Power corridor disaster risk assessment method and system based on entropy weight method

Abstract

The invention discloses a power corridor disaster risk assessment method and system based on an entropy weight method, and particularly relates to the technical field of power corridor disaster risk assessment, and the method comprises the steps: obtaining original multi-source point location data, related to disaster risk assessment, of a power corridor region in the southwest region, carrying out the coordinate unification and time alignment, and carrying out the coordinate unification and time alignment; an original multi-source point location data set is obtained and subjected to quality control processing, the consistency and stability of the original multi-source point location data set are checked, missing values and abnormal values are removed, and the multi-source point location data set is output; on the basis of electric power corridor axis segmentation, double-threshold amplitude limiting conditions of an information amount retention rate and an extreme value order are set, a segmentation optimal interpolation kernel and a time window parameter set are constructed, and an interpolation scale and time aggregation granularity are dynamically backtracked and adjusted, so that risk integral offset caused by terrain fracture and window averaging is eliminated, and the power corridor axis segmentation accuracy is improved. And the credibility of the information amount of entropy weight evaluation is ensured.

Description

Technical Field

[0001] This invention relates to the field of disaster risk assessment technology for power corridors, and more specifically, to a method and system for disaster risk assessment of power corridors based on the entropy weight method. Background Technology

[0002] Given the complex terrain and multiple climate zones in Southwest China, the existing process typically converts elevation, slope, vegetation, soil, precipitation, temperature, wind speed, and historical disaster-causing point data into grid data within the power corridor buffer zone, and then performs unified standardization and resampling within a fixed time window. Because ridges and valleys are adjacent, stations are sparse, and resolutions are inconsistent, the interpolation kernel crosses terrain faults and introduces neighboring values, causing high values ​​near the tower base or crossing section to diffuse to adjacent slopes or be diluted by low values. On a time scale, short-duration extreme signals are weakened by averaging over a longer window, resulting in spatial diffusion and temporal weakening. Together, these factors cause a systematic residual in the risk quantity of the integral along the same corridor before and after gridding, which is significantly amplified in sections with obvious terrain changes and uneven observation. Consequently, in the entropy weight calculation stage, diluted indicators are judged as low information content, and diffused indicators are exaggerated as high information content, ultimately changing the shape and level of high-risk zones on the zoning map. This shows that the current technology lacks integral conservation constraints and drift monitoring mechanisms for corridor domains, which makes it impossible to identify and correct the spatial and temporal risk deviations generated during the preprocessing and standardization stages, thus undermining the reliable determination of information content in subsequent entropy weight assessments. Summary of the Invention

[0003] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a method and system for assessing disaster risks in power corridors based on the entropy weight method. By setting dual thresholding conditions of information retention rate and extreme value order on the basis of segmenting the power corridor axis, a set of segmented optimal interpolation kernels and time window parameters is constructed. The interpolation scale and time aggregation granularity are dynamically adjusted backtrackingly, thereby eliminating the risk integral offset caused by terrain fracture and window averaging, and ensuring the credibility of the information in the entropy weight assessment.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a method for assessing the disaster risk of power corridors based on the entropy weight method, characterized in that it includes: S1. Obtain the original multi-source point data related to the power corridor area and disaster risk assessment in Southwest China, unify the coordinates and align the time, obtain the original multi-source point dataset and perform quality control processing, check the consistency and stability of the original multi-source point dataset and remove missing values ​​and outliers, and output the multi-source point dataset. S2. Based on the multi-source point dataset, the facility coordinates are connected sequentially to construct the power corridor axis. On the power corridor axis, the tangential direction is determined according to the direction of the connection between adjacent facility coordinates, and the normal direction is determined according to the direction perpendicular to the tangential direction. The system is then segmented according to preset rules, and the segmented point dataset is output. S3. For each segment of data, establish reference conditions for the distribution characteristics and extreme value sorting of historical disaster-causing records within the segment of data, and set the information retention rate threshold and the relative order of peak values ​​as dual-threshold conditions. S4. Based on the distribution of facility coordinate spacing, the half-width of the interpolation kernel is determined, and a candidate set of interpolation kernel half-widths is obtained. A candidate set of time window lengths is constructed based on the duration of meteorological events. The optimal interpolation kernel half-width and time window length are selected under the condition of double gating using the safe trust region Bayesian optimization, and the piecewise optimal parameter set is output. S5. Based on the segmented optimal parameter set, interpolate the multi-source point dataset, convert the point data into grid data and aggregate it according to the corresponding time window length to obtain a standardized evaluation index set, and perform entropy weight method calculation to generate disaster risk assessment values, and summarize them to form a disaster zoning map of the power corridor.

[0005] In a preferred embodiment, in S1, raw multi-source location data related to the disaster risk assessment of the power corridor area in Southwest China are acquired. The raw multi-source location data includes geographic location data, infrastructure location data and meteorological back-calculation location data. Spatial reference system identifier resolution is performed on the original multi-source point data to extract the projection coordinate parameters of the original multi-source point data. Then, with the preset unified coordinates as the target, coordinate transformation operation is performed to unify the spatial coordinates of all the original multi-source point data to the preset unified coordinates and output the original multi-source point data under the unified coordinate system. Read the time attribute field of each original multi-source point data under a unified coordinate system, extract the timestamps of various types of original multi-source point data, sort the extracted timestamps in chronological order, and output the sorted timestamp sequence. Based on the sorted timestamp sequence, the lower limit time interval between adjacent timestamps is determined, and a unified time axis covering the entire research time range is constructed using the lower limit time interval as the step size. For raw multi-source point data with sampling frequency lower than the unified time axis, an interpolation method is used. The number of points before and after the missing point is taken as the benchmark. The ratio of the time interval between the missing point and the previous known point to the total time interval between the two known points is used as the scaling factor. Starting from the number of points of the previous known point, the product of the scaling factor and the difference in point values ​​between the two known points is superimposed to obtain the point value of the missing point. The low-frequency point data after interpolation is output. For original multi-source point data with a sampling frequency higher than the unified time axis, the sliding window width is an integer multiple of the lower limit time interval. The sliding window is used to slide sequentially on the original multi-source point data with a sampling frequency higher than the unified time axis. The average value of the point data in each sliding window is calculated, and the average value is used as the representative value of the corresponding time point of the corresponding sliding window. The high-frequency point data after the sliding average is output. The low-frequency and high-frequency point data are realigned according to the time nodes of a unified time axis, and the original multi-source point dataset aligned with the unified time axis is output.

