Geological disaster safety risk assessment method and system for power grid region

Abstract

The invention discloses a geological disaster safety risk assessment method and system for a power grid region, which is applied to the field of power system safety and disaster risk assessment, and comprises the steps: obtaining remote sensing deformation monitoring data and power grid operation monitoring data of a target power grid region; performing time sequence analysis and deformation calculation on the remote sensing deformation monitoring data to generate deformation time sequence data; converting the deformation data into a deformation characteristic quantity associated with the power grid key facility; performing time sequence alignment and fusion processing on the multi-source data, and constructing a comprehensive operation risk fusion data set; and obtaining a risk critical value through a time sequence prediction model based on the fused data set, and performing risk quantitative evaluation and load adjustment according to the risk critical value. According to the invention, through fusion of radar remote sensing deformation monitoring and power grid operation data and combination of the deep learning prediction model, accurate assessment and advanced early warning of geological disaster risks of key facilities of the power grid are realized, and the active defense capability and operation toughness of the power grid to geological disasters are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of power system safety and disaster risk assessment technology, and in particular to a method and system for assessing geological disaster safety risks in power grid areas. Background Technology

[0002] With the expansion of energy resource allocation, a large number of critical facilities such as transmission lines, power poles, and substations are being built and operating in mountainous and hilly areas with complex geological environments. These areas are frequently affected by geological disasters such as landslides and collapses due to factors like topography and rainfall, posing a continuous threat to power grid facilities. Accidents such as tower collapses and line breaks can lead to large-scale power outages. Therefore, accurate assessment of geological disaster risks in power grid areas is crucial for enhancing the resilience and disaster prevention capabilities of the power system.

[0003] Currently, the assessment method for power grid geological disaster risks relies on inspection personnel judging the results of regular manual inspections based on their own experience. This method is highly subjective, has low accuracy, and, due to the limited coverage of the power grid area by manual inspections, it is difficult to capture early landslide warnings in a timely manner, resulting in a significant lag in risk identification. This makes it difficult for the assessment results to support the forward-looking adjustment of power grid operation modes and the formulation of proactive defense strategies, severely hindering the technological transformation of power grid geological disaster risk prevention and control from passive response to proactive early warning. Summary of the Invention

[0004] This invention provides a method and system for assessing geological disaster safety risks in power grid areas, which addresses the technical problem of insufficient risk assessment accuracy due to the lack of dynamic correlation between geological deformation processes and the operating status of power grid facilities. By integrating multi-source data and applying prediction models, it promotes the intelligent process of geological disaster safety risk assessment in power grid areas.

[0005] To address the aforementioned technical problems, embodiments of the present invention provide a method for assessing geological disaster safety risks in power grid areas, comprising: Acquire multi-source monitoring data for the target power grid area, wherein the multi-source monitoring data includes at least remote sensing deformation monitoring data and power grid operation monitoring data; Time-series analysis is performed on the remote sensing deformation monitoring data, and the deformation time-series data of the target power grid area is obtained based on the analysis results; Based on the geographical location of key facilities in the target power grid area, the deformation time series data is converted into deformation feature quantities associated with the key facilities; Based on the deformation time series data, the power grid operation monitoring data is time-aligned, and based on the deformation characteristic quantities and the aligned power grid operation monitoring data, a fusion dataset is constructed to characterize the comprehensive operational risk of the key facilities. The fused dataset is input into the constructed time-series prediction model to obtain risk thresholds that indicate the safety status of the critical facilities. Based on the risk thresholds, the critical facilities in the target power grid area are subjected to risk quantification assessment, and the assessment results are used to adjust the power load of the target power grid area.

[0006] As one preferred embodiment, the step of performing time-series analysis on the remote sensing deformation monitoring data and obtaining the deformation time-series data of the target power grid area based on the analysis results includes: Extract the time-series radar image set of the target power grid area from the remote sensing deformation monitoring data, and select image combinations from the time-series radar image set to form interferometric units; The stage-specific deformation is extracted from each of the interference units using differential interferometry. The deformation time series data are obtained by performing time series integration on each of the aforementioned stage deformation variables.

[0007] As one preferred embodiment, the step of converting the deformation time-series data into deformation feature quantities associated with the key facilities in the target power grid area, based on the geographical location of the key facilities, includes: Based on the coordinate information of the geographical location of the key facility, the spatial image range of the key facility is determined; Deformation information related to the spatial image range is extracted from the deformation time series data, and the deformation information is processed to generate the deformation feature quantity used to quantify the degree of deformation of the geographical location.

[0008] As one preferred embodiment, the step of performing time-series alignment processing on the power grid operation monitoring data based on the deformation time-series data includes: Based on the time series of the deformation time series data, the target time series is determined; The corresponding values ​​of the power grid operation monitoring data on the target time series are obtained by interpolation, and the corresponding values ​​are paired with the deformation time series data to complete the time series alignment process.

[0009] As one preferred embodiment, the risk quantification assessment of the critical facilities in the target power grid area based on the risk threshold includes: The risk threshold is compared with a preset risk threshold. Based on the comparison results, the range of the preset risk threshold where the risk threshold is located is determined, the risk level of the key facility is obtained, and the risk quantification assessment is completed.

