorological modal components at different time scales by combining multi-scale signal decomposition. Based on this, a fractional integral operator and a spatial distance kernel function are introduced to nonlinearly accumulate the instantaneous hazard intensity of micrometeorology on both the time series and spatial topology, perfectly aligning with the "historical memory" and "spatial spread" effects of continuous severe environmental conditions causing physical fatigue and icing damage to power line equipment. In particular, by calculating the coupling factor that deeply relates micrometeorology and equipment characteristics, and performing the judgment under the joint constraint of dual physical critical thresholds, the underlying physical disaster attribution such as "wind deflection frequency resonance" or "icing overload" can be directly revealed when outputting high-risk warnings, thereby greatly improving the interpretability of the assessment results and the acceptance and trust of front-line dispatchers.

[0129] In terms of engineering verification and performance evaluation, offline simulations based on real power grid extreme imbalance fault samples fully demonstrate that the proposed method significantly outperforms traditional statistical inference and deep learning benchmark models in core evaluation indicators such as the area under the precision-recall curve (AUC-PR) and average early warning time. This method not only enables precise spatial location and thermal evolution tracking of hidden disaster sources by generating multi-dimensional risk maps, but also objectively quantifies the system computational overhead introduced by high-frequency trajectory reconstruction and rigorously demonstrates the engineering constraints of sensor deployment density (e.g., 3km spacing at the red line boundary) on manifold reconstruction error and false alarm rate. This objective boundary condition analysis provides valuable scientific guidance for the deployment of high-efficiency, differentiated monitoring networks for transmission lines in complex micro-topographic areas.

[0130] It should be understood that the dual-drive framework of physical mechanism and data-driven approach constructed in this application can be further extended to the field of collaborative risk analysis of multi-energy power systems in the future. To address the high computational cost of complex fractional tensor integration operations, edge computing architectures can be introduced in future implementations to accelerate model inference in local grids, further meeting the ultra-fast response requirements of microsecond-level relay protection. Furthermore, the integration of multi-source observation data, such as low-orbit microsatellite remote sensing, can be explored to effectively compensate for the lack of local upper-air wind field information, thereby continuously expanding the applicability of this prediction model in wide-area complex power grids and its all-weather real-time defense capabilities.

[0131] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of combined actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Secondly, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application. In the above embodiments, the descriptions of each embodiment have their own emphasis; for parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0132] Figure 6 A structural block diagram of an example of a power outage risk prediction system based on micrometeorological evolution trajectory according to an embodiment of this application is shown.

[0133] like Figure 6 As shown, the power outage risk prediction system 600 based on micro-meteorological evolution trajectory includes a data acquisition unit 610, a spatiotemporal alignment unit 620, a trajectory reconstruction unit 630, a modal analysis unit 640, a risk accumulation unit 650, and an early warning output unit 660.

[0134] The data acquisition unit 610 is used to acquire micro-meteorological monitoring data, micro-topographic data and equipment characteristic data of the transmission line along the line, and divide the micro-topographic data into multiple micro-topographic units along the line.

[0135] The spatiotemporal alignment unit 620 is used to perform spatiotemporal alignment of the micro-meteorological monitoring data and the micro-topographic data by combining the time synchronization signal, so as to construct a multi-dimensional basic dataset.

[0136] The trajectory reconstruction unit 630 is used to reconstruct the spatiotemporal evolution trajectory of micro-meteorological factors continuously distributed along the transmission line based on the multidimensional basic dataset; the micro-meteorological factors include at least one of the following: wind speed, wind direction, temperature, humidity, air pressure, and rainfall.

[0137] The modal analysis unit 640 is used to extract the micro-meteorological modal components in the spatiotemporal evolution trajectory and perform matching analysis between the micro-meteorological modal components and the equipment characteristic data to calculate the coupling factor between the micro-meteorological factor and the equipment of the transmission line; wherein, the micro-meteorological modal components characterize the oscillation components of the micro-meteorological factor along the spatiotemporal evolution trajectory at different time scales.

[0138] The risk accumulation unit 650 is used to determine the instantaneous hazard intensity of the micro-meteorological factors in the spatiotemporal evolution trajectory on each of the micro-topographic units, and to perform a nonlinear accumulation operation based on a historical time window on the instantaneous hazard intensity using a fractional integral operator to generate a comprehensive risk index for each of the micro-topographic units that has a historical memory effect.

[0139] The early warning output unit 660 is used to determine the power outage risk level of the transmission line based on the comprehensive risk index of each of the micro-terrain units and the coupling factor, and to generate early warning information that includes the physical causes of the risk.

[0140] In some embodiments, this application provides a non-volatile computer-readable storage medium storing one or more programs including execution instructions. The execution instructions can be read and executed by electronic devices (including but not limited to computers, servers, or network devices) to perform the steps of any of the above-described power outage risk prediction methods based on micro-meteorological evolution trajectories.