[0006] In a preferred embodiment, S1 further includes performing a consistency check on the original multi-source point dataset aligned with a unified time axis, extracting the unit system information of different point data for the same physical quantity, converting point values ​​with non-uniform units according to a preset International System of Units (SI), and outputting point data with consistent units. The numerical range of point data with consistent units is checked, and a reasonable threshold range of physical quantities is called. When the point value exceeds the reasonable threshold range, it is marked as abnormal and removed. The point data that has passed the threshold test is output. The spatial distance and time interval between adjacent points of the data detected by the threshold are compared, and the change in unit distance or unit time is calculated. When the change exceeds the preset reasonable fluctuation range, it is judged as an unreasonable observation, the corresponding point is marked as abnormal and removed, and the point data after spatiotemporal reasonableness comparison is output. The point data that has passed the unit unification, threshold detection and spatiotemporal rationality comparison are summarized and the point dataset after consistency verification is output. The stability of the point dataset after consistency verification is checked. A fixed-length sliding window is constructed on the time series. The difference between the value of adjacent points within the fixed-length sliding window is calculated. When the difference exceeds a preset threshold, the corresponding point record is determined to be unstable and removed. The point dataset after stability verification is output. Missing values ​​are identified and processed in the point dataset after stability testing. Continuous missing intervals of point data are detected on a unified time axis, and the length of continuous missing intervals is compared with a preset threshold. When the length of consecutive missing values ​​is less than or equal to a preset threshold, interpolation is used to fill in the missing values. When the length of consecutive missing values ​​is greater than the preset threshold, the corresponding data points are directly removed, and a multi-source dataset for parameter optimization and power corridor disaster risk assessment is output.

[0007] In a preferred embodiment, in S2, based on the multi-source point dataset, the coordinates of the facilities are read according to the arrangement order of the facilities on the power corridor, and the coordinates of adjacent facilities are connected in sequence to form a polyline composed of the lines connecting adjacent facilities, which serves as the central axis of the power corridor. For each segment of the central axis of the power corridor, a local coordinate system is established with the direction of the line connecting the start and end points as the tangential direction and the direction perpendicular to the tangential direction as the normal direction. Using the central axis of the power corridor as a reference line, and combining the tangential and normal directions, boundary lines are generated on both sides of the central axis at a preset equidistant offset distance, and the area enclosed by the central axis and the two boundary lines is taken as the permissible spatial domain of the power corridor. The permissible spatial domain of the power corridor is segmented according to preset rules, and the point datasets in each segment are aggregated to output the segmented point datasets.

[0008] In a preferred embodiment, in S3, for the precipitation, temperature and wind speed in the segmented point data, the corresponding historical disaster-causing records are called, and the values ​​of each meteorological element at the time of disaster and at the time of non-disaster are statistically analyzed to form disaster-triggered samples and non-disaster samples. The upper and lower limits of the observed values ​​of each meteorological element in the disaster-triggered sample and the non-disaster sample are identified respectively as the numerical interval. The observed values ​​of the corresponding meteorological elements within the numerical interval are added together and divided by the number of observations to obtain the average value. The numerical interval and the average value together constitute the distribution characteristics. Using the upper limit of the observed value of each meteorological element in the disaster-triggered sample as a reference threshold, a table of correspondence between each meteorological element and the disaster type is generated. In the correspondence table, the proportion of disaster-triggered samples to the total number of samples is used as the information retention rate threshold, and the order of appearance of the upper limit of disaster-triggered samples is used as the peak relative order constraint. The information retention rate threshold and the peak relative order constraint together constitute a double threshold condition.

[0009] In a preferred embodiment, in S4, based on the spatial spacing between the coordinates of adjacent facilities in the segmented point data, all spatial spacing between adjacent facilities is collected to form a spacing sample set. The lower limit and upper limit of the spacing sample set are statistically analyzed to determine the distribution range of the spacing sample set. The mean of the spatial spacing within the distribution range is used as the candidate value of the interpolation kernel half-width to form an interpolation kernel half-width candidate set. Based on the duration of historical meteorological events corresponding to segmented point data, the start and end times of each meteorological event are identified and the duration of a single meteorological event is calculated. The duration of all meteorological events is collected to form a duration sample set. The lower and upper limits of the duration sample set are statistically determined to determine the distribution range of the duration sample set. The mean of the duration within the distribution range is used as a candidate value for the time window length, forming a candidate set of time window lengths.

[0010] In a preferred embodiment, S4 further includes inputting the interpolation kernel half-width candidate set and the time window length candidate set into the secure trust domain Bayesian optimization architecture. The secure trust domain Bayesian optimization architecture includes a prediction distribution model that predicts the relationship between the parameter combination and the degree of satisfaction of the double gate limiting condition, and a secure boundary condition for limiting the range of parameter combination values. The average value of the interval sample set is used as the initial interpolation kernel half-width, and the average value of the meteorological event duration sample set is used as the initial time window length to form the initial parameter combination. The corresponding information retention rate and peak relative order are calculated, and the calculation results are used as the objective function value to input the prediction distribution model. In the predictive distribution model, the initial parameter combination is used as the benchmark, and the corresponding spatial spacing distribution range and meteorological event duration range are used as the search range. Several candidate parameter combinations are divided within the corresponding search range according to the preset step size. Calculate the expected objective function value for each candidate parameter combination, compare the candidate parameter combinations according to the magnitude of the expected objective function value, and select the candidate parameter combination with the upper limit of the expected objective function value as the new parameter combination; The information retention rate and peak relative order under the new parameter combination are calculated, and the calculation results are used as the new objective function to re-input into the prediction distribution model, triggering the calculation of the new parameter combination, eliminating candidate parameter combinations whose expected objective function value is lower than the preset threshold, narrowing the search range of parameter combinations, and completing the correction of the prediction distribution model. Repeat the correction process in the predicted distribution model until the calculated parameter combination no longer changes. Then, determine the corresponding parameter combination as the piecewise optimal parameter combination and output the piecewise optimal parameter set.

[0011] In a preferred embodiment, in S5, the interpolation kernel half-width of the segmented optimal parameter set is called, and the difference method is used to expand and extrapolate the data of each point on the central axis in space, and fill the missing positions along the normal direction to obtain grid data covering the allowable domain of the power corridor space. Call the time window length in the optimal parameter set of segments, divide the continuous time window on the unified time axis, take the values ​​corresponding to the grid data falling into the same time window as the candidate set of the time window, perform an arithmetic mean on all values ​​in the candidate set, take the average as the aggregate representative value of the corresponding time window, and output the grid aggregate data under the corresponding time window. The values ​​of geographic location data, infrastructure location data and meteorological back-calculation location data in the grid aggregated data are used as indicators, and normalization processing is performed to eliminate the differences in the numerical range between different types of indicators, and output a standardized set of evaluation indicators. The weight of each standardized evaluation index in the standardized evaluation index set is calculated based on the entropy weight method. The weight of each index is multiplied by the corresponding standardized index value and accumulated to obtain the disaster risk evaluation value of the corresponding segment, and the segmented disaster risk sequence of the power corridor is output. The disaster risk assessment values ​​of each segment within the segmented disaster risk sequence are bound to their corresponding geographical coordinates and spliced ​​together within the permissible spatial domain of the power corridor to form a risk distribution sequence. Based on preset risk level thresholds, the risk distribution sequence is divided into multiple risk level intervals, and displayed in partitions corresponding to different risk levels to generate a disaster zoning map of the power corridor.