[0010] Another embodiment of the present invention provides a geological disaster safety risk assessment system for power grid areas, comprising: The acquisition module is used to acquire multi-source monitoring data of the target power grid area, wherein the multi-source monitoring data includes at least remote sensing deformation monitoring data and power grid operation monitoring data; The analysis module is used to perform time-series analysis on the remote sensing deformation monitoring data, and obtain the deformation time-series data of the target power grid area based on the analysis results; The spatial association module is used to convert the deformation time series data into deformation feature quantities associated with the key facilities, based on the geographical location of the key facilities in the target power grid area. The time fusion module is used to perform time-series alignment processing on the power grid operation monitoring data based on the deformation time-series data, and to construct a fusion dataset to characterize the comprehensive operational risk of the key facilities based on the deformation feature quantity and the aligned power grid operation monitoring data. The risk assessment module is used to input the fused dataset into the constructed time-series prediction model to obtain a risk threshold value that indicates the safety status of the critical facility, and to perform a risk quantification assessment of the critical facility in the target power grid area based on the risk threshold value. The assessment result is used to adjust the power load of the target power grid area.

[0011] As one preferred embodiment, the analysis module is further configured to: Extract the time-series radar image set of the target power grid area from the remote sensing deformation monitoring data, and select image combinations from the time-series radar image set to form interferometric units; The stage-specific deformation is extracted from each of the interference units using differential interferometry. The deformation time series data are obtained by performing time series integration on each of the aforementioned stage deformation variables.

[0012] As one preferred embodiment, the spatial association module is further configured to: Based on the coordinate information of the geographical location of the key facility, the spatial image range of the key facility is determined; Deformation information related to the spatial image range is extracted from the deformation time series data, and the deformation information is processed to generate the deformation feature quantity used to quantify the degree of deformation of the geographical location.

[0013] As one preferred embodiment, the time fusion module is further configured to: Based on the time series of the deformation time series data, the target time series is determined; The corresponding values ​​of the power grid operation monitoring data on the target time series are obtained by interpolation, and the corresponding values ​​are paired with the deformation time series data to complete the time series alignment process.

[0014] As one preferred embodiment, the risk assessment module is further used for: The risk threshold is compared with a preset risk threshold. Based on the comparison results, the range of the preset risk threshold where the risk threshold is located is determined, the risk level of the key facility is obtained, and the risk quantification assessment is completed.

[0015] Compared with the prior art, the beneficial effects of the embodiments of the present invention are at least one of the following: (1) This invention acquires remote sensing deformation monitoring data and power grid operation monitoring data, performs time-series analysis on the remote sensing deformation monitoring data to obtain deformation time-series data, and then converts it into deformation characteristic quantities based on the geographical location of key power grid facilities. This data is then fused with time-series aligned power grid operation monitoring data to construct a dataset, which is input into a time-series prediction model to obtain risk threshold values ​​for risk assessment. This process deeply couples geological deformation information with the power grid operation status, overcoming the limitations of isolated analysis of the two types of data in existing technologies, and achieving a more accurate quantitative assessment of the geological hazard threats to power grid facilities.

[0016] (2) This invention constructs a complete technical system from multi-source heterogeneous data collaborative perception and dynamic risk prediction to active regulation of power grid operation, realizing early identification and accurate quantification of geological disaster risks. This method significantly improves the credibility and timeliness of risk assessment results, provides key technical support for power grid dispatching decisions, and effectively enhances the resilience and safe and stable operation level of power grid infrastructure in complex geological environments. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a method for assessing geological disaster safety risks in a power grid area, according to one embodiment of the present invention. Figure 2 This is a schematic diagram of a geological disaster safety risk assessment system for a power grid area in one embodiment of the present invention.

[0018] Figure label: The module includes: acquisition module 11, analysis module 12, spatial association module 13, temporal fusion module 14, and risk assessment module 15. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0020] In the description of this application, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined with "first," "second," "third," etc., may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0021] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to the internal communication between two components. The terms "vertical," "horizontal," "left," "right," "upper," "lower," and similar expressions used herein are for illustrative purposes only and do not indicate or imply that the device or component referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as limiting the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0022] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this specification is for the purpose of describing specific embodiments only and is not intended to limit the invention. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0023] One embodiment of the present invention provides a method for assessing geological disaster safety risks in power grid areas. For details, please refer to [link / reference needed]. Figure 1 , Figure 1 The method for assessing geological disaster safety risks in a power grid area, as shown in one embodiment of the present invention, includes steps S1-S5: S1: Acquire multi-source monitoring data for the target power grid area. The multi-source monitoring data shall include at least remote sensing deformation monitoring data and power grid operation monitoring data. Remote sensing deformation monitoring data refers to a set of information that quantitatively reflects the changes in the Earth's surface location over time, obtained through repeated observations of the Earth's surface using spaceborne or airborne synthetic aperture radar platforms. Its core data source is single-view complex imagery, a format that simultaneously records the backscattering amplitude and phase information of ground features. Power grid operation monitoring data refers to a set of physical parameters directly collected by various sensors deployed on power grid physical facilities, used to characterize the real-time operating status and health of power transmission equipment and supporting structures.

[0024] Step S1 obtains the raw data required for geological disaster safety risk assessment in the power grid area, thereby constructing a multi-dimensional initial dataset to provide information input for subsequent accurate deformation calculation, data fusion and risk prediction.