[0141] In some embodiments, this application also provides a computer program product, the computer program product including a computer program stored on a non-volatile computer-readable storage medium, the computer program including program instructions, which, when executed by a computer, cause the computer to perform the steps of any of the above-described methods for predicting power outage risks based on micro-meteorological evolution trajectories.

[0142] In some embodiments, this application also provides an electronic device comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform steps of a power outage risk prediction method based on micrometeorological evolution trajectories.

[0143] The above-described product can perform the methods provided in the embodiments of this application, and has the corresponding functional modules and beneficial effects for performing the methods. Technical details not described in detail in this embodiment can be found in the methods provided in the embodiments of this application.

[0144] The electronic devices in this application can exist in various forms, including but not limited to: mobile communication devices, ultra-mobile personal computer devices, portable entertainment devices, or other airborne electronic devices with data interaction functions.

[0145] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0146] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0147] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for predicting power outage risk based on micro-meteorological evolution trajectories, characterized in that, The method includes: Acquire micro-meteorological monitoring data, micro-topographic data, and equipment characteristic data of the transmission line along the line, and divide the micro-topographic data into multiple micro-topographic units along the line; The micro-meteorological monitoring data and the micro-topographic data are spatiotemporally aligned by combining time synchronization signals to construct a multi-dimensional basic dataset; Based on the aforementioned multidimensional basic dataset, the spatiotemporal evolution trajectory of micro-meteorological factors continuously distributed along the transmission line is reconstructed; the micro-meteorological factors include at least one of the following: wind speed, wind direction, temperature, humidity, air pressure, and rainfall. Micrometeorological modal components are extracted from the spatiotemporal evolution trajectory, and the micrometeorological modal components are matched and analyzed with the equipment characteristic data to calculate the coupling factor between the micrometeorological factor and the equipment of the transmission line; wherein, the micrometeorological modal components characterize the oscillation components of the micrometeorological factor along the spatiotemporal evolution trajectory at different time scales; The instantaneous hazard intensity of the micro-meteorological factors in the spatiotemporal evolution trajectory on each micro-topographic unit is determined, and a nonlinear accumulation operation based on a historical time window is performed on the instantaneous hazard intensity using a fractional integral operator to generate a comprehensive risk index for each micro-topographic unit with a historical memory effect. Based on the comprehensive risk index of each micro-topographic unit and the coupling factor, the power outage risk level of the transmission line is determined, and early warning information containing the physical causes of the risk is generated.

2. The method according to claim 1, characterized in that, The method of combining time synchronization signals to perform spatiotemporal alignment of the micro-meteorological monitoring data and the micro-topographic data to construct a multi-dimensional basic dataset includes: Using the Network Time Protocol timestamp as a unified time synchronization signal, the time series labels of the micro-meteorological monitoring data collected by each sensor node are extracted, and spline interpolation algorithm is used to fill in the missing time series values ​​in the micro-meteorological monitoring data to generate time-aligned meteorological data. The digital elevation model and digital surface model features of each micro-topographic unit are extracted, and the local slope and canyon index are calculated to construct a three-dimensional spatial topological mesh for each micro-topographic unit based on the local slope and canyon index. A spatial resampling algorithm based on Delaunay triangulation is used to map the time-series aligned meteorological data to the grid nodes of the corresponding three-dimensional spatial topology grid of the micro-topographic unit, so as to construct a multidimensional basic dataset that is strictly aligned in both time and spatial dimensions.

3. The method according to claim 2, characterized in that, The process of reconstructing the spatiotemporal evolution trajectory of micro-meteorological factors continuously distributed along the transmission line based on the multidimensional basic dataset includes: For any number Each micro-meteorological factor is used to extract data sequences from the grid nodes of the three-dimensional spatial topological grid in the multi-dimensional basic dataset to construct a state vector. Based on the local slope and canyon index as micro-topographical environmental constraints, a nonlinear state transition equation and observation equation are established: ， ， In the formula, To characterize the nonlinear transfer function of the dynamic evolution of micrometeorological states under topographic constraints, For nonlinear observation functions, For process noise, To observe the noise, The sampling time interval, For a moment The observation vector; Indicates time The state vector; Indicates time The state vector; For the aforementioned nonlinear state transition equation and observation equation, the unscented Kalman filter algorithm is used to construct a Sigma point set for the state vector. Lossless transfer of nonlinear probability distribution is performed to estimate the posterior mean and covariance at each time point, thereby obtaining a discrete meteorological field distribution with temporal smoothness characteristics. A manifold learning algorithm based on Laplace eigenmaps is used to map the discrete meteorological field distribution to a low-dimensional embedding space to extract the meteorological feature flow axis that fits the terrain orientation characterized by the local slope and canyon index. Based on the meteorological feature flow axis, spatial continuous interpolation is performed on the three-dimensional spatial topological grid of each micro-topographic unit to reconstruct the spatiotemporal evolution trajectory of the micro-meteorological factors continuously distributed along the transmission line.