[0012] A disaster risk assessment system for power corridors based on the entropy weight method includes a data quality control module, an axis construction module, a fidelity control module, a parameter optimization module, and a zoning output module. The data quality control module is used to acquire the original multi-source point data related to the power corridor area and disaster risk assessment in Southwest China, perform coordinate unification and time alignment, obtain the original multi-source point dataset and perform quality control processing, check the consistency and stability of the original multi-source point dataset, remove missing values ​​and outliers, and output the multi-source point dataset. The axis construction module is based on a multi-source point dataset. It connects the facility coordinates in sequence to construct the power corridor axis. On the power corridor axis, the tangential direction is determined according to the direction of the connection between adjacent facility coordinates, and the normal direction is determined according to the direction perpendicular to the tangential direction. The module is then segmented according to preset rules and outputs a segmented point dataset. The fidelity control module is used to establish reference conditions for the distribution characteristics and extreme value sorting of historical disaster-causing records within each segment of data, and to set the information retention rate threshold and the relative order of peak values ​​as dual-gate limiting conditions. The parameter optimization module determines the half-width of the interpolation kernel based on the distribution of facility coordinate spacing, obtains a candidate set of interpolation kernel half-widths, and constructs a candidate set of time window lengths based on the duration of meteorological events. It then uses safe trust region Bayesian optimization to select the optimal interpolation kernel half-width and time window length under double gating conditions, and outputs the piecewise optimal parameter set. The zoning output module interpolates the multi-source point dataset based on the segmented optimal parameter set, converts the point data into grid data, aggregates it according to the corresponding time window length, obtains a standardized evaluation index set, performs entropy weight calculation, generates disaster risk assessment values, and summarizes them to form a disaster zoning map of the power corridor.

[0013] The technical effects and advantages of this invention are as follows: This scheme sets dual-gate limiting conditions and constructs a segmented optimal parameter set to dynamically optimize the interpolation kernel half-width and time window length, ensuring the stability of spatial distribution and extreme value sorting. It solves the risk integral drift and level misjudgment caused by interpolation mismatch and window averaging, and improves the accuracy and consistency of disaster zoning maps. A parameter candidate set is constructed by utilizing facility spacing and weather duration. Under double threshold constraints, a safe trust region Bayesian optimization method is adopted to dynamically adjust the interpolation kernel and time window, thereby improving the retention capacity of short-term extreme events and the accuracy of interpolation response, and adapting to the spatiotemporal differences under multiple terrain changes in the power corridor. Geographic location, infrastructure and meteorological back-calculation data are uniformly aligned, time-registered and quality-controlled, and then standardized after segmented aggregation to ensure the uniformity of multiple indicators in terms of dimensions, scale and distribution characteristics, forming a complete set of standardized evaluation indicators. Based on the facility coordinates, the power corridor axis is constructed, the tangential and normal directions are extracted to form the spatial allowable domain, and functional segmentation is performed in combination with the terrain and facility attributes to provide a structural boundary basis for the construction of interpolation paths, thereby improving the disaster signal response and feature analysis capabilities in areas of spatial discontinuity and terrain fracture. Attached Figure Description

[0014] Figure 1 This is a flowchart outlining the method steps of the present invention; Figure 2 This is a schematic diagram of the system module structure of the present invention; Figure 3 This is the original data processing flowchart of the present invention; Figure 4 This is a flowchart of the spatial segmentation construction process of the present invention; Figure 5 This is a flowchart of the parameter optimization and threshold determination process of the present invention; Figure 6 This is a flowchart illustrating the standardization of indicators and risk output for this invention. Detailed Implementation

[0015] 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. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Refer to the instruction manual appendix Figure 1-6 An embodiment of the present invention provides a method for assessing the disaster risk of power corridors based on the entropy weight method, comprising: S1. Obtain the original multi-source location data related to the disaster risk assessment of the power corridor in Southwest China. The original multi-source location data includes geographic location data, infrastructure location data and meteorological back-calculation location data. Perform coordinate unification and time alignment, output the original multi-source location dataset for disaster risk assessment of the power corridor, and perform quality control processing to check the consistency and stability of the original multi-source location dataset and remove missing values ​​and outliers. Output the multi-source location dataset for parameter optimization and disaster risk assessment of the power corridor. S2. Based on the multi-source point dataset, the facility coordinates are connected sequentially to construct the power corridor axis. On the power corridor axis, the tangential direction is determined according to the direction of the connection between adjacent facility coordinates, and the normal direction is determined by the direction perpendicular to the tangential direction. The permissible spatial domain of the power corridor is generated and segmented according to preset rules. The segmented point dataset is output. The segmented point dataset is used for segmented calculation and feature extraction of power corridor disaster risk assessment. S3. For each segment of data, establish reference conditions for the distribution characteristics and extreme value ranking of precipitation, temperature, wind speed and corresponding historical disaster records within the segment of data. Set the information retention rate threshold and the relative order of peak values ​​as dual-gate limiting conditions. The dual-gate limiting conditions are used for data fidelity control in the disaster risk assessment of power corridors. S4. Based on the distribution of facility coordinate spacing, the half-width of the interpolation kernel is determined, and a candidate set of interpolation kernel half-widths is obtained. Based on the distribution of the duration of the corresponding meteorological events, a candidate set of time window lengths is constructed. The optimal interpolation kernel half-width and time window length are selected under the double-thresholding condition using the secure trust region Bayesian optimization. When the information retention rate is detected to be lower than the preset threshold or the relative order of peak values ​​is changed, the interpolation kernel half-width is shrunk or the time window is shortened and the optimization is re-entered until the double-thresholding condition is met. The piecewise optimal parameter set that passes the double-thresholding condition is output. The piecewise optimal parameter set is used for gridding and standardization processing of power corridor disaster risk assessment. S5. Based on the segmented optimal parameter set, interpolate the multi-source point dataset to convert the point data into grid data spatially distributed along the power corridor and aggregate it according to the corresponding time window length to obtain a standardized evaluation index set including geographic point data, infrastructure point data and meteorological back-calculation point data. Then, perform entropy weight method calculation to generate disaster risk assessment values ​​for each power corridor segment and summarize them to form a disaster zoning map of the power corridor.