[0025] Specifically, the spatial scope of the monitoring unit is first determined. This scope centers on the transmission towers, transmission line corridors, and substation sites within the target power grid area, extending to surrounding landslide hazard areas that may threaten these critical facilities. Then, remote sensing deformation monitoring data and power grid operation monitoring data are acquired. The remote sensing deformation monitoring data is primarily obtained through space-based Earth observation technology, with its core data source being time-series single-view complex images acquired by synthetic aperture radar satellites. Power grid operation monitoring data is directly acquired through a physical sensor network deployed on critical power grid facilities. In this embodiment, L-band satellite imagery from the "LuTan-1" satellite can be used. Due to its longer wavelength, this type of data has better penetration through dense vegetation in mountainous areas, effectively reducing signal decoherence, and its short revisit period of four to eight days ensures the density of the time series. Typically, at least twenty images covering the monitoring period are required to construct a reliable time-series dataset. Simultaneously, reference digital elevation model (DEM) data for the monitoring area needs to be obtained from a publicly available data platform. Specifically, a DEM model with a spatial resolution of approximately 30 meters, derived from the Space Shuttle Radar Topographic Mapping Mission, can be used. This model will serve as a crucial input for simulating and removing terrain phases in subsequent interferometric processing. In this embodiment, the power grid operation monitoring data includes at least: tower tilt angle data measured by tilt sensors, which directly reflects changes in the verticality of the tower structure; tower foundation stress data measured by stress sensors, which reflects the mechanical load borne by the foundation structure; conductor tension data measured by tension sensors, which relates to conductor sag and safety margin; and line current load data measured by current transformers, which are core electrical parameters for assessing the transmission status of the lines. These data collectively constitute the key criteria for assessing whether the power grid infrastructure is at risk of abnormality or near-failure due to external geological disasters.

[0026] After acquiring the raw remote sensing deformation monitoring data and power grid operation monitoring data, they are standardized to ensure data quality and usability. For remote sensing deformation monitoring data, preprocessing includes necessary radiometric calibration and noise filtering using interferometric radar processing software. Radiometric calibration converts the raw values ​​into standardized backscattering coefficients; noise filtering can employ various mature interferogram filtering algorithms, including but not limited to mean filtering or adaptive filtering algorithms. These algorithms can effectively suppress speckle noise, among which the Goldstein filtering algorithm, due to its ability to adaptively adjust the filtering intensity based on the local coherence of the interferogram, achieves a good balance between noise suppression and phase preservation. For power grid operation monitoring data, preprocessing includes outlier removal and data smoothing, specifically using a statistically based sliding window method: setting a window containing a specific number of sampling points, calculating the median and its absolute deviation within the window; observations deviating from the median by more than three times the absolute deviation of the median are identified as outliers and removed. Subsequently, a moving average algorithm is applied to the processed data sequence for smoothing to eliminate irregular fluctuations. The preprocessed remote sensing deformation monitoring data will be input into the analysis module as input for three-track differential interferometry, used to accurately calculate the landslide deformation of the monitoring unit at different time periods. The power grid operation monitoring data will then be further processed along with the deformation time-series data output from step S2.

[0027] S2: Perform time-series analysis on remote sensing deformation monitoring data, and obtain deformation time-series data of the target power grid area based on the analysis results; Step S2 involves time-series analysis of remote sensing deformation monitoring data. Based on the physical principle of coherence between multi-temporal synthetic aperture radar (SAR) images, and using the "three-track method" differential interferometry technique, it processes and analyzes radar image datasets covering the same geographical area acquired at different times. The aim is to quantitatively extract and reconstruct the temporal evolution of the displacement of surface targets along the radar line of sight from the original phase signals, which contain noise. The deformation time-series data, the output of the time-series analysis, is a structured dataset. Its core content is the cumulative surface deformation value corresponding to specific monitoring points or areas within the target power grid region at a series of regular time points. This dataset comprehensively depicts the entire process of deformation accumulation or evolution of the monitored target from its initial state over time. Its form is typically represented as a curve or sequence with time as the horizontal axis and cumulative deformation as the vertical axis.

[0028] Step S2 performs time-series analysis on the remote sensing deformation monitoring data, ultimately generating deformation time-series data that reflects the cumulative deformation process of landslides within the target power grid area. This step primarily transforms the original radar image sequence into quantitative and intuitive surface deformation information, providing a direct data foundation for subsequent deformation prediction.

[0029] Preferably, in one embodiment of the present invention, time-series analysis is performed on remote sensing deformation monitoring data, and deformation time-series data of the target power grid area is obtained based on the analysis results, including: Extract time-series radar image sets from remote sensing deformation monitoring data within the target power grid area, and select image combinations from the time-series radar image sets to form interferometric units; The stage-specific deformation is extracted from each interference unit using differential interferometry. The deformation time series data are obtained by performing time series integration on the deformation variables at each stage.

[0030] In this embodiment, the temporal radar image set refers to the data set composed of multiple synthetic aperture radar single-view complex images acquired continuously at specific time intervals to monitor the deformation of the target power grid area. Differential interferometry is a core data processing method in synthetic aperture radar interferometry. Its technical purpose is to accurately separate the phase component contributed by surface deformation from the total interferometric phase containing multiple phase components. The basic principle of this technique is to generate an interferogram using radar images acquired at different times to obtain the total phase value containing terrain phase, deformation phase, flat land phase, and noise phase. Then, non-deformation phases such as terrain and flat land are removed through precise data processing steps. In the embodiment of this invention, a specific technical route of three-track differential interferometry is adopted. This route does not rely on an external digital elevation model. Instead, it selects three temporally adjacent images from the temporal radar image set and internally combines them into "terrain interferometric pairs" and "deformation interferometric pairs" to self-consistently estimate and remove the terrain phase.