4. The method according to claim 3, characterized in that, The extraction of micro-meteorological modal components from the spatiotemporal evolution trajectory includes: For the micro-meteorological factor function in the spatiotemporal evolution trajectory A Gaussian white noise sequence with a defined standard deviation is introduced and ensemble averaging is performed. The ensemble empirical mode decomposition algorithm is then used to analyze the micrometeorological factor function. Multi-scale signal decomposition is performed to eliminate mode aliasing of complex meteorological signals, and the micro-meteorological factor function is then used. Decomposed into multiple eigenmode functions and residual terms of different frequencies: ， In the formula, Represents the spatial topological coordinates of the transmission line. Indicates the first Micrometeorological factors in spatial topological coordinates and time coordinates Microclimate factor function; The total number of layers of the intrinsic mode functions obtained from the decomposition. This represents the residual trend term reflecting the long-term evolution of meteorological patterns. Indicates the first Each intrinsic mode function characterizes the oscillation component of the micrometeorological factor along the spatiotemporal evolution trajectory at different time scales, and serves as the corresponding micrometeorological mode component; For each of the aforementioned micrometeorological modal components, a corresponding analytical signal is constructed through Hilbert transform. The instantaneous phase derivative and amplitude square of the analytical signal are calculated. The instantaneous phase derivative is used as the instantaneous excitation frequency of the micrometeorological modal component, and the amplitude square is used as the energy density distribution of the micrometeorological modal component.

5. The method according to claim 4, characterized in that, The step of matching and analyzing the micro-meteorological modal components with the equipment characteristic data to calculate the coupling factor between the micro-meteorological factors and the equipment of the transmission line includes: The inherent natural frequencies of the equipment in the transmission line and the material sensitivity coefficients characterizing the sensitivity of the equipment in the transmission line to different micro-meteorological factors are extracted from the equipment characteristic data. For each of the aforementioned micro-terrain units, the first The first micro-topography unit is extracted. Spatial topological coordinates of each micro-topographic unit and based on the spatial topological coordinates The energy density distribution extracted from the source and the instantaneous excitation frequency are used to calculate the coupling factor between the micro-meteorological factor and the equipment of the transmission line using the following formula: ， In the formula, Represents the spatial topological coordinates The first The coupling factor between micro-meteorological factors on a micro-topographic unit and the equipment of the transmission line; This represents the total number of micrometeorological factors. This indicates the corresponding number extracted from the device characteristic data. Material sensitivity coefficients for individual micro-meteorological factors; Indicates the first The first micrometeorological factor Each micrometeorological modal component is located in the spatial topological coordinates. Time variables of integrator and integral The energy density distribution; To characterize the exponential term of nonlinear strengthening and ; This represents the natural frequency of the equipment in the transmission line extracted from the equipment characteristic data; Indicates the first The first micrometeorological factor Each micrometeorological modal component is located in the spatial topological coordinates. Time variables of integrator and integral The instantaneous excitation frequency; The length of the observation time window; To prevent the adjustment term of a positive constant with a denominator of zero; Wherein, when the instantaneous excitation frequency Approximating the natural frequency At that time, the coupling factor Nonlinear jumps are generated to quantitatively characterize the physical resonance intensity.

6. The method according to claim 5, characterized in that, The process of determining the instantaneous hazard intensity of the micro-meteorological factors in the spatiotemporal evolution trajectory on each of the micro-topographic units, and performing a nonlinear accumulation operation based on a historical time window on the instantaneous hazard intensity using a fractional integral operator to generate a comprehensive risk index for each of the micro-topographic units that possesses a historical memory effect, includes: For the micrometeorological factors in the aforementioned spatiotemporal evolution trajectory, nonlinear hazard transformation functions corresponding to wind pressure-induced disasters, icing and freezing, and insulation dampness are constructed to calculate the th On the first micro-topographic unit Micrometeorological factors in integral time variables The instantaneous hazard intensity generated ; A fractional integral accumulation model based on historical memory decay weights is used to perform multi-factor joint integral fusion on each of the instantaneous hazard intensities to calculate the comprehensive risk index: ， In the formula, Indicates the first Each micro-topographic unit at the current assessment time The aforementioned comprehensive risk indicators; Indicates the first Micrometeorological factors in the integral time variable For the The instantaneous hazard intensity generated by each micro-topographic unit; Indicates the first Disaster-causing weights of micrometeorological factors; Denotes the order parameter of the fractional integral operator and It is used to control the memory decay rate of physical fatigue or icing damage to transmission lines caused by historical meteorological events; It is a gamma function; The time decay kernel function characterizes the historical memory effect.