[0017] In S1, raw multi-source location data related to the disaster risk assessment of the power corridor area in Southwest China are obtained. The raw multi-source location data includes geographic location data, infrastructure location data and meteorological back calculation location data. The process involves resolving the spatial reference system identifier of the original multi-source point data, extracting the projected coordinate parameters, and simultaneously reading the central meridian, projection zone number, ellipsoid parameters, and coordinate units from the header or metadata of the original multi-source point data. Using a preset unified coordinate system as the target, coordinate transformation operations are performed to unify the spatial coordinates of all original multi-source point data to the preset unified coordinate system, outputting the original multi-source point data in a unified coordinate system. Here, the spatial reference system identifier refers to the coordinate description information carried by the original multi-source point data; the coordinate transformation operation refers to performing affine transformations or geodetic transformations on the latitude and longitude or projected coordinates of the original multi-source point data based on the projection method and ellipsoid parameters determined by the spatial reference system identifier resolution, so that the spatial coordinates of various point data are unified to the same Cartesian coordinate system or unified geodetic coordinate system. Read the time attribute field of each original multi-source point data under a unified coordinate system, extract the timestamps of various types of original multi-source point data, sort the extracted timestamps in chronological order, and output the sorted timestamp sequence. The lower limit time interval between adjacent timestamps is determined based on the sorted timestamp sequence, and a unified time axis covering the entire research time range is constructed using this lower limit time interval as the step size. For raw multi-source point data with sampling frequencies lower than the unified time axis, an interpolation method is used. The number of points before and after the missing point is taken as the benchmark. The ratio of the time interval between the missing point and the previous known point to the total time interval between the two known points is used as the scaling factor. Starting from the number of points of the previous known point, the product of the scaling factor and the difference in point values ​​between the two known points is added to obtain the point value of the missing point. The low-frequency point data with interpolation is then output. The value refers to the meteorological element value corresponding to the point data at a certain moment under the unified time axis, including known values ​​directly obtained from observation records and missing point values ​​calculated by interpolation methods. For original multi-source point data with a sampling frequency higher than the unified time axis, the sliding window width is an integer multiple of the lower limit time interval. The sliding window is used to slide sequentially on the original multi-source point data with a sampling frequency higher than the unified time axis. The average value of the point data in each sliding window is calculated, and the average value is used as the representative value of the corresponding time point of the corresponding sliding window. The high-frequency point data after the sliding average is output. The low-frequency and high-frequency point data are realigned according to the time nodes of a unified time axis, and the original multi-source point dataset aligned with the unified time axis is output.

[0018] S1 also includes performing consistency checks on the original multi-source point dataset aligned with a unified time axis, extracting unit system information of different point data for the same physical quantity, converting point values ​​with non-uniform units according to a preset International System of Units (SI), and outputting point data with consistent units. The unit system information refers to the measurement unit and system of measurement of the physical quantity recorded in the point data, such as temperature in Fahrenheit and Celsius, and wind speed in miles per hour and meters per second. The SI conversion refers to converting point values ​​with non-uniform units into standard numerical representations under the International System of Units (SI) according to a preset conversion formula, such as converting Fahrenheit temperature to Celsius temperature and miles per hour to meters per second. The numerical range of point data with consistent units is checked, and the reasonable threshold range of the physical quantity is called. When the point value exceeds the reasonable threshold range, it is marked as abnormal and removed. The point data that has passed the threshold test is output. The spatial distance and time interval between adjacent points in the threshold-detected point data are compared to calculate the change amplitude per unit distance or per unit time. When the change amplitude exceeds the preset reasonable fluctuation range, it is judged as an unreasonable observation, the corresponding point is marked as abnormal and removed, and the point data after spatiotemporal reasonableness comparison is output. The change amplitude per unit time is the numerical difference between adjacent points divided by the time interval; the change amplitude per unit distance is the numerical difference between adjacent points divided by the spatial distance. The point data that has passed the unit unification, threshold detection and spatiotemporal rationality comparison are summarized and the point dataset after consistency verification is output. The stability of the point dataset after consistency verification is checked. A fixed-length sliding window is constructed on the time series. The difference between the value of adjacent points within the fixed-length sliding window is calculated. When the difference exceeds a preset threshold, the corresponding point record is determined to be unstable and removed. The point dataset after stability verification is output. Missing values ​​are identified and processed in the point dataset after stability testing. Continuous missing intervals of point data are detected on a unified time axis, and the length of continuous missing intervals is compared with a preset threshold. When the length of consecutive missing data is less than or equal to a preset threshold, the missing values ​​are filled by interpolation. When the length of consecutive missing data is greater than the preset threshold, the corresponding data points are directly removed. The multi-source data point dataset is output for parameter optimization and power corridor disaster risk assessment. The interpolation method is the same as that used for interpolation of low-frequency data points, but it is used to correct occasional gaps caused by observation anomalies.

[0019] In S2, based on the multi-source point dataset, the coordinates of the facilities are read according to the arrangement order of the facilities on the power corridor, and the coordinates of adjacent facilities are connected in sequence to form a polyline composed of the lines connecting adjacent facilities, which serves as the central axis of the power corridor. For each segment of the central axis of the power corridor, a local coordinate system is established with the direction of the line connecting the start and end points of the segment as the tangential direction and the direction perpendicular to the tangential direction as the normal direction, to ensure the geometric definition of the lateral spatial constraints of the power corridor. Using the central axis of the power corridor as a reference line, and combining the tangential and normal directions, boundary lines are generated on both sides of the central axis at a preset equidistant offset distance, and the area enclosed by the central axis and the two boundary lines is taken as the permissible spatial domain of the power corridor. The permissible spatial domain of the power corridor is segmented according to preset rules. The point datasets in each segment are aggregated and output as segmented point datasets. The preset rules refer to dividing the permissible spatial domain of the power corridor into multiple segmented areas with clear boundaries based on differences in facility type, geographical environment, and historical disaster distribution. This ensures that each segmented area maintains relative consistency in facility attributes, geographical features, or disaster features, facilitating segmented calculations and feature extraction in the disaster risk assessment of the power corridor.

[0020] In S3, for the precipitation, temperature and wind speed in the segmented point data, the corresponding historical disaster-causing records are called, and the values ​​of each meteorological element at the time of disaster and the values ​​at the time of non-disaster are statistically analyzed to form disaster-triggered samples and non-disaster samples. The upper and lower limits of the observed values ​​of each meteorological element in the disaster-triggered sample and the non-disaster sample are identified separately as numerical intervals. The observed values ​​of the corresponding meteorological elements within the numerical intervals are added together and divided by the number of observations to obtain the observed average value. The numerical intervals and the observed average value together constitute the distribution characteristics. The observed value refers to the set of actual values ​​of meteorological elements recorded by the point data in the disaster-triggered sample and the non-disaster sample, which is used to statistically analyze the distribution characteristics such as the upper limit value, the lower limit value, and the average value. Using the upper limit of the observed value of each meteorological element in the disaster-triggered sample as a reference threshold, a table of correspondence between each meteorological element and the disaster type is generated. In the correspondence table, the proportion of disaster-triggered samples to the total number of samples is used as the information retention rate threshold, and the order of appearance of the upper limit of disaster-triggered samples is used as the peak relative order constraint. The information retention rate threshold and the peak relative order constraint together constitute a double threshold condition.