[0031] Specifically, core data for deformation calculation is extracted from remote sensing deformation monitoring data, namely, a time-series radar image set within the target power grid area. Subsequently, image combinations are selected from this time-series radar image set to form interferometric processing units: referencing the spatial baseline between images, three temporally adjacent synthetic aperture radar images are used as a group of interferometric units, where the spatial baseline needs to be controlled within one-third of the critical baseline to maintain good interferometric coherence.

[0032] Taking one interferometric unit as an example, the three images it contains are numbered in chronological order as images ①, ②, and ③. The processing flow is as follows: First, images ① and ② are selected to form a terrain interferometric pair, with image ① as the primary image and image ② as the secondary image. This interferometric pair is precisely registered, with a registration accuracy requirement reaching the sub-pixel level. Next, multi-view processing is performed to suppress speckle noise. Then, an initial interferogram is generated, and the flat-ground effect is removed using precise orbital data. Subsequently, the Goldstein filtering algorithm is applied to filter the interferogram. This algorithm adaptively adjusts the filtering intensity based on the local coherence of the interferogram, effectively suppressing noise while preserving phase details. Finally, regions with coherence higher than 0.3 are selected, and the minimum cost flow algorithm is used for phase unwrapping to obtain the true terrain phase changes.

[0033] Subsequently, images ① and ③ were selected to form a deformation interferometric pair, with image ① as the primary image and image ③ as the secondary image, assuming that minor surface deformation may have occurred during this time period. Then, using the same procedure as for the topographic interferometric pair, the composite interferometric phase of the deformation interferometric pair was obtained: In the formula, These represent the deformation phase, flat terrain phase, atmospheric phase, orbital error phase, topographic phase, and decoherence / thermal noise phase, respectively; among them, the flat terrain phase, atmospheric phase, orbital error phase, and decoherence / thermal noise phase can be removed. Then, differential interferometry is performed, that is, the topographic phase provided by the topographic interferometer is subtracted from the composite interferometric phase. The core calculation formula is: This calculation effectively removes the dominant influence of topographic relief on the phase, allowing the deformation phase to be highlighted.

[0034] Finally, the differentially derived deformation phase is unwrapped to convert it into the true phase value. Since the obtained phase value is a principal value wrapped in the interval [-π, π], the purpose of phase unwrapping is to recover its true, continuous phase value through path integration in space or time. This invention uses a minimum cost flow algorithm for this operation. This algorithm constructs a phase gradient field and solves for the absolute phase under the constraint of minimizing the gradient closure error, thereby obtaining the true, continuous deformation phase value of each pixel in the monitoring area.

[0035] Based on the radar system's operating wavelength (λ, which is approximately 24 cm for the L-band "LuTan-1" satellite) and the basic principles of interferometry, the unwrapped absolute deformation phase value is converted into the staged deformation ΔD of the monitoring unit along the radar line of sight during the time interval from image ① to image ③. The physical conversion formula is: The derivation of this formula stems from the physical fact that the phase changes twice due to the round-trip path of radar waves.

[0036] After traversing all interferometric units and obtaining a series of time-series deformation variables, the deformation variables of each stage are gradually accumulated to form deformation time series data. This data is the absolute benchmark and core input for spatial correlation and risk assessment of geological deformation with power grid facilities in subsequent steps.

[0037] S3: Based on the geographical location of key facilities in the target power grid area, the deformation time series data is converted into deformation characteristic quantities associated with the key facilities; Deformation characteristic quantities refer to a set of indicators extracted from deformation time-series data through spatial statistical analysis, which can quantify the degree of deformation at the geographical location of key facilities. These characteristic quantities are not single values, but a set of statistics describing the deformation field around the facility from different dimensions, mainly including: deformation average, deformation maximum, and deformation gradient. The deformation average is the arithmetic mean of all deformation raster values ​​within a certain range around the facility, reflecting the overall deformation level of the area; the deformation maximum is the largest deformation within the range around the facility, indicating the most severe local deformation; the deformation gradient is the spatial rate of change of deformation around the facility, obtained by calculating the differential of the deformation field in the spatial direction, reflecting the non-uniformity of deformation. Converting deformation time-series data into deformation characteristic quantities involves overlaying deformation raster data with facility location data using spatial analysis techniques. First, an analysis buffer is established centered on the facility coordinates. Then, regional statistical algorithms are used to extract the deformation statistical characteristics within the buffer, ultimately generating a deformation characteristic vector corresponding to each facility.

[0038] Step S3 transforms the macroscopic deformation time-series data covering the entire monitoring area into quantifiable local deformation indicators directly related to the safety of each specific key power grid facility (such as transmission towers and substations). Before the transformation, the deformation time-series data needs to be cleaned, specifically using the sliding window method. Monitoring landslide deformation in mountainous areas with complex terrain can lead to anomalously large errors in the cumulative deformation time-series data due to atmospheric delay effects and phase unwrapping errors. A sliding window is set up, and observations within the sliding window whose difference from the median of the sliding window exceeds three times the Locally Transformed Median Absolute Deviation (SMAD) are considered anomalously large errors. These anomalously large errors in the landslide cumulative deformation time-series data are detected and removed. The median of the sliding window is the median of all observations within the sliding window. The formula for calculating the Locally Transformed Median Absolute Deviation (SMAD) is: In the formula, Let i represent the cumulative deformation observation value at time i, and S represent all cumulative deformation observation values ​​within the sliding window.