7. The method according to claim 6, characterized in that, After generating a comprehensive risk index for each of the micro-topographic units that possesses a historical memory effect, the method further includes performing spatial adjacency coupling smoothing processing on the comprehensive risk index, specifically including: A one-dimensional spatial topological coordinate axis is established along the transmission line, and the comprehensive risk index of each micro-topographic unit is mapped onto the one-dimensional spatial topological coordinate axis to construct a continuously distributed spatial risk field along the line. The spatial risk field is spatially adjacent coupled and smoothed by a spatial kernel function based on distance decay to obtain the comprehensive risk index that incorporates spatial spread effects and has undergone spatial smoothing. ， In the formula, Represents the position of the topological coordinate axis in one-dimensional space. Location, Current Assessment Time The comprehensive risk index after spatial smoothing; The spatial risk field, mapped from the comprehensive risk index of each of the micro-topographic units, is represented by the coordinates of the integral variable. The value at that location; The pre-defined spatial coupling influence radius; The normalized distance decay kernel function is used to characterize the degree of spatial coupling and transmission of meteorological disasters between adjacent micro-topographic units; Let be the coordinates of the integral variable along the one-dimensional spatial topological coordinate axis.

8. The method according to claim 7, characterized in that, The method, based on the comprehensive risk index of each micro-topographic unit and the coupling factor, determines the outage risk level of the transmission line and generates early warning information containing physical causes of the risk, including: Based on the equipment characteristic data of the transmission line, a matching critical cumulative damage threshold and critical resonant shock wave threshold are set. For each micro-topographic unit, the spatially smoothed comprehensive risk index of the micro-topographic unit at the corresponding position on the one-dimensional spatial topological coordinate axis is extracted, and the spatially smoothed comprehensive risk index is compared with the critical cumulative damage threshold, and the coupling factor is compared with the critical resonant shock wave threshold. If the comprehensive risk index after spatial smoothing is less than the critical cumulative damage threshold and the coupling factor is lower than the critical resonant shock threshold, the power outage risk level of the corresponding micro-topography unit is determined to be safe. If the comprehensive risk index after spatial smoothing is greater than or equal to the critical cumulative damage threshold and the coupling factor is lower than the critical resonant shock threshold, the power outage risk level of the corresponding micro-topography unit is determined to be dangerous, and the physical cause of the risk is determined to be icing overload or continuous stress fatigue. If the coupling factor is greater than or equal to the critical resonant shock threshold, the power outage risk level of the corresponding micro-topography unit is determined to be dangerous, and the physical cause of the risk is determined to be wind deflection frequency resonance or conductor galloping. By combining the power outage risk level and the physical cause of the risk of each micro-terrain unit, a multi-dimensional risk map covering the transmission line and including spatial location and disaster mechanism is generated. Based on the multi-dimensional risk map, early warning information containing differentiated scheduling and control strategies for high-risk areas is generated.

9. A power outage risk prediction system based on micro-meteorological evolution trajectory, characterized in that, The system includes: The data acquisition unit is used to acquire micro-meteorological monitoring data, micro-topographic data, and equipment characteristic data of the transmission line along the line, and to divide the micro-topographic data into multiple micro-topographic units along the line. The spatiotemporal alignment unit is used to perform spatiotemporal alignment of the micro-meteorological monitoring data and the micro-topographic data by combining the time synchronization signal, so as to construct a multi-dimensional basic dataset; The trajectory reconstruction unit is used to reconstruct the spatiotemporal evolution trajectory of micro-meteorological factors continuously distributed along the transmission line based on the multidimensional basic dataset; the micro-meteorological factors include at least one of the following: wind speed, wind direction, temperature, humidity, air pressure, and rainfall. The modal analysis unit is used to extract the micro-meteorological modal components in the spatiotemporal evolution trajectory and perform matching analysis between the micro-meteorological modal components and the equipment characteristic data to calculate the coupling factor between the micro-meteorological factor and the equipment of the transmission line; wherein, the micro-meteorological modal components characterize the oscillation components of the micro-meteorological factor along the spatiotemporal evolution trajectory at different time scales; A risk accumulation unit is used to determine the instantaneous hazard intensity of the micro-meteorological factors in the spatiotemporal evolution trajectory on each of the micro-topographic units, and to perform a nonlinear accumulation operation based on a historical time window on the instantaneous hazard intensity using a fractional integral operator to generate a comprehensive risk index for each of the micro-topographic units that has a historical memory effect. The early warning output unit is used to determine the power outage risk level of the transmission line based on the comprehensive risk index of each micro-terrain unit and the coupling factor, and to generate early warning information that includes the physical causes of the risk.

Images