[0021] In S4, based on the spatial spacing between the coordinates of adjacent facilities in the segmented point data, all spatial spacing between adjacent facilities is collected to form a spacing sample set. The lower limit and upper limit of the spacing sample set are statistically analyzed to determine the distribution range of the spacing sample set. The mean of the spatial spacing within the distribution range is used as the candidate value of the interpolation kernel half-width to form the interpolation kernel half-width candidate set. Based on the duration of historical meteorological events corresponding to segmented point data, the start and end times of each meteorological event are identified and the duration of a single meteorological event is calculated. The durations of all meteorological events are collected to form a duration sample set. The lower and upper limits of the duration sample set are statistically determined to define the distribution range of the duration sample set. The mean of the duration within the distribution range is used as a candidate value for the time window length, forming a candidate set of time window lengths. Here, a meteorological event refers to a disastrous weather process composed of continuous changes in meteorological elements.

[0022] S4 also includes inputting the candidate set of interpolation kernel half-width and the candidate set of time window length into the safe trust region Bayesian optimization architecture. The safe trust region Bayesian optimization architecture includes a predictive distribution model that predicts the relationship between the parameter combination and the degree of satisfaction of the double-gating condition, and a safe boundary condition for limiting the range of parameter combination values. The safe boundary condition refers to the value restriction set based on the upper and lower limits formed by the distribution of facility coordinate spacing and the distribution of meteorological event duration, which is used to avoid the parameter combination from exceeding the reasonable calculation range. The parameter combination refers to the two-dimensional parameter vector composed of the interpolation kernel half-width and the time window length. The average value of the interval sample set is used as the initial interpolation kernel half-width, and the average value of the meteorological event duration sample set is used as the initial time window length to form the initial parameter combination. The corresponding information retention rate and peak relative order are calculated, and the calculation results are used as the objective function value to input the prediction distribution model. In the predictive distribution model, the initial parameter combination is used as the benchmark, and the corresponding spatial spacing distribution range and meteorological event duration range are used as the search range. Several candidate parameter combinations are divided within the corresponding search range according to the preset step size. The expected objective function value is calculated for each candidate parameter combination, and the candidate parameter combinations are compared according to the magnitude of the expected objective function value. The candidate parameter combination with the upper limit of the expected objective function value is selected as the new parameter combination. The expected objective function value refers to the predicted value obtained by the prediction distribution model based on the relative difference between the objective function value of the initial parameter combination and the candidate parameter combination. It is used to characterize the strength of the candidate parameter combination that may satisfy the double threshold condition. The information retention rate and peak relative order under the new parameter combination are calculated, and the calculation results are used as the new objective function to re-input into the prediction distribution model, triggering the calculation of the new parameter combination, eliminating candidate parameter combinations whose expected objective function value is lower than the preset threshold, narrowing the search range of parameter combinations, and completing the correction of the prediction distribution model. Repeat the correction process in the predicted distribution model until the calculated parameter combination no longer changes. Then, determine the corresponding parameter combination as the piecewise optimal parameter combination and output the piecewise optimal parameter set.

[0023] It should be noted that in the formula structure involved in this scheme, dimensionless terms can be used as proportional or structural adjustment factors. When combined with quantities with units, they only play a role in numerical scaling and do not introduce new physical dimensions. Therefore, they will not change or confuse the overall unit system. This combination of "dimensionless terms and terms with units" can be understood as a composite structural expression commonly used in mathematical physics modeling. It conforms to the principle of dimensional consistency and has a clear physical interpretation basis. Secondly, in the formula structure of this scheme, if multiple variables with different physical units are involved, including but not limited to time, mass or energy variables, their joint appearance is to express the collaborative modeling relationship of multiple physical mechanisms. Each variable can form a unified structure through function mapping, ratio combination or normalization adjustment, with clear units and clear meaning. The overall expression conforms to the principle of dimensional consistency and the conventional formula of engineering modeling. In this solution, constants, weights, adjustment factors, threshold parameters, proportional coefficients, etc., are all adjustable control parameters for different application environments. Their values ​​depend on the target equipment configuration, data input characteristics, and performance optimization goals. During the implementation phase, they are set to converge within a reasonable range through model verification, performance constraints, or engineering calibration. Although these parameters do not have a unique preset value, they have clear adjustment logic and calculation paths. They belong to the deterministic setting process in engineering implementation. The purpose of this setting is to ensure that the solution is both universally adaptable and reproducible and operable, without affecting its technical clarity and feasibility. In S5, the interpolation kernel half-width of the segmented optimal parameter set is called, and the difference method is used to expand and extrapolate the data of each point on the central axis in space. The missing positions are filled along the normal direction to obtain grid data covering the allowable domain of the power corridor space. Call the time window length in the optimal parameter set of segments, divide the continuous time window on the unified time axis, take the values ​​corresponding to the grid data falling into the same time window as the candidate set of the time window, perform an arithmetic mean on all the values ​​in the candidate set, take the average as the aggregate representative value of the corresponding time window, and output the grid aggregate data under the corresponding time window. The values ​​of geographic location data, infrastructure location data and meteorological back-calculation location data reflected in the grid aggregated data are used as indicators. Normalization processing is performed to eliminate the differences in the numerical range between different types of indicators and output a standardized evaluation indicator set. The normalization processing refers to subtracting the lower limit value of each type of indicator from the value of each type of indicator, and then dividing by the difference between the upper limit value and the lower limit value, so as to convert the result into a dimensionless value between zero and one, which is used to eliminate the differences in the numerical range between different indicators. The weight of each standardized evaluation index in the standardized evaluation index set is calculated based on the entropy weight method. The weight of each index is multiplied by the corresponding standardized index value and accumulated to obtain the disaster risk evaluation value of the corresponding segment, and the segmented disaster risk sequence of the power corridor is output. Define disaster risk assessment value : in Indicates the first The global weight value of each evaluation indicator. It reflects the distinguishing ability and information content of the corresponding indicator across all segments; high information entropy corresponds to low weight. Indicates the first The segment in the first Standardized index values ​​under each evaluation indicator The values ​​derived from the interpolated and aggregated grid data are transformed through normalization. Dimensionless values ​​within an interval; Indicates the total number of evaluation indicators; Furthermore, in In the formula, the first Global weight values ​​of each evaluation indicator With the The segment in the first Multiply the standardized index values ​​under the first evaluation index to obtain the second... Disaster risk assessment values ​​for each segment ; in Indicates the first The information entropy value of each evaluation index is the information entropy value across all segments. The larger the information entropy value, the more dispersed the information. This represents the loop variable of the indicators during the normalization process, used to iterate through all evaluation indicators. Indicates the first The information entropy value of each evaluation index across all segments; Furthermore, in In the formula, the first The information entropy value of each evaluation indicator across all segments through The result of the interpolation operation is divided by the sum of the results of the same interpolation operation performed on all evaluation indicators to obtain the result of the first interpolation operation. Global weight values ​​of each evaluation indicator ; in Indicates the first The first segment The proportion of each evaluation indicator in all segments; Represent the natural logarithm function; This indicates the total number of segments in the spatial division of the power corridor; The constant coefficients representing the normalized entropy value Used to ensure The range is in interval; Furthermore, in In the formula, the first The first segment The proportion of each evaluation indicator in all segments With logarithm Multiply by summing the product over all segments, then multiply by the negative constant coefficient of the normalized entropy value. , obtained the The information entropy value of each evaluation indicator across all segments ; in This represents a segmented loop variable used to iterate through all segments. Indicates the first The segment in the first Standardized indicator values ​​under each evaluation indicator; Furthermore, in In the formula, the first The segment in the first Standardized index values ​​under each evaluation indicator , and all segments in the The ratio of the sum of the standardized index values ​​under the evaluation index is used to obtain the i-th evaluation index. The first segment The proportion of each evaluation indicator in all segments ; Furthermore, in In the formula, the total number of segments into which the power corridor space is divided is denoted. Substituting the natural logarithm function Take the reciprocal of the logarithm to obtain the constant coefficient of the normalized entropy value. ; The disaster risk assessment values ​​of each segment within the segmented disaster risk sequence are bound to their corresponding geographical coordinates and spliced ​​together within the permissible spatial domain of the power corridor to form a risk distribution sequence. Based on preset risk level thresholds, the risk distribution sequence is divided into multiple risk level intervals and displayed in partitions corresponding to different risk levels, generating a disaster zoning map of the power corridor. The disaster zoning map of the power corridor serves as a result of the disaster risk assessment of the power corridor.