[0039] Take a moving window size of m, gradually move it along the time series, calculate the mean deformation within the window, and eliminate the influence of irregular or periodic fluctuations in the time series data. The formula for calculating the mean deformation within the window is: In the formula, t = m, m+1, ..., N, where N represents the number of time series observations. and These represent the actual cumulative deformation observations at times t and tm, respectively. and Let represent the smoothed cumulative deformation at times t and t-1, respectively.

[0040] Preferably, in one embodiment of the present invention, based on the geographical location of key facilities in the target power grid area, deformation time series data is converted into deformation feature quantities associated with the key facilities, including: Based on the coordinate information of the geographical location of the key facilities, determine the spatial image range of the key facilities; Deformation information related to the spatial image range is extracted from deformation time series data, and the deformation information is processed to generate deformation feature quantities for quantifying the degree of deformation of geographical location.

[0041] The coordinate information of the geographical location of key facilities refers to the absolute location data of core equipment in the power grid system (such as transmission towers, substations, and important line nodes) in geographic space. This information comes from the power grid asset management system or the power grid geographic information system and is usually stored in the form of planar coordinates of latitude and longitude. Its positioning accuracy is usually required to be better than 10 meters to ensure the accuracy of spatial analysis. The spatial image range refers to a regular geographical area delineated with the geographical coordinates of the key facilities as the center. This range is used to define the spatial boundary from which analytical data needs to be extracted from the macroscopic deformation field. In this embodiment, this range is preferably defined as a circular buffer zone with a specific radius (usually set to 50 to 100 meters) centered on the facility location. The specific value is determined according to the facility type and its engineering geological conditions. Deformation feature quantity refers to a structured data set used to quantify the degree of deformation at the facility location. It is usually represented in the form of a feature vector, which integrates indicators from multiple dimensions.

[0042] Specifically, based on the geographic coordinates of key facilities obtained from the power grid geographic information system, the spatial image range of each facility is determined. After defining the spatial image range, deformation information related to that range is extracted from the deformation time-series data. This process is achieved through spatial overlay analysis, that is, using the regional statistical tools of the geographic information system, the deformation raster data is overlaid with the buffers of all facilities, thereby extracting the set of all deformation raster values ​​within each buffer at each time point. Subsequently, these extracted deformation information are processed to generate deformation feature quantities used to quantify the degree of deformation of the geographic location of key facilities. Specifically, this involves a series of standardized statistical and spatial analysis calculations, the core of which is to construct a multi-dimensional feature vector, which includes at least the following three key indicators: calculating the arithmetic mean of all deformation raster values ​​within the buffer range as the deformation average reflecting the overall deformation level of the region; identifying the maximum value of the deformation raster within the buffer as a risk signal indicating the most severe local deformation; and calculating the rate of change of the deformation field in the east-west and north-south directions using a spatial difference algorithm and synthesizing the maximum gradient as the deformation gradient characterizing the surface inhomogeneous deformation or potential shear surface. Ultimately, each critical facility corresponds to a deformation characteristic quantity composed of these specific indicators at each point in time.

[0043] S4: Based on the deformation time series data, perform time series alignment processing on the power grid operation monitoring data, and construct a fusion dataset to characterize the comprehensive operational risks of key facilities based on the deformation characteristic quantities and the aligned power grid operation monitoring data. The fused dataset refers to data sorted by the target time series in the row direction and containing all feature variables extracted from deformation data and power grid data in the column direction. Step S4 synchronizes the deformation feature quantities obtained from different sources and at different time frequencies with the power grid operation monitoring data in time to construct a spatiotemporally consistent fused dataset that can characterize the comprehensive operational risks of critical facilities. The technical purpose of this step is to solve the problem that multi-source heterogeneous data cannot be directly used for joint analysis due to different time bases and sampling frequencies, and to provide standardized input for subsequent risk prediction models based on the same time series.

[0044] Preferably, in one embodiment of the present invention, time-series alignment processing of power grid operation monitoring data is performed based on deformation time-series data, including: The target time series is determined based on the time series of deformation time series data; The corresponding values ​​of the power grid operation monitoring data on the target time series are obtained by interpolation, and the corresponding values ​​are paired with the deformation time series data to complete the time series alignment process.

[0045] The target time series refers to a specific set of time points determined based on the aforementioned benchmark, i.e., the precise acquisition time of all satellite images. In this embodiment, the interpolation method refers to a linear interpolation algorithm used to estimate the values ​​of power grid operation monitoring data on the target time series.

[0046] Specifically, the first step is time-series alignment. Since deformation time-series data originates from satellite observations, its time points are determined by the satellite's transit time, and are typically discrete and discontinuous; while power grid operation monitoring data originates from ground sensors, and is typically continuously acquired or recorded at higher frequencies. To achieve accurate fusion of the two, the target time series, i.e., the acquisition time points of all satellite images, must be determined based on the time series of the deformation time-series data. In this embodiment, linear interpolation is used to obtain the corresponding values ​​of the power grid operation monitoring data on the time series of the deformation time-series data. The basic principle is: for each target time point (i.e., the satellite transit time), find the two nearest adjacent data points in the time series of the power grid operation monitoring data, assuming that their values ​​change linearly with time between these two points, and calculate the estimated value of the target time point using a linear function. The calculation formula is: In the formula, It is the target time point (the satellite's transit time). and It is the data of the power grid that is earlier and later than The most recent time point, and These are the power grid parameter values ​​(such as tower tilt angle, stress, etc.) corresponding to these two time points. The calculated y is the power grid parameter at time [time]. The interpolated estimate.

[0047] After interpolation, the corresponding values ​​are paired with the deformation time-series data. Specifically, the deformation characteristic quantity at each satellite transit time (target time point) is combined with the interpolated power grid operation monitoring data at the same time to form a unified multi-dimensional data vector. Each vector represents a snapshot of the coupling between the external geological environmental stress faced by the critical facility and its internal operating state at a specific observation time. Arranging the paired data vectors of all time points in chronological order constructs a fusion dataset characterizing the comprehensive operational risk of the critical facility. This dataset is strictly synchronized with deformation observations in the time dimension and simultaneously includes geological hazard driving factors and facility status response variables in the feature dimension, laying a solid data foundation for accurately predicting the overall risk status of the facility in the next step.

[0048] S5: Input the fused dataset into the constructed time series prediction model to obtain the risk threshold value used to indicate the safety status of critical facilities, and perform risk quantification assessment of critical facilities in the target power grid area based on the risk threshold value. The assessment results are used to adjust the power load of the target power grid area.

[0049] In this context, the risk threshold value refers to the quantitative indicator output by the time-series prediction model, used to directly determine the safety status of critical facilities. Risk quantification assessment is the final output of step S5, a structured assessment conclusion that goes beyond simply determining the risk level. It is a complete data record containing the risk threshold value, the corresponding risk level, the corresponding facility number, and the assessment timestamp. This record is sent to the power grid dispatching system, providing direct, quantitative decision-making basis for subsequent load regulation. Step S5 is the final execution stage of this embodiment. Its technical purpose is to utilize a trained time-series prediction model to make a forward-looking judgment on the future safety status of critical power grid facilities, and based on this judgment, generate executable power grid control instructions, thereby achieving closed-loop management from geological disaster monitoring to proactive power grid protection.

[0050] In one embodiment of the present invention, a Long Short-Term Memory (LSTM) network is preferably used as the basic architecture of the time series prediction model. LSTM is a variant of a recurrent neural network used for time series prediction tasks. Its technical purpose is to solve the gradient vanishing or gradient exploding problems encountered by traditional recurrent neural networks during training, enabling it to effectively learn and memorize the long-distance, complex, nonlinear temporal dependencies between deformation features and power grid operating parameters in the fused dataset.

[0051] In practice, Long Short-Term Memory (LSTM) networks control information flow through three gating units: the forget gate, the input gate, and the output gate. The forget gate uses the sigmoid function to determine the degree of retention of historical information; the input gate, through the collaborative work of the sigmoid and Tanh functions, controls the amount of new information updated; and the output gate uses the sigmoid function to adjust the output of the current state. The network takes the hidden state from the previous time step and the standardized fused data from the current time step as input, updates the cell state through the gating mechanism, and calculates the current output.

[0052] During model training, the root mean square error is used as the loss function, and its calculation formula is as follows: in, This represents the true value of the foundation stress at time t. This represents the model's predicted value, and N is the number of samples. The Adam optimization algorithm is used to iteratively update the network weights, aiming to minimize the loss function, until the model converges.

[0053] The model's prediction accuracy is evaluated by calculating the mean absolute error and root mean square error. The formula for calculating the mean absolute error is: The model is considered trained when the root mean square error (RMSE) on the validation set no longer decreases significantly and the mean absolute error (MAE) is below 5% of the material's yield strength. Inputting the real-time fused dataset into the trained model allows for the output of predicted stress values ​​for critical infrastructure foundations in future periods, providing accurate risk thresholds for power grid safety assessment. This approach ensures both prediction accuracy and meets the reliability requirements of engineering applications.

[0054] Before making predictions, the fused dataset output from step S4 is normalized using the Z-score normalization method. This transforms the data for each feature dimension into a distribution with a mean of 0 and a standard deviation of 1, eliminating the influence of different physical units and accelerating model convergence. Z-score normalization is a data normalization technique performed on each feature dimension of the dataset before inputting the fused dataset into the time-series prediction model for training or prediction. Its purpose is to eliminate the negative impact of different physical units and numerical ranges on model training, transforming the feature data to a uniform order of magnitude, thereby accelerating the model convergence process and improving prediction stability.

[0055] After inputting the processed fusion dataset into the constructed prediction model, the risk of critical facilities is quantitatively assessed based on the output risk threshold.

[0056] Preferably, in one embodiment of the present invention, a risk quantification assessment of critical facilities in a target power grid area based on a risk threshold includes: Compare the risk threshold with the preset risk threshold; Based on the comparison results, the range of the preset risk threshold where the risk threshold is located is determined, the risk level of the critical facility is obtained, and the risk quantification assessment is completed.