[0024] A disaster risk assessment system for power corridors based on the entropy weight method includes a data quality control module, an axis construction module, a fidelity control module, a parameter optimization module, and a zoning output module. The data quality control module is used to acquire the original multi-source point data related to the power corridor area and disaster risk assessment in Southwest China, perform coordinate unification and time alignment, obtain the original multi-source point dataset and perform quality control processing, check the consistency and stability of the original multi-source point dataset, remove missing values ​​and outliers, and output the multi-source point dataset. The axis construction module is based on a multi-source point dataset. It connects the facility coordinates in sequence to construct the power corridor axis. On the power corridor axis, the tangential direction is determined according to the direction of the connection between adjacent facility coordinates, and the normal direction is determined according to the direction perpendicular to the tangential direction. The module is then segmented according to preset rules and outputs a segmented point dataset. The fidelity control module is used to establish reference conditions for the distribution characteristics and extreme value sorting of historical disaster-causing records within each segment of data, and to set the information retention rate threshold and the relative order of peak values ​​as dual-gate limiting conditions. The parameter optimization module determines the half-width of the interpolation kernel based on the distribution of facility coordinate spacing, obtains a candidate set of interpolation kernel half-widths, and constructs a candidate set of time window lengths based on the duration of meteorological events. It then uses safe trust region Bayesian optimization to select the optimal interpolation kernel half-width and time window length under double gating conditions, and outputs the piecewise optimal parameter set. The zoning output module interpolates the multi-source point dataset based on the segmented optimal parameter set, converts the point data into grid data, aggregates it according to the corresponding time window length, obtains a standardized evaluation index set, performs entropy weight calculation, generates disaster risk assessment values, and summarizes them to form a disaster zoning map of the power corridor.

[0025] It should be noted that, including but not limited to: This solution addresses the spatial diffusion and temporal attenuation problems of existing power corridor risk assessment methods under complex terrain and multi-climate conditions. Current methods are prone to introducing non-true neighborhood values ​​and averaging extreme signals during the interpolation and resampling stages, resulting in a systematic shift in the risk integral along the route. This shift is particularly significant in areas with abrupt terrain changes and uneven observation, which significantly affects the effectiveness of entropy weight calculation and leads to distortion of high-risk zoning results. To address the above issues, six closely linked execution steps were designed: The first step involves performing unit unification, time alignment, and missing and outlier removal operations on the original observation datasets covering geographic location data, infrastructure location data, and meteorological back-calculation location data within the power corridor area. This ensures that the multi-source observation data has a processing foundation under a unified scale, thus avoiding processing bias caused by structural differences from the source. The second step is to construct a normal calculation path based on the central axis of the power corridor, and divide the power corridor into multiple segmented areas according to the terrain change intervals and the differences in infrastructure functions. Instead of using fixed interpolation kernel parameters for each segmented area, a candidate parameter set consisting of the interpolation kernel half-width and the time window length is dynamically generated by combining the fluctuation characteristics of each indicator in the historical disaster event samples, so as to enhance the response capability to areas of extreme change. The third step involves constructing a dual-objective optimization mechanism for the candidate parameter set, aiming to minimize interpolation error and maximize the preservation of fluctuations. This mechanism combines spatial interpolation effect with temporal aggregation sensitivity to perform parameter screening, extracting the optimal interpolation kernel half-width and time window length that can reconstruct the spatial structure without weakening the fluctuation signal, thus forming a piecewise optimal parameter set. The fourth step involves calling the interpolation kernel half-width from the optimal parameter set for each segment, performing spatial extrapolation along the normal direction for each point on the central axis, filling in missing positions, and outputting grid data covering the power corridor area. Then, the time window length is called to unify the time axis division, and the grid data is aggregated according to the time window to generate grid aggregate data. The values ​​reflecting geographic point data, infrastructure point data, and meteorological back-calculation point data in the grid aggregate data are standardized to eliminate dimensional differences and output a set of standardized evaluation indicators. The fifth step involves constructing a mapping structure between indicators and spatial segments based on a standardized set of evaluation indicators. Information entropy calculation is performed on the standardized values ​​of each type of indicator in all spatial segments to obtain the degree of dispersion of each type of indicator in all segments. The greater the degree of dispersion, the stronger its ability to distinguish different risk states. Then, information entropy weights are constructed based on information entropy, and weighted processing is performed on all indicator values ​​in each spatial segment to generate a disaster risk score that objectively reflects the actual impact of various indicators. The sixth step is to link the score values ​​of each segment with the geographical location, divide the level range according to the risk level threshold, and stitch them together to obtain a disaster risk map covering the entire power corridor. This disaster risk map can not only reflect the risk distribution of each segment, but also has spatial display capabilities, which facilitates the regulatory authorities to carry out key control and risk early warning deployment.