[0057] Specifically, the risk threshold refers to the predicted stress value of the critical facility foundation, measured in megapascals (MPa). During the assessment, the risk threshold is compared with preset risk thresholds. These preset risk thresholds are set based on the facility's material properties and safety regulations, and include three key thresholds: the first threshold is 50% of the material's yield strength, the second threshold is 65%, and the third threshold is 80%. The risk thresholds are based on a comprehensive analysis of material mechanics, power industry safety regulations, and historical operating data, with the technical objective of establishing a quantitative safety standard that strictly corresponds to the facility's physical characteristics. Specifically, the first threshold is based on the principle of elastic design. In engineering practice, the allowable stress is usually set far below the yield strength to ensure the structure operates safely within the elastic range. A 50% threshold is a common conservative value in engineering, ensuring the facility is in a fully elastic state. The second threshold follows the principle of engineering safety factor. According to relevant industry standards for overhead transmission line operation, tower components typically need to maintain a safety factor of not less than 2.0. This embodiment sets a threshold of 65%, which not only meets the industry's general requirements for structural safety but also reserves an appropriate safety margin for the system to cope with temporary load fluctuations. The third threshold is based on the plastic deformation early warning mechanism. When the stress reaches this level, the material will begin to enter a significant plastic deformation stage. The setting of this threshold refers to the technical requirements for power equipment condition monitoring, ensuring effective early warning before substantial damage to the facility, and buying valuable time for protective measures. The specific values ​​of each threshold are determined by the formula: Threshold = Material Yield Strength × Percentage Factor, where the material yield strength is determined according to the actual steel grade used in the facility (such as Q235, Q345, etc.). This tiered threshold system establishes quantitative safety standards that strictly correspond to the physical characteristics of the facilities, ensuring the accuracy and reliability of risk assessments.

[0058] The risk level is determined based on the comparison results: when the risk threshold is below the first threshold, it is assessed as a safe level; when it reaches the first threshold but does not exceed the second threshold, it is assessed as a caution level; when it reaches the second threshold but does not exceed the third threshold, it is assessed as a warning level; and when it exceeds the third threshold, it is assessed as a dangerous level. This tiered assessment mechanism enables precise quantification of the risk status of power grid facilities, providing a clear basis for subsequent load regulation decisions.

[0059] After the assessment is completed, the system automatically triggers corresponding countermeasures based on the risk level. When the assessment is at the warning level, an early warning message is sent to the dispatch center and it is recommended to strengthen monitoring; when the assessment is at the danger level, a load adjustment command is automatically generated, and the transmission power of the relevant lines is reduced through the power grid energy management system to control the load rate within a safe range. If necessary, load transfer operations are performed to ensure the safe and stable operation of the power grid.

[0060] Another embodiment of the present invention provides a geological disaster safety risk assessment system for power grid areas. For details, please refer to [link to relevant documentation]. Figure 2 , Figure 2 The geological hazard safety risk assessment system for a power grid area shown in one embodiment of the present invention includes: The acquisition module 11 is used to acquire multi-source monitoring data of the target power grid area. The multi-source monitoring data includes at least remote sensing deformation monitoring data and power grid operation monitoring data. Analysis module 12 is used to perform time series analysis on remote sensing deformation monitoring data and obtain deformation time series data of the target power grid area based on the analysis results; The spatial association module 13 is used to convert deformation time series data into deformation characteristic quantities associated with key facilities based on the geographical location of key facilities in the target power grid area. The time fusion module 14 is used to perform time-series alignment processing on the power grid operation monitoring data based on the deformation time-series data, and to construct a fusion dataset to characterize the comprehensive operational risks of key facilities based on the deformation characteristic quantities and the aligned power grid operation monitoring data. The risk assessment module 15 is used to input the fused dataset into the constructed time series prediction model to obtain the risk threshold value used to indicate the safety status of critical facilities, and to perform risk quantification assessment of critical facilities in the target power grid area based on the risk threshold value. The assessment results are used to adjust the power load of the target power grid area.

[0061] Preferably, in one embodiment of the present invention, the analysis module is further configured to: Extract time-series radar image sets from remote sensing deformation monitoring data within the target power grid area, and select image combinations from the time-series radar image sets to form interferometric units; The stage-specific deformation is extracted from each interference unit using differential interferometry. The deformation time series data are obtained by performing time series integration on the deformation variables at each stage.

[0062] Preferably, in one embodiment of the present invention, the spatial association module is further configured to: Based on the coordinate information of the geographical location of the key facilities, determine the spatial image range of the key facilities; Deformation information related to the spatial image range is extracted from deformation time series data, and the deformation information is processed to generate deformation feature quantities for quantifying the degree of deformation of geographical location.

[0063] Preferably, in one embodiment of the present invention, the time fusion module is further configured to: The target time series is determined based on the time series of deformation time series data; The corresponding values ​​of the power grid operation monitoring data on the target time series are obtained by interpolation, and the corresponding values ​​are paired with the deformation time series data to complete the time series alignment process.

[0064] Preferably, in one embodiment of the present invention, the risk assessment module is further configured to: Compare the risk threshold with the preset risk threshold; Based on the comparison results, the range of the preset risk threshold where the risk threshold is located is determined, the risk level of the critical facility is obtained, and the risk quantification assessment is completed.

[0065] Compared with the prior art, the beneficial effects of the embodiments of the present invention are at least one of the following: (1) This invention acquires remote sensing deformation monitoring data and power grid operation monitoring data, performs time-series analysis on the remote sensing deformation monitoring data to obtain deformation time-series data, and then converts it into deformation characteristic quantities based on the geographical location of key power grid facilities. This data is then fused with time-series aligned power grid operation monitoring data to construct a dataset, which is input into a time-series prediction model to obtain risk threshold values ​​for risk assessment. This process deeply couples geological deformation information with the power grid operation status, overcoming the limitations of isolated analysis of the two types of data in existing technologies, and achieving a more accurate quantitative assessment of the geological hazard threats to power grid facilities.

[0066] (2) This invention constructs a complete technical system from multi-source heterogeneous data collaborative perception and dynamic risk prediction to active regulation of power grid operation, realizing early identification and accurate quantification of geological disaster risks. This method significantly improves the credibility and timeliness of risk assessment results, provides key technical support for power grid dispatching decisions, and effectively enhances the resilience and safe and stable operation level of power grid infrastructure in complex geological environments.