[0026] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A disaster risk assessment method for power corridors based on the entropy weight method, characterized in that, include: S1. Obtain the original multi-source point data related to the power corridor area and disaster risk assessment in Southwest China, unify the coordinates and align the time, obtain the original multi-source point dataset and perform quality control processing, check the consistency and stability of the original multi-source point dataset and remove missing values ​​and outliers, and output the multi-source point dataset. S2. Based on the multi-source point dataset, the facility coordinates are connected sequentially to construct the power corridor axis. On the power corridor axis, the tangential direction is determined according to the direction of the connection between adjacent facility coordinates, and the normal direction is determined according to the direction perpendicular to the tangential direction. The system is then segmented according to preset rules, and the segmented point dataset is output. S3. For each segment of data, establish reference conditions for the distribution characteristics and extreme value sorting of historical disaster-causing records within the segment of data, and set the information retention rate threshold and the relative order of peak values ​​as dual-threshold conditions. S4. Based on the distribution of facility coordinate spacing, the half-width of the interpolation kernel is determined, and a candidate set of interpolation kernel half-widths is obtained. A candidate set of time window lengths is constructed based on the duration of meteorological events. The optimal interpolation kernel half-width and time window length are selected under the condition of double gating using the safe trust region Bayesian optimization, and the piecewise optimal parameter set is output. S5. Based on the segmented optimal parameter set, interpolate the multi-source point dataset, convert the point data into grid data and aggregate it according to the corresponding time window length to obtain a standardized evaluation index set, and perform entropy weight method calculation to generate disaster risk assessment values, and summarize them to form a disaster zoning map of the power corridor.

2. The method for assessing the disaster risk of power corridors based on the entropy weight method according to claim 1, characterized in that: In S1, raw multi-source location data related to the disaster risk assessment of the power corridor area in Southwest China are obtained. The raw multi-source location data includes geographic location data, infrastructure location data and meteorological back calculation location data. Spatial reference system identifier resolution is performed on the original multi-source point data to extract the projection coordinate parameters of the original multi-source point data. Then, with the preset unified coordinates as the target, coordinate transformation operation is performed to unify the spatial coordinates of all the original multi-source point data to the preset unified coordinates and output the original multi-source point data under the unified coordinate system. Read the time attribute field of each original multi-source point data under a unified coordinate system, extract the timestamps of various types of original multi-source point data, sort the extracted timestamps in chronological order, and output the sorted timestamp sequence. Based on the sorted timestamp sequence, the lower limit time interval between adjacent timestamps is determined, and a unified time axis covering the entire research time range is constructed using the lower limit time interval as the step size. For raw multi-source point data with sampling frequency lower than the unified time axis, an interpolation method is used. The number of points before and after the missing point is taken as the benchmark. The ratio of the time interval between the missing point and the previous known point to the total time interval between the two known points is used as the scaling factor. Starting from the number of points of the previous known point, the product of the scaling factor and the difference in point values ​​between the two known points is superimposed to obtain the point value of the missing point. The low-frequency point data after interpolation is output. For original multi-source point data with a sampling frequency higher than the unified time axis, the sliding window width is an integer multiple of the lower limit time interval. The sliding window is used to slide sequentially on the original multi-source point data with a sampling frequency higher than the unified time axis. The average value of the point data in each sliding window is calculated, and the average value is used as the representative value of the corresponding time point of the corresponding sliding window. The high-frequency point data after the sliding average is output. The low-frequency and high-frequency point data are realigned according to the time nodes of a unified time axis, and the original multi-source point dataset aligned with the unified time axis is output.

3. The method for assessing the disaster risk of power corridors based on the entropy weight method according to claim 2, characterized in that: S1 also includes performing consistency checks on the original multi-source point dataset aligned with a unified time axis, extracting the unit system information of different point data for the same physical quantity, converting point values ​​with non-uniform units according to the preset International System of Units, and outputting point data with consistent units. The numerical range of point data with consistent units is checked, and a reasonable threshold range of physical quantities is called. When the point value exceeds the reasonable threshold range, it is marked as abnormal and removed. The point data that has passed the threshold test is output. The spatial distance and time interval between adjacent points of the data detected by the threshold are compared, and the change in unit distance or unit time is calculated. When the change exceeds the preset reasonable fluctuation range, it is judged as an unreasonable observation, the corresponding point is marked as abnormal and removed, and the point data after spatiotemporal reasonableness comparison is output. The point data that has passed the unit unification, threshold detection and spatiotemporal rationality comparison are summarized and the point dataset after consistency verification is output. The stability of the point dataset after consistency verification is checked. A fixed-length sliding window is constructed on the time series. The difference between the value of adjacent points within the fixed-length sliding window is calculated. When the difference exceeds a preset threshold, the corresponding point record is determined to be unstable and removed. The point dataset after stability verification is output. Missing values ​​are identified and processed in the point dataset after stability testing. Continuous missing intervals of point data are detected on a unified time axis, and the length of continuous missing intervals is compared with a preset threshold. When the length of consecutive missing values ​​is less than or equal to a preset threshold, interpolation is used to fill in the missing values. When the length of consecutive missing values ​​is greater than the preset threshold, the corresponding data points are directly removed, and a multi-source dataset for parameter optimization and power corridor disaster risk assessment is output.

4. The method for assessing the disaster risk of power corridors based on the entropy weight method according to claim 3, characterized in that: In S2, based on the multi-source point dataset, the coordinates of the facilities are read according to the arrangement order of the facilities on the power corridor, and the coordinates of adjacent facilities are connected in sequence to form a polyline composed of the lines connecting adjacent facilities, which serves as the central axis of the power corridor. For each segment of the central axis of the power corridor, a local coordinate system is established with the direction of the line connecting the start and end points as the tangential direction and the direction perpendicular to the tangential direction as the normal direction. Using the central axis of the power corridor as a reference line, and combining the tangential and normal directions, boundary lines are generated on both sides of the central axis at a preset equidistant offset distance, and the area enclosed by the central axis and the two boundary lines is taken as the permissible spatial domain of the power corridor. The permissible spatial domain of the power corridor is segmented according to preset rules, and the point datasets in each segment are aggregated to output the segmented point datasets.

5. The method for assessing the disaster risk of power corridors based on the entropy weight method according to claim 4, characterized in that: In S3, for the precipitation, temperature and wind speed in the segmented point data, the corresponding historical disaster-causing records are called, and the values ​​of each meteorological element at the time of disaster and the values ​​at the time of non-disaster are statistically analyzed to form disaster-triggered samples and non-disaster samples. The upper and lower limits of the observed values ​​of each meteorological element in the disaster-triggered sample and the non-disaster sample are identified respectively as the numerical interval. The observed values ​​of the corresponding meteorological elements within the numerical interval are added together and divided by the number of observations to obtain the average value. The numerical interval and the average value together constitute the distribution characteristics. Using the upper limit of the observed value of each meteorological element in the disaster-triggered sample as a reference threshold, a table of correspondence between each meteorological element and the disaster type is generated. In the correspondence table, the proportion of disaster-triggered samples to the total number of samples is used as the information retention rate threshold, and the order of appearance of the upper limit of disaster-triggered samples is used as the peak relative order constraint. The information retention rate threshold and the peak relative order constraint together constitute a double threshold condition.

6. The method for assessing the disaster risk of power corridors based on the entropy weight method according to claim 5, characterized in that: In S4, based on the spatial spacing between the coordinates of adjacent facilities in the segmented point data, all spatial spacing between adjacent facilities is collected to form a spacing sample set. The lower limit and upper limit of the spacing sample set are statistically analyzed to determine the distribution range of the spacing sample set. The mean of the spatial spacing within the distribution range is used as the candidate value of the interpolation kernel half-width to form the interpolation kernel half-width candidate set. Based on the duration of historical meteorological events corresponding to segmented point data, the start and end times of each meteorological event are identified and the duration of a single meteorological event is calculated. The duration of all meteorological events is collected to form a duration sample set. The lower and upper limits of the duration sample set are statistically determined to determine the distribution range of the duration sample set. The mean of the duration within the distribution range is used as a candidate value for the time window length, forming a candidate set of time window lengths.

7. The method for assessing the disaster risk of power corridors based on the entropy weight method according to claim 6, characterized in that: S4 also includes inputting the candidate set of interpolation kernel half-width and the candidate set of time window length into the safe trust domain Bayesian optimization architecture. The safe trust domain Bayesian optimization architecture includes a prediction distribution model that predicts the relationship between the parameter combination and the degree of satisfaction of the double gate limiting condition, as well as a safe boundary condition for limiting the range of parameter combination values. The average value of the interval sample set is used as the initial interpolation kernel half-width, and the average value of the meteorological event duration sample set is used as the initial time window length to form the initial parameter combination. The corresponding information retention rate and peak relative order are calculated, and the calculation results are used as the objective function value to input the prediction distribution model. In the predictive distribution model, the initial parameter combination is used as the benchmark, and the corresponding spatial spacing distribution range and meteorological event duration range are used as the search range. Several candidate parameter combinations are divided within the corresponding search range according to the preset step size. Calculate the expected objective function value for each candidate parameter combination, compare the candidate parameter combinations according to the magnitude of the expected objective function value, and select the candidate parameter combination with the upper limit of the expected objective function value as the new parameter combination; The information retention rate and peak relative order under the new parameter combination are calculated, and the calculation results are used as the new objective function to re-input into the prediction distribution model, triggering the calculation of the new parameter combination, eliminating candidate parameter combinations whose expected objective function value is lower than the preset threshold, narrowing the search range of parameter combinations, and completing the correction of the prediction distribution model. Repeat the correction process in the predicted distribution model until the calculated parameter combination no longer changes. Then, determine the corresponding parameter combination as the piecewise optimal parameter combination and output the piecewise optimal parameter set.

8. The method for assessing the disaster risk of power corridors based on the entropy weight method according to claim 7, characterized in that: In S5, the interpolation kernel half-width of the segmented optimal parameter set is called, and the difference method is used to expand and extrapolate the data of each point on the central axis in space. The missing positions are filled along the normal direction to obtain grid data covering the allowable domain of the power corridor space. Call the time window length in the optimal parameter set of segments, divide the continuous time window on the unified time axis, take the values ​​corresponding to the grid data falling into the same time window as the candidate set of the time window, perform an arithmetic mean on all values ​​in the candidate set, take the average as the aggregate representative value of the corresponding time window, and output the grid aggregate data under the corresponding time window. The values ​​of geographic location data, infrastructure location data and meteorological back-calculation location data in the grid aggregated data are used as indicators, and normalization processing is performed to eliminate the differences in the numerical range between different types of indicators, and output a standardized set of evaluation indicators. The weight of each standardized evaluation index in the standardized evaluation index set is calculated based on the entropy weight method. The weight of each index is multiplied by the corresponding standardized index value and accumulated to obtain the disaster risk evaluation value of the corresponding segment, and the segmented disaster risk sequence of the power corridor is output. The disaster risk assessment values ​​of each segment within the segmented disaster risk sequence are bound to their corresponding geographical coordinates and spliced ​​together within the permissible spatial domain of the power corridor to form a risk distribution sequence. Based on preset risk level thresholds, the risk distribution sequence is divided into multiple risk level intervals, and displayed in partitions corresponding to different risk levels to generate a disaster zoning map of the power corridor.

9. A power corridor disaster risk assessment system based on the entropy weight method, comprising applying the power corridor disaster risk assessment method based on the entropy weight method as described in claim 8, including a data quality control module, an axis construction module, a fidelity control module, a parameter optimization module, and a zoning output module, characterized in that: The data quality control module is used to acquire the original multi-source point data related to the power corridor area and disaster risk assessment in Southwest China, perform coordinate unification and time alignment, obtain the original multi-source point dataset and perform quality control processing, check the consistency and stability of the original multi-source point dataset, remove missing values ​​and outliers, and output the multi-source point dataset. The axis construction module is based on a multi-source point dataset. It connects the facility coordinates in sequence to construct the power corridor axis. On the power corridor axis, the tangential direction is determined according to the direction of the connection between adjacent facility coordinates, and the normal direction is determined according to the direction perpendicular to the tangential direction. The module is then segmented according to preset rules and outputs a segmented point dataset. The fidelity control module is used to establish reference conditions for the distribution characteristics and extreme value sorting of historical disaster-causing records within each segment of data, and to set the information retention rate threshold and the relative order of peak values ​​as dual-gate limiting conditions. The parameter optimization module determines the half-width of the interpolation kernel based on the distribution of facility coordinate spacing, obtains a candidate set of interpolation kernel half-widths, and constructs a candidate set of time window lengths based on the duration of meteorological events. It then uses safe trust region Bayesian optimization to select the optimal interpolation kernel half-width and time window length under double gating conditions, and outputs the piecewise optimal parameter set. The zoning output module interpolates the multi-source point dataset based on the segmented optimal parameter set, converts the point data into grid data, aggregates it according to the corresponding time window length, obtains a standardized evaluation index set, performs entropy weight calculation, generates disaster risk assessment values, and summarizes them to form a disaster zoning map of the power corridor.

Images