[0067] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for assessing geological hazard safety risks in power grid areas, characterized in that, include: Acquire multi-source monitoring data for the target power grid area, wherein the multi-source monitoring data includes at least remote sensing deformation monitoring data and power grid operation monitoring data; Time-series analysis is performed on the remote sensing deformation monitoring data, and the deformation time-series data of the target power grid area is obtained based on the analysis results; Based on the geographical location of key facilities in the target power grid area, the deformation time series data is converted into deformation feature quantities associated with the key facilities; Based on the deformation time series data, the power grid operation monitoring data is time-aligned, and based on the deformation feature quantity and the aligned power grid operation monitoring data, a fusion dataset is constructed to characterize the comprehensive operational risk of the key facilities. The fused dataset is input into the constructed time-series prediction model to obtain risk thresholds that indicate the safety status of the critical facilities. Based on the risk thresholds, the critical facilities in the target power grid area are subjected to risk quantification assessment, and the assessment results are used to adjust the power load of the target power grid area.

2. The method for assessing geological disaster safety risks in a power grid area as described in claim 1, characterized in that, The step of performing time-series analysis on the remote sensing deformation monitoring data, and obtaining the deformation time-series data of the target power grid area based on the analysis results, includes: Extract the time-series radar image set of the target power grid area from the remote sensing deformation monitoring data, and select image combinations from the time-series radar image set to form interferometric units; The stage-specific deformation is extracted from each of the interference units using differential interferometry. The deformation time series data are obtained by performing time series integration on each of the aforementioned stage deformation variables.

3. The method for assessing geological disaster safety risks in a power grid area as described in claim 1, characterized in that, The step of converting the deformation time-series data into deformation feature quantities associated with the key facilities, based on the geographical location of the key facilities in the target power grid area, includes: Based on the coordinate information of the geographical location of the key facility, the spatial image range of the key facility is determined; Deformation information related to the spatial image range is extracted from the deformation time series data, and the deformation information is processed to generate the deformation feature quantity used to quantify the degree of deformation of the geographical location.

4. The method for assessing geological disaster safety risks in a power grid area as described in claim 1, characterized in that, The step of performing time-series alignment processing on the power grid operation monitoring data based on the deformation time-series data includes: Based on the time series of the deformation time series data, the target time series is determined; The corresponding values ​​of the power grid operation monitoring data on the target time series are obtained by interpolation, and the corresponding values ​​are paired with the deformation time series data to complete the time series alignment process.

5. The method for assessing geological disaster safety risks in a power grid area as described in claim 1, characterized in that, The risk quantification assessment of the critical facilities in the target power grid area based on the risk threshold includes: The risk threshold is compared with a preset risk threshold. Based on the comparison results, the range of the preset risk threshold where the risk threshold is located is determined, the risk level of the key facility is obtained, and the risk quantification assessment is completed.

6. A geological hazard safety risk assessment system for power grid areas, characterized in that, include: The acquisition module is used to acquire multi-source monitoring data of the target power grid area, wherein the multi-source monitoring data includes at least remote sensing deformation monitoring data and power grid operation monitoring data. The analysis module is used to perform time-series analysis on the remote sensing deformation monitoring data, and obtain the deformation time-series data of the target power grid area based on the analysis results; The spatial association module is used to convert the deformation time series data into deformation feature quantities associated with the key facilities, based on the geographical location of the key facilities in the target power grid area. The time fusion module is used to perform time-series alignment processing on the power grid operation monitoring data based on the deformation time-series data, and to construct a fusion dataset to characterize the comprehensive operational risk of the key facilities based on the deformation feature quantity and the aligned power grid operation monitoring data. The risk assessment module is used to input the fused dataset into the constructed time-series prediction model to obtain a risk threshold value that indicates the safety status of the critical facility, and to perform a risk quantification assessment of the critical facility in the target power grid area based on the risk threshold value. The assessment result is used to adjust the power load of the target power grid area.

7. The geological hazard safety risk assessment system for power grid areas as described in claim 6, characterized in that, The analysis module is also used for: Extract the time-series radar image set of the target power grid area from the remote sensing deformation monitoring data, and select image combinations from the time-series radar image set to form interferometric units; The stage-specific deformation is extracted from each of the interference units using differential interferometry. The deformation time series data are obtained by performing time series integration on each of the aforementioned stage deformation variables.

8. The geological disaster safety risk assessment system for power grid areas as described in claim 6, characterized in that, The spatial association module is also used for: Based on the coordinate information of the geographical location of the key facility, the spatial image range of the key facility is determined; Deformation information related to the spatial image range is extracted from the deformation time series data, and the deformation information is processed to generate the deformation feature quantity used to quantify the degree of deformation of the geographical location.

9. A geological hazard safety risk assessment system for power grid areas as described in claim 6, characterized in that, The time fusion module is also used for: Based on the time series of the deformation time series data, the target time series is determined; The corresponding values ​​of the power grid operation monitoring data on the target time series are obtained by interpolation, and the corresponding values ​​are paired with the deformation time series data to complete the time series alignment process.

10. A geological hazard safety risk assessment system for power grid areas as described in claim 6, characterized in that, The risk assessment module is also used for: The risk threshold is compared with a preset risk threshold. Based on the comparison results, the range of the preset risk threshold where the risk threshold is located is determined, the risk level of the key facility is obtained, and the risk quantification assessment is completed.

